diff --git "a/inference_metadata.annotated.yaml" "b/inference_metadata.annotated.yaml" new file mode 100644--- /dev/null +++ "b/inference_metadata.annotated.yaml" @@ -0,0 +1,6975 @@ +# Annotated review companion for justinchuby/qwen2.5-14b-instruct-int4-zp-onnx. +# Model form: INT4 zero-point decoder; migrated from model.io to serialized pipeline.workflow. Per-layer sliding-window resolution verified; token-exact parity with a reference ORT decode. +# Canonical authority: inference_metadata.yaml at revision 753817320d232b0205a7971e8ea25068453fb393. +# Comments explain authored semantics, derived structure, and fail-closed omissions. +# After YAML parsing this file is mechanically equal to the canonical metadata; comments do not create a second contract. +# Backend implementation, recorded execution, and performance evidence remain separate from this portable metadata. +# Schema spelling `1.0` normalizes under the current reader; versioning gates unknown fields before typed parsing. No component authors batch_capacity, so independent requests must execute per item at every component. +schema_version: '1.0' +# The pipeline contains executable package semantics. Deployment/QoS policy and backend-derived execution plans live outside this portable contract. +pipeline: + # This typed SSA workflow is the authoritative execution contract for the package. + workflow: + # Manifest vocabulary declares interpreter/adapter ABI requirements. It is not ORT/native readiness or performance evidence. + manifest: + # Current v1 capability strings mirror typed workflow structure; the structure remains the semantic source of truth. + capabilities: + - workflow_ssa + - linear_effects + - nested_control_flow + - typed_emit + - emit_valid_length + - loop_induction_values + - serving_service_contract + - bounded_state_recurrence + # Workflow inputs are the complete typed boundary. Roles and sources are authored; absent required values must fail before execution. + inputs: + # Input `request.input_ids` is required, provided by the request through a typed runtime role, and carries runtime role prompt_tokens@1.0. Unsupported required roles must fail closed. + request.input_ids: + # Authored tensor ABI: dtype int64, rank 2, shape [batch, sequence_len]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 2 + shape: + - batch + - sequence_len + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: runtime + version: '1.0' + role: prompt_tokens + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: request + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: true + # Input `request.max_iterations` is required, provided by the request through a typed runtime role, and carries runtime role max_output_tokens@1.0. Unsupported required roles must fail closed. + request.max_iterations: + # Authored tensor ABI: dtype int64, rank 1, shape [1]. No grouping permission is implied by this shape. + contract: + dtype: int64 + rank: 1 + shape: + - 1 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: runtime + version: '1.0' + role: max_output_tokens + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: request + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: true + # Input `package.eos_ids` is required; omission uses semantic default 151645, a package-authored literal/default, and carries opaque. Unsupported required roles must fail closed. + package.eos_ids: + # Authored tensor ABI: dtype int64, rank 1, shape [E]. No grouping permission is implied by this shape. + contract: + dtype: int64 + rank: 1 + shape: + - E + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: literal + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: true + # This is a semantic tensor default used only when the optional input is omitted. + default: 151645 + # Input `package.one_token` is optional; omission uses semantic default 1, a package-authored literal/default, and carries opaque. Unsupported required roles must fail closed. + package.one_token: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: literal + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 1 + # Input `package.one_step` is optional; omission uses semantic default 1, a package-authored literal/default, and carries opaque. Unsupported required roles must fail closed. + package.one_step: + # Authored tensor ABI: dtype int64, rank 1, shape [1]. No grouping permission is implied by this shape. + contract: + dtype: int64 + rank: 1 + shape: + - 1 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: literal + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 1 + # Input `package.max_context` is optional; omission uses semantic default 32768, a package-authored literal/default, and carries opaque. Unsupported required roles must fail closed. + package.max_context: + # Authored tensor ABI: dtype int64, rank 1, shape [1]. No grouping permission is implied by this shape. + contract: + dtype: int64 + rank: 1 + shape: + - 1 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: literal + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 32768 + # Input `request.prompt_lengths` is optional; omission uses semantic default -1, provided by the application as prompt_lengths, and carries opaque. Unsupported required roles must fail closed. + request.prompt_lengths: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: application + name: prompt_lengths + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: -1 + # Input `request.eos_ids` is optional; omission uses semantic default 151645, provided by the application as eos_ids, and carries opaque. Unsupported required roles must fail closed. + request.eos_ids: + # Authored tensor ABI: dtype int64, rank 2, shape [batch, num_eos]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 2 + shape: + - batch + - num_eos + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: application + name: eos_ids + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 151645 + # Input `request.eos_lengths` is optional; omission uses semantic default 1, provided by the application as eos_lengths, and carries opaque. Unsupported required roles must fail closed. + request.eos_lengths: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: application + name: eos_lengths + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 1 + # Input `request.row_max_iterations` is optional; omission uses semantic default -1, provided by the application as row_max_iterations, and carries opaque. Unsupported required roles must fail closed. + request.row_max_iterations: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: application + name: row_max_iterations + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: -1 + # Input `request.temperature` is optional; omission uses semantic default 1.0, provided by the request through a typed runtime role, and carries runtime role sampling_temperature@1.0. Unsupported required roles must fail closed. + request.temperature: + # Authored tensor ABI: dtype float32, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float32 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: runtime + version: '1.0' + role: sampling_temperature + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: request + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 1.0 + # Input `request.top_k` is optional; omission uses semantic default 0, provided by the request through a typed runtime role, and carries runtime role sampling_top_k@1.0. Unsupported required roles must fail closed. + request.top_k: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: runtime + version: '1.0' + role: sampling_top_k + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: request + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 0 + # Input `request.top_p` is optional; omission uses semantic default 1.0, provided by the request through a typed runtime role, and carries runtime role sampling_top_p@1.0. Unsupported required roles must fail closed. + request.top_p: + # Authored tensor ABI: dtype float32, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float32 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: runtime + version: '1.0' + role: sampling_top_p + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: request + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 1.0 + # Input `request.min_p` is optional; omission uses semantic default 0.0, provided by the request through a typed runtime role, and carries runtime role sampling_min_p@1.0. Unsupported required roles must fail closed. + request.min_p: + # Authored tensor ABI: dtype float32, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float32 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: runtime + version: '1.0' + role: sampling_min_p + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: request + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 0.0 + # Input `request.seed` is optional; omission uses semantic default 0, provided by the request through a typed runtime role, and carries runtime role seed@1.0. Unsupported required roles must fail closed. + request.seed: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: runtime + version: '1.0' + role: seed + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: request + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 0 + # Input `request.rng_counter` is optional; omission uses semantic default 0, provided by the application as rng_counter, and carries opaque. Unsupported required roles must fail closed. + request.rng_counter: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: application + name: rng_counter + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 0 + # Input `package.active` is optional; omission uses semantic default True, a package-authored literal/default, and carries opaque. Unsupported required roles must fail closed. + package.active: + # Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: literal + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: true + # Input `package.not_done` is optional; omission uses semantic default False, a package-authored literal/default, and carries opaque. Unsupported required roles must fail closed. + package.not_done: + # Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: literal + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: false + # Input `package.cache_lengths` is optional; omission uses semantic default 0, a package-authored literal/default, and carries opaque. Unsupported required roles must fail closed. + package.cache_lengths: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: literal + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 0 + # Input `package.zero_batch` is optional; omission uses semantic default 0, a package-authored literal/default, and carries opaque. Unsupported required roles must fail closed. + package.zero_batch: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Role identifies runtime semantics without relying on a model-family name. + role: + kind: opaque + # Source states who must provide this value; the runtime must not invent application data. + source: + kind: literal + # Required values fail closed when absent; optional values may use only their authored default/presence rule. + required: false + # This is a semantic tensor default used only when the optional input is omitted. + default: 0 + # Workflow outputs are the complete public result boundary and retain explicit tensor contracts. + outputs: + # Output `tokens` publishes the authored tokens result at the pre_adapter stage; it is not backend-readiness evidence. + tokens: + # Authored tensor ABI: dtype int64, rank 2, shape [batch, generated_sequence]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 2 + shape: + - batch + - generated_sequence + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + role: tokens + # Stage records whether adapters/postprocessing have run before this value is exposed. + stage: pre_adapter + # Components are the executable artifacts and semantic policy programs. Their presence does not prove ORT/native readiness or performance. + components: + # Component `model` is a primary model compute graph using onnx from `model.onnx`. No batch_capacity is authored, so the runtime must not group independent requests at this component. + model: + # Selects a packaged implementation/artifact only; provider/backend readiness requires separate recorded evidence. + implementation: + kind: onnx + artifact: model.onnx + # Ports bind workflow names to the artifact's exact graph ABI; model-family inference is forbidden. + ports: + # Semantic port roles identify tokens, logits, positions, masks, caches, or media without architecture-name conditionals. + roles: + input_ids: token_ids + attention_mask: attention_mask + logits: logits + # Component `token_sampler` is an executable semantic policy component implementing `onnx-genai.token-sampler` using onnx from `policies/token_sampler.onnx`. Policy graphs compute tensors and semantics; they are not deployment/QoS knobs or backend readiness claims. No batch_capacity is authored, so the runtime must not group independent requests at this component. + token_sampler: + # Selects a packaged implementation/artifact only; provider/backend readiness requires separate recorded evidence. + implementation: + kind: onnx + artifact: policies/token_sampler.onnx + # Ports bind workflow names to the artifact's exact graph ABI; model-family inference is forbidden. + ports: + # Exact graph inputs; each tensor shape/dtype is authored. + inputs: + # Port `logits`: Authored tensor ABI: dtype float32, rank 2, shape [batch, vocabulary]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + logits: + dtype: float32 + rank: 2 + shape: + - batch + - vocabulary + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `temperature`: Authored tensor ABI: dtype float32, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + temperature: + dtype: float32 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `top_k`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + top_k: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `top_p`: Authored tensor ABI: dtype float32, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + top_p: + dtype: float32 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `min_p`: Authored tensor ABI: dtype float32, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + min_p: + dtype: float32 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `seed`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + seed: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `counter`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + counter: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `active`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + active: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `done`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + done: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Exact graph outputs; each tensor shape/dtype is authored. + outputs: + # Port `token`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + token: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `next_counter`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + next_counter: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Semantic component ABI `onnx-genai.token-sampler`; its bindings/parameters affect output correctness. + contract: + id: onnx-genai.token-sampler + version: '2' + bindings: + logits: logits + token: token + temperature: temperature + top_k: top_k + top_p: top_p + min_p: min_p + active: active + done: done + seed: seed + counter: counter + next_counter: next_counter + parameters: + mode: seeded_stochastic + batching: per_row + inactive_rows: preserve + # An override may replace this component only through the same validated semantic contract. + # Application replacement is allowed only through the same typed component contract. + application_overridable: true + # Component `termination` is an executable semantic policy component implementing `onnx-genai.termination-predicate` using onnx from `policies/termination.onnx`. Policy graphs compute tensors and semantics; they are not deployment/QoS knobs or backend readiness claims. No batch_capacity is authored, so the runtime must not group independent requests at this component. + termination: + # Selects a packaged implementation/artifact only; provider/backend readiness requires separate recorded evidence. + implementation: + kind: onnx + artifact: policies/termination.onnx + # Ports bind workflow names to the artifact's exact graph ABI; model-family inference is forbidden. + ports: + # Exact graph inputs; each tensor shape/dtype is authored. + inputs: + # Port `tokens`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + tokens: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `eos_ids`: Authored tensor ABI: dtype int64, rank 2, shape [batch, num_eos]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + eos_ids: + dtype: int64 + rank: 2 + shape: + - batch + - num_eos + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `eos_lengths`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + eos_lengths: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `iteration`: Authored tensor ABI: dtype int64, rank 1, shape [1]. No grouping permission is implied by this shape. + iteration: + dtype: int64 + rank: 1 + shape: + - 1 + # Port `max_iterations`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + max_iterations: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `active`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + active: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Exact graph outputs; each tensor shape/dtype is authored. + outputs: + # Port `done`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + done: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `next_active`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + next_active: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `continue`: Authored tensor ABI: dtype bool, rank 1, shape [1]. No grouping permission is implied by this shape. + continue: + dtype: bool + rank: 1 + shape: + - 1 + # Semantic component ABI `onnx-genai.termination-predicate`; its bindings/parameters affect output correctness. + contract: + id: onnx-genai.termination-predicate + version: '2' + bindings: + tokens: tokens + eos_ids: eos_ids + iteration: iteration + max_iterations: max_iterations + eos_lengths: eos_lengths + active: active + done: done + next_active: next_active + continue: continue + parameters: + batching: per_row + inactive_rows: preserve + # Component `token_state_update` is an executable semantic policy component implementing `onnx-genai.state-update` using onnx from `policies/token_state_update.onnx`. Policy graphs compute tensors and semantics; they are not deployment/QoS knobs or backend readiness claims. No batch_capacity is authored, so the runtime must not group independent requests at this component. + token_state_update: + # Selects a packaged implementation/artifact only; provider/backend readiness requires separate recorded evidence. + implementation: + kind: onnx + artifact: policies/token_state_update.onnx + # Ports bind workflow names to the artifact's exact graph ABI; model-family inference is forbidden. + ports: + # Exact graph inputs; each tensor shape/dtype is authored. + inputs: + # Port `current`: Authored tensor ABI: dtype int64, rank 2, shape [batch, 1]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + current: + dtype: int64 + rank: 2 + shape: + - batch + - 1 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `update`: Authored tensor ABI: dtype int64, rank 2, shape [batch, 1]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + update: + dtype: int64 + rank: 2 + shape: + - batch + - 1 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `active`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + active: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `done`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + done: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Exact graph outputs; each tensor shape/dtype is authored. + outputs: + # Port `next`: Authored tensor ABI: dtype int64, rank 2, shape [batch, 1]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + next: + dtype: int64 + rank: 2 + shape: + - batch + - 1 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Semantic component ABI `onnx-genai.state-update`; its bindings/parameters affect output correctness. + contract: + id: onnx-genai.state-update + version: '2' + bindings: + current: current + update: update + active: active + done: done + next: next + parameters: + batching: per_row + inactive_rows: preserve + # Component `last_token_logits` is a logit-selection policy graph using onnx from `policies/last_token_logits.onnx`. Policy graphs compute tensors and semantics; they are not deployment/QoS knobs or backend readiness claims. No batch_capacity is authored, so the runtime must not group independent requests at this component. + last_token_logits: + # Selects a packaged implementation/artifact only; provider/backend readiness requires separate recorded evidence. + implementation: + kind: onnx + artifact: policies/last_token_logits.onnx + # Ports bind workflow names to the artifact's exact graph ABI; model-family inference is forbidden. + ports: + # Exact graph inputs; each tensor shape/dtype is authored. + inputs: + # Port `logits`: Authored tensor ABI: dtype float16, rank 3, shape [batch, sequence, vocabulary]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + logits: + dtype: float16 + rank: 3 + shape: + - batch + - sequence + - vocabulary + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Exact graph outputs; each tensor shape/dtype is authored. + outputs: + # Port `last_logits`: Authored tensor ABI: dtype float32, rank 2, shape [batch, vocabulary]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + last_logits: + dtype: float32 + rank: 2 + shape: + - batch + - vocabulary + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Component `decoder_state_initializer` is a primary model compute graph using onnx from `policies/decoder_state_initializer.onnx`. No batch_capacity is authored, so the runtime must not group independent requests at this component. + decoder_state_initializer: + # Selects a packaged implementation/artifact only; provider/backend readiness requires separate recorded evidence. + implementation: + kind: onnx + artifact: policies/decoder_state_initializer.onnx + # Ports bind workflow names to the artifact's exact graph ABI; model-family inference is forbidden. + ports: + # Exact graph inputs; each tensor shape/dtype is authored. + inputs: + # Port `prompt_tokens`: Authored tensor ABI: dtype int64, rank 2, shape [batch, prompt_sequence]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + prompt_tokens: + dtype: int64 + rank: 2 + shape: + - batch + - prompt_sequence + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `prompt_lengths`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + prompt_lengths: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `max_iterations`: Authored tensor ABI: dtype int64, rank 1, shape [1]. No grouping permission is implied by this shape. + max_iterations: + dtype: int64 + rank: 1 + shape: + - 1 + # Exact graph outputs; each tensor shape/dtype is authored. + outputs: + # Port `attention_mask`: Authored tensor ABI: dtype int64, rank 2, shape [batch, capacity]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + attention_mask: + dtype: int64 + rank: 2 + shape: + - batch + - capacity + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `body_attention_mask`: Authored tensor ABI: dtype int64, rank 2, shape [batch, capacity]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + body_attention_mask: + dtype: int64 + rank: 2 + shape: + - batch + - capacity + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `token_slot`: Authored tensor ABI: dtype int64, rank 2, shape [batch, 1]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + token_slot: + dtype: int64 + rank: 2 + shape: + - batch + - 1 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `generated_lengths`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + generated_lengths: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `cache_lengths`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + cache_lengths: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.0.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.0.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.0.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.0.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.1.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.1.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.1.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.1.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.10.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.10.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.10.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.10.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.11.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.11.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.11.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.11.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.12.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.12.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.12.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.12.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.13.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.13.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.13.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.13.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.14.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.14.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.14.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.14.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.15.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.15.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.15.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.15.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.16.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.16.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.16.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.16.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.17.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.17.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.17.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.17.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.18.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.18.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.18.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.18.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.19.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.19.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.19.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.19.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.2.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.2.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.2.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.2.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.20.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.20.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.20.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.20.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.21.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.21.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.21.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.21.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.22.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.22.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.22.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.22.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.23.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.23.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.23.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.23.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.24.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.24.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.24.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.24.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.25.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.25.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.25.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.25.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.26.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.26.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.26.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.26.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.27.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.27.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.27.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.27.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.28.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.28.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.28.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.28.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.29.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.29.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.29.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.29.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.3.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.3.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.3.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.3.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.30.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.30.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.30.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.30.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.31.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.31.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.31.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.31.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.32.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.32.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.32.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.32.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.33.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.33.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.33.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.33.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.34.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.34.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.34.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.34.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.35.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.35.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.35.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.35.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.36.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.36.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.36.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.36.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.37.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.37.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.37.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.37.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.38.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.38.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.38.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.38.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.39.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.39.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.39.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.39.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.4.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.4.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.4.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.4.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.40.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.40.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.40.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.40.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.41.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.41.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.41.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.41.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.42.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.42.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.42.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.42.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.43.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.43.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.43.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.43.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.44.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.44.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.44.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.44.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.45.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.45.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.45.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.45.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.46.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.46.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.46.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.46.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.47.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.47.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.47.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.47.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.5.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.5.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.5.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.5.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.6.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.6.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.6.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.6.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.7.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.7.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.7.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.7.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.8.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.8.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.8.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.8.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.9.key`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.9.key: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `past_key_values.9.value`: Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, capacity, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + past_key_values.9.value: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - capacity + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Component `decoder_step_update` is a primary model compute graph using onnx from `policies/decoder_step_update.onnx`. No batch_capacity is authored, so the runtime must not group independent requests at this component. + decoder_step_update: + # Selects a packaged implementation/artifact only; provider/backend readiness requires separate recorded evidence. + implementation: + kind: onnx + artifact: policies/decoder_step_update.onnx + # Ports bind workflow names to the artifact's exact graph ABI; model-family inference is forbidden. + ports: + # Exact graph inputs; each tensor shape/dtype is authored. + inputs: + # Port `attention_mask`: Authored tensor ABI: dtype int64, rank 2, shape [batch, context]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + attention_mask: + dtype: int64 + rank: 2 + shape: + - batch + - context + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `logical_length`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + logical_length: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Exact graph outputs; each tensor shape/dtype is authored. + outputs: + # Port `next_attention_mask`: Authored tensor ABI: dtype int64, rank 2, shape [batch, context]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + next_attention_mask: + dtype: int64 + rank: 2 + shape: + - batch + - context + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Component `cache_length_update` is a state-transition policy graph using onnx from `policies/cache_length_update.onnx`. Policy graphs compute tensors and semantics; they are not deployment/QoS knobs or backend readiness claims. No batch_capacity is authored, so the runtime must not group independent requests at this component. + cache_length_update: + # Selects a packaged implementation/artifact only; provider/backend readiness requires separate recorded evidence. + implementation: + kind: onnx + artifact: policies/cache_length_update.onnx + # Ports bind workflow names to the artifact's exact graph ABI; model-family inference is forbidden. + ports: + # Exact graph inputs; each tensor shape/dtype is authored. + inputs: + # Port `left`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + left: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `right`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + right: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `active`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + active: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `done`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + done: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Exact graph outputs; each tensor shape/dtype is authored. + outputs: + # Port `total`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + total: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Component `termination_batch_initializer` is a termination policy graph using onnx from `policies/termination_batch_initializer.onnx`. Policy graphs compute tensors and semantics; they are not deployment/QoS knobs or backend readiness claims. No batch_capacity is authored, so the runtime must not group independent requests at this component. + termination_batch_initializer: + # Selects a packaged implementation/artifact only; provider/backend readiness requires separate recorded evidence. + implementation: + kind: onnx + artifact: policies/termination_batch_initializer.onnx + # Ports bind workflow names to the artifact's exact graph ABI; model-family inference is forbidden. + ports: + # Exact graph inputs; each tensor shape/dtype is authored. + inputs: + # Port `input_eos_ids`: Authored tensor ABI: dtype int64, rank 2, shape [batch, num_eos]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + input_eos_ids: + dtype: int64 + rank: 2 + shape: + - batch + - num_eos + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `input_eos_lengths`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + input_eos_lengths: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `input_max_iterations`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + input_max_iterations: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `fallback_max_iterations`: Authored tensor ABI: dtype int64, rank 1, shape [1]. No grouping permission is implied by this shape. + fallback_max_iterations: + dtype: int64 + rank: 1 + shape: + - 1 + # Port `active`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + active: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Exact graph outputs; each tensor shape/dtype is authored. + outputs: + # Port `row_eos_ids`: Authored tensor ABI: dtype int64, rank 2, shape [batch, num_eos]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + row_eos_ids: + dtype: int64 + rank: 2 + shape: + - batch + - num_eos + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `eos_lengths`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + eos_lengths: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `max_iterations`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + max_iterations: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Component `token_to_slot` is an authored workflow compute component using onnx from `policies/token_to_slot.onnx`. No batch_capacity is authored, so the runtime must not group independent requests at this component. + token_to_slot: + # Selects a packaged implementation/artifact only; provider/backend readiness requires separate recorded evidence. + implementation: + kind: onnx + artifact: policies/token_to_slot.onnx + # Ports bind workflow names to the artifact's exact graph ABI; model-family inference is forbidden. + ports: + # Exact graph inputs; each tensor shape/dtype is authored. + inputs: + # Port `token`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + token: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Exact graph outputs; each tensor shape/dtype is authored. + outputs: + # Port `slot`: Authored tensor ABI: dtype int64, rank 2, shape [batch, 1]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + slot: + dtype: int64 + rank: 2 + shape: + - batch + - 1 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Component `generated_length_update` is a state-transition policy graph using onnx from `policies/generated_length_update.onnx`. Policy graphs compute tensors and semantics; they are not deployment/QoS knobs or backend readiness claims. No batch_capacity is authored, so the runtime must not group independent requests at this component. + generated_length_update: + # Selects a packaged implementation/artifact only; provider/backend readiness requires separate recorded evidence. + implementation: + kind: onnx + artifact: policies/generated_length_update.onnx + # Ports bind workflow names to the artifact's exact graph ABI; model-family inference is forbidden. + ports: + # Exact graph inputs; each tensor shape/dtype is authored. + inputs: + # Port `left`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + left: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `right`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + right: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `active`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + active: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Port `done`: Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + done: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # Exact graph outputs; each tensor shape/dtype is authored. + outputs: + # Port `total`: Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + total: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # State cells make recurrence, cache mutation, lifetime, and ownership explicit. Physical paging/tiering/placement are not encoded by tensor names. + state: + # State `token` is semantic tensor carry between workflow steps. Scope=invocation, recurrence=invariant. The authored recurrence determines how it evolves. + token: + # Authored tensor ABI: dtype int64, rank 2, shape [batch, 1]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 2 + shape: + - batch + - 1 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: initializer.token_slot + # Recurrence defines the next-value rule: {kind=invariant}. + recurrence: + kind: invariant + # State `logits` is semantic tensor carry between workflow steps. Scope=invocation, recurrence=invariant. The authored recurrence determines how it evolves. + logits: + # Authored tensor ABI: dtype float32, rank 2, shape [batch, 152064]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float32 + rank: 2 + shape: + - batch + - 152064 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.last_logits + # Recurrence defines the next-value rule: {kind=invariant}. + recurrence: + kind: invariant + # State `generated_lengths` is loop/scheduler control state. Scope=invocation, recurrence=invariant. The values affect correctness, not merely performance. + generated_lengths: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # State class distinguishes semantic state from advisory/runtime-only bookkeeping. + class: semantic + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: initializer.generated_lengths + # Recurrence defines the next-value rule: {kind=invariant}. + recurrence: + kind: invariant + # State `active` is loop/scheduler control state. Scope=invocation, recurrence=invariant. The values affect correctness, not merely performance. + active: + # Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # State class distinguishes semantic state from advisory/runtime-only bookkeeping. + class: semantic + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: package.active + # Recurrence defines the next-value rule: {kind=invariant}. + recurrence: + kind: invariant + # State `done` is loop/scheduler control state. Scope=invocation, recurrence=invariant. The values affect correctness, not merely performance. + done: + # Authored tensor ABI: dtype bool, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: bool + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # State class distinguishes semantic state from advisory/runtime-only bookkeeping. + class: semantic + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: package.not_done + # Recurrence defines the next-value rule: {kind=invariant}. + recurrence: + kind: invariant + # State `accepted_len` is loop/scheduler control state. Scope=invocation, recurrence=invariant. The values affect correctness, not merely performance. + accepted_len: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # State class distinguishes semantic state from advisory/runtime-only bookkeeping. + class: semantic + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: package.zero_batch + # Recurrence defines the next-value rule: {kind=invariant}. + recurrence: + kind: invariant + # State `cache_lengths` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=invariant. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_lengths: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # State class distinguishes semantic state from advisory/runtime-only bookkeeping. + class: semantic + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: initializer.cache_lengths + # Recurrence defines the next-value rule: {kind=invariant}. + recurrence: + kind: invariant + # State `rng_counter` is deterministic random-stream state. Scope=invocation, recurrence=invariant. Its recurrence preserves reproducibility across loop steps. + rng_counter: + # Authored tensor ABI: dtype int64, rank 1, shape [batch]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 1 + shape: + - batch + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # State class distinguishes semantic state from advisory/runtime-only bookkeeping. + class: semantic + # Initializer names the exact first SSA value for this state cell. + initializer: request.rng_counter + # Recurrence defines the next-value rule: {kind=invariant}. + recurrence: + kind: invariant + # State `attention_mask` is semantic tensor carry between workflow steps. Scope=invocation, recurrence=invariant. The authored recurrence determines how it evolves. + attention_mask: + # Authored tensor ABI: dtype int64, rank 2, shape [batch, context]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: int64 + rank: 2 + shape: + - batch + - context + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: initializer.attention_mask + # Recurrence defines the next-value rule: {kind=invariant}. + recurrence: + kind: invariant + # State `cache_0` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_0: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.0.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_1` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_1: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.0.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_2` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_2: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.1.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_3` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_3: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.1.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_4` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_4: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.2.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_5` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_5: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.2.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_6` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_6: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.3.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_7` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_7: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.3.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_8` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_8: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.4.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_9` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_9: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.4.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_10` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_10: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.5.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_11` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_11: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.5.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_12` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_12: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.6.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_13` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_13: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.6.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_14` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_14: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.7.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_15` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_15: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.7.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_16` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_16: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.8.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_17` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_17: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.8.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_18` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_18: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.9.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_19` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_19: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.9.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_20` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_20: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.10.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_21` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_21: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.10.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_22` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_22: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.11.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_23` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_23: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.11.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_24` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_24: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.12.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_25` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_25: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.12.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_26` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_26: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.13.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_27` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_27: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.13.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_28` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_28: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.14.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_29` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_29: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.14.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_30` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_30: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.15.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_31` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_31: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.15.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_32` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_32: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.16.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_33` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_33: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.16.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_34` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_34: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.17.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_35` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_35: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.17.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_36` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_36: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.18.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_37` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_37: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.18.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_38` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_38: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.19.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_39` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_39: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.19.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_40` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_40: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.20.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_41` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_41: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.20.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_42` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_42: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.21.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_43` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_43: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.21.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_44` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_44: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.22.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_45` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_45: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.22.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_46` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_46: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.23.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_47` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_47: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.23.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_48` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_48: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.24.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_49` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_49: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.24.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_50` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_50: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.25.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_51` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_51: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.25.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_52` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_52: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.26.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_53` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_53: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.26.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_54` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_54: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.27.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_55` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_55: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.27.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_56` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_56: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.28.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_57` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_57: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.28.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_58` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_58: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.29.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_59` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_59: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.29.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_60` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_60: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.30.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_61` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_61: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.30.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_62` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_62: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.31.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_63` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_63: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.31.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_64` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_64: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.32.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_65` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_65: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.32.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_66` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_66: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.33.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_67` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_67: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.33.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_68` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_68: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.34.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_69` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_69: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.34.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_70` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_70: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.35.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_71` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_71: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.35.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_72` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_72: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.36.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_73` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_73: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.36.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_74` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_74: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.37.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_75` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_75: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.37.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_76` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_76: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.38.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_77` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_77: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.38.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_78` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_78: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.39.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_79` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_79: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.39.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_80` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_80: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.40.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_81` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_81: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.40.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_82` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_82: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.41.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_83` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_83: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.41.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_84` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_84: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.42.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_85` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_85: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.42.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_86` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_86: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.43.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_87` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_87: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.43.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_88` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_88: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.44.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_89` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_89: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.44.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_90` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_90: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.45.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_91` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_91: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.45.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_92` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_92: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.46.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_93` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_93: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.46.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_94` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_94: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.47.key + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # State `cache_95` is semantic key/value or recurrent cache state. Scope=invocation, recurrence=bounded. Physical paging, tiering, prefix storage, and mirror dtype remain runtime policy. + cache_95: + # Authored tensor ABI: dtype float16, rank 4, shape [batch, 8, past_sequence_len, 128]. Request rows stay aligned on axis 0; this preserves identity but does not by itself authorize grouping independent requests. + contract: + dtype: float16 + rank: 4 + shape: + - batch + - 8 + - past_sequence_len + - 128 + # Tensor row layout is structural. Only component batch_capacity can authorize grouping independent requests. + batch_layout: + kind: request_aligned + axis: 0 + # `invocation` defines the semantic lifetime boundary; it is not a worker-placement policy. + scope: invocation + # Initializer names the exact first SSA value for this state cell. + initializer: decoder.setup.present.47.value + # Recurrence defines the next-value rule: {kind=bounded, axis=2, max=package.max_context}. + recurrence: + kind: bounded + axis: 2 + max: package.max_context + # Runtime management means the engine preserves this semantic state; storage layout remains backend policy. + management: runtime + # The state may be released only at this correctness boundary. + release_boundary: invocation + # Service-group membership links this cell to the typed cache/state service below. + service_group: decoder_cache + # Serving binds workflow tensors to scheduler/session semantics. It does not choose worker count, provider, QoS, or claim thread safety. + serving: + # Active-row binding tells the scheduler which request rows still execute. + active: active + # Done-row binding is the semantic termination signal. + done: done + # Accepted-length binding controls cache/state advancement after verification. + accepted_len: accepted_len + # State service groups connect semantic cells to cache/recurrent update contracts used by serving admission. + state_service: + # Each group declares semantic update/rollback behavior; physical cache storage remains deployment/runtime policy. + groups: + # Group `decoder_cache` is semantic `full_attention` state in `bnsh` layout with update None; rollback/fork/snapshot are correctness capabilities. + decoder_cache: + kind: full_attention + # Sequence axis identifies logical history growth for append/rollback validation. + sequence_axis: 2 + # Layout names tensor axis order, not physical paging or device placement. + layout: bnsh + logical_lengths: cache_lengths + aliasing: permitted + reuse: + prefix_reusable: true + evictable_prefix: false + # Port bindings connect service state to exact component input/output names. + ports: + model: + cache_0: + input: past_key_values.0.key + output: present.0.key + role: key + layer: 0 + cache_1: + input: past_key_values.0.value + output: present.0.value + role: value + layer: 0 + cache_2: + input: past_key_values.1.key + output: present.1.key + role: key + layer: 1 + cache_3: + input: past_key_values.1.value + output: present.1.value + role: value + layer: 1 + cache_4: + input: past_key_values.2.key + output: present.2.key + role: key + layer: 2 + cache_5: + input: past_key_values.2.value + output: present.2.value + role: value + layer: 2 + cache_6: + input: past_key_values.3.key + output: present.3.key + role: key + layer: 3 + cache_7: + input: past_key_values.3.value + output: present.3.value + role: value + layer: 3 + cache_8: + input: past_key_values.4.key + output: present.4.key + role: key + layer: 4 + cache_9: + input: past_key_values.4.value + output: present.4.value + role: value + layer: 4 + cache_10: + input: past_key_values.5.key + output: present.5.key + role: key + layer: 5 + cache_11: + input: past_key_values.5.value + output: present.5.value + role: value + layer: 5 + cache_12: + input: past_key_values.6.key + output: present.6.key + role: key + layer: 6 + cache_13: + input: past_key_values.6.value + output: present.6.value + role: value + layer: 6 + cache_14: + input: past_key_values.7.key + output: present.7.key + role: key + layer: 7 + cache_15: + input: past_key_values.7.value + output: present.7.value + role: value + layer: 7 + cache_16: + input: past_key_values.8.key + output: present.8.key + role: key + layer: 8 + cache_17: + input: past_key_values.8.value + output: present.8.value + role: value + layer: 8 + cache_18: + input: past_key_values.9.key + output: present.9.key + role: key + layer: 9 + cache_19: + input: past_key_values.9.value + output: present.9.value + role: value + layer: 9 + cache_20: + input: past_key_values.10.key + output: present.10.key + role: key + layer: 10 + cache_21: + input: past_key_values.10.value + output: present.10.value + role: value + layer: 10 + cache_22: + input: past_key_values.11.key + output: present.11.key + role: key + layer: 11 + cache_23: + input: past_key_values.11.value + output: present.11.value + role: value + layer: 11 + cache_24: + input: past_key_values.12.key + output: present.12.key + role: key + layer: 12 + cache_25: + input: past_key_values.12.value + output: present.12.value + role: value + layer: 12 + cache_26: + input: past_key_values.13.key + output: present.13.key + role: key + layer: 13 + cache_27: + input: past_key_values.13.value + output: present.13.value + role: value + layer: 13 + cache_28: + input: past_key_values.14.key + output: present.14.key + role: key + layer: 14 + cache_29: + input: past_key_values.14.value + output: present.14.value + role: value + layer: 14 + cache_30: + input: past_key_values.15.key + output: present.15.key + role: key + layer: 15 + cache_31: + input: past_key_values.15.value + output: present.15.value + role: value + layer: 15 + cache_32: + input: past_key_values.16.key + output: present.16.key + role: key + layer: 16 + cache_33: + input: past_key_values.16.value + output: present.16.value + role: value + layer: 16 + cache_34: + input: past_key_values.17.key + output: present.17.key + role: key + layer: 17 + cache_35: + input: past_key_values.17.value + output: present.17.value + role: value + layer: 17 + cache_36: + input: past_key_values.18.key + output: present.18.key + role: key + layer: 18 + cache_37: + input: past_key_values.18.value + output: present.18.value + role: value + layer: 18 + cache_38: + input: past_key_values.19.key + output: present.19.key + role: key + layer: 19 + cache_39: + input: past_key_values.19.value + output: present.19.value + role: value + layer: 19 + cache_40: + input: past_key_values.20.key + output: present.20.key + role: key + layer: 20 + cache_41: + input: past_key_values.20.value + output: present.20.value + role: value + layer: 20 + cache_42: + input: past_key_values.21.key + output: present.21.key + role: key + layer: 21 + cache_43: + input: past_key_values.21.value + output: present.21.value + role: value + layer: 21 + cache_44: + input: past_key_values.22.key + output: present.22.key + role: key + layer: 22 + cache_45: + input: past_key_values.22.value + output: present.22.value + role: value + layer: 22 + cache_46: + input: past_key_values.23.key + output: present.23.key + role: key + layer: 23 + cache_47: + input: past_key_values.23.value + output: present.23.value + role: value + layer: 23 + cache_48: + input: past_key_values.24.key + output: present.24.key + role: key + layer: 24 + cache_49: + input: past_key_values.24.value + output: present.24.value + role: value + layer: 24 + cache_50: + input: past_key_values.25.key + output: present.25.key + role: key + layer: 25 + cache_51: + input: past_key_values.25.value + output: present.25.value + role: value + layer: 25 + cache_52: + input: past_key_values.26.key + output: present.26.key + role: key + layer: 26 + cache_53: + input: past_key_values.26.value + output: present.26.value + role: value + layer: 26 + cache_54: + input: past_key_values.27.key + output: present.27.key + role: key + layer: 27 + cache_55: + input: past_key_values.27.value + output: present.27.value + role: value + layer: 27 + cache_56: + input: past_key_values.28.key + output: present.28.key + role: key + layer: 28 + cache_57: + input: past_key_values.28.value + output: present.28.value + role: value + layer: 28 + cache_58: + input: past_key_values.29.key + output: present.29.key + role: key + layer: 29 + cache_59: + input: past_key_values.29.value + output: present.29.value + role: value + layer: 29 + cache_60: + input: past_key_values.30.key + output: present.30.key + role: key + layer: 30 + cache_61: + input: past_key_values.30.value + output: present.30.value + role: value + layer: 30 + cache_62: + input: past_key_values.31.key + output: present.31.key + role: key + layer: 31 + cache_63: + input: past_key_values.31.value + output: present.31.value + role: value + layer: 31 + cache_64: + input: past_key_values.32.key + output: present.32.key + role: key + layer: 32 + cache_65: + input: past_key_values.32.value + output: present.32.value + role: value + layer: 32 + cache_66: + input: past_key_values.33.key + output: present.33.key + role: key + layer: 33 + cache_67: + input: past_key_values.33.value + output: present.33.value + role: value + layer: 33 + cache_68: + input: past_key_values.34.key + output: present.34.key + role: key + layer: 34 + cache_69: + input: past_key_values.34.value + output: present.34.value + role: value + layer: 34 + cache_70: + input: past_key_values.35.key + output: present.35.key + role: key + layer: 35 + cache_71: + input: past_key_values.35.value + output: present.35.value + role: value + layer: 35 + cache_72: + input: past_key_values.36.key + output: present.36.key + role: key + layer: 36 + cache_73: + input: past_key_values.36.value + output: present.36.value + role: value + layer: 36 + cache_74: + input: past_key_values.37.key + output: present.37.key + role: key + layer: 37 + cache_75: + input: past_key_values.37.value + output: present.37.value + role: value + layer: 37 + cache_76: + input: past_key_values.38.key + output: present.38.key + role: key + layer: 38 + cache_77: + input: past_key_values.38.value + output: present.38.value + role: value + layer: 38 + cache_78: + input: past_key_values.39.key + output: present.39.key + role: key + layer: 39 + cache_79: + input: past_key_values.39.value + output: present.39.value + role: value + layer: 39 + cache_80: + input: past_key_values.40.key + output: present.40.key + role: key + layer: 40 + cache_81: + input: past_key_values.40.value + output: present.40.value + role: value + layer: 40 + cache_82: + input: past_key_values.41.key + output: present.41.key + role: key + layer: 41 + cache_83: + input: past_key_values.41.value + output: present.41.value + role: value + layer: 41 + cache_84: + input: past_key_values.42.key + output: present.42.key + role: key + layer: 42 + cache_85: + input: past_key_values.42.value + output: present.42.value + role: value + layer: 42 + cache_86: + input: past_key_values.43.key + output: present.43.key + role: key + layer: 43 + cache_87: + input: past_key_values.43.value + output: present.43.value + role: value + layer: 43 + cache_88: + input: past_key_values.44.key + output: present.44.key + role: key + layer: 44 + cache_89: + input: past_key_values.44.value + output: present.44.value + role: value + layer: 44 + cache_90: + input: past_key_values.45.key + output: present.45.key + role: key + layer: 45 + cache_91: + input: past_key_values.45.value + output: present.45.value + role: value + layer: 45 + cache_92: + input: past_key_values.46.key + output: present.46.key + role: key + layer: 46 + cache_93: + input: past_key_values.46.value + output: present.46.value + role: value + layer: 46 + cache_94: + input: past_key_values.47.key + output: present.47.key + role: key + layer: 47 + cache_95: + input: past_key_values.47.value + output: present.47.value + role: value + layer: 47 + # The ordered SSA program defines actual execution and dataflow; capability strings are not a substitute for these steps. + steps: + # Run a bounded semantic loop (request.max_iterations); carried state and termination values make progress explicit and fail closed if unsupported. + - kind: loop + setup: + # Invoke component `decoder_state_initializer` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: decoder_state_initializer + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + prompt_tokens: request.input_ids + prompt_lengths: request.prompt_lengths + max_iterations: request.max_iterations + # SSA output names become the only values subsequent steps may consume. + outputs: + attention_mask: initializer.attention_mask + body_attention_mask: initializer.body_attention_mask + token_slot: initializer.token_slot + generated_lengths: initializer.generated_lengths + cache_lengths: initializer.cache_lengths + past_key_values.0.key: initializer.past_key_values.0.key + past_key_values.0.value: initializer.past_key_values.0.value + past_key_values.1.key: initializer.past_key_values.1.key + past_key_values.1.value: initializer.past_key_values.1.value + past_key_values.10.key: initializer.past_key_values.10.key + past_key_values.10.value: initializer.past_key_values.10.value + past_key_values.11.key: initializer.past_key_values.11.key + past_key_values.11.value: initializer.past_key_values.11.value + past_key_values.12.key: initializer.past_key_values.12.key + past_key_values.12.value: initializer.past_key_values.12.value + past_key_values.13.key: initializer.past_key_values.13.key + past_key_values.13.value: initializer.past_key_values.13.value + past_key_values.14.key: initializer.past_key_values.14.key + past_key_values.14.value: initializer.past_key_values.14.value + past_key_values.15.key: initializer.past_key_values.15.key + past_key_values.15.value: initializer.past_key_values.15.value + past_key_values.16.key: initializer.past_key_values.16.key + past_key_values.16.value: initializer.past_key_values.16.value + past_key_values.17.key: initializer.past_key_values.17.key + past_key_values.17.value: initializer.past_key_values.17.value + past_key_values.18.key: initializer.past_key_values.18.key + past_key_values.18.value: initializer.past_key_values.18.value + past_key_values.19.key: initializer.past_key_values.19.key + past_key_values.19.value: initializer.past_key_values.19.value + past_key_values.2.key: initializer.past_key_values.2.key + past_key_values.2.value: initializer.past_key_values.2.value + past_key_values.20.key: initializer.past_key_values.20.key + past_key_values.20.value: initializer.past_key_values.20.value + past_key_values.21.key: initializer.past_key_values.21.key + past_key_values.21.value: initializer.past_key_values.21.value + past_key_values.22.key: initializer.past_key_values.22.key + past_key_values.22.value: initializer.past_key_values.22.value + past_key_values.23.key: initializer.past_key_values.23.key + past_key_values.23.value: initializer.past_key_values.23.value + past_key_values.24.key: initializer.past_key_values.24.key + past_key_values.24.value: initializer.past_key_values.24.value + past_key_values.25.key: initializer.past_key_values.25.key + past_key_values.25.value: initializer.past_key_values.25.value + past_key_values.26.key: initializer.past_key_values.26.key + past_key_values.26.value: initializer.past_key_values.26.value + past_key_values.27.key: initializer.past_key_values.27.key + past_key_values.27.value: initializer.past_key_values.27.value + past_key_values.28.key: initializer.past_key_values.28.key + past_key_values.28.value: initializer.past_key_values.28.value + past_key_values.29.key: initializer.past_key_values.29.key + past_key_values.29.value: initializer.past_key_values.29.value + past_key_values.3.key: initializer.past_key_values.3.key + past_key_values.3.value: initializer.past_key_values.3.value + past_key_values.30.key: initializer.past_key_values.30.key + past_key_values.30.value: initializer.past_key_values.30.value + past_key_values.31.key: initializer.past_key_values.31.key + past_key_values.31.value: initializer.past_key_values.31.value + past_key_values.32.key: initializer.past_key_values.32.key + past_key_values.32.value: initializer.past_key_values.32.value + past_key_values.33.key: initializer.past_key_values.33.key + past_key_values.33.value: initializer.past_key_values.33.value + past_key_values.34.key: initializer.past_key_values.34.key + past_key_values.34.value: initializer.past_key_values.34.value + past_key_values.35.key: initializer.past_key_values.35.key + past_key_values.35.value: initializer.past_key_values.35.value + past_key_values.36.key: initializer.past_key_values.36.key + past_key_values.36.value: initializer.past_key_values.36.value + past_key_values.37.key: initializer.past_key_values.37.key + past_key_values.37.value: initializer.past_key_values.37.value + past_key_values.38.key: initializer.past_key_values.38.key + past_key_values.38.value: initializer.past_key_values.38.value + past_key_values.39.key: initializer.past_key_values.39.key + past_key_values.39.value: initializer.past_key_values.39.value + past_key_values.4.key: initializer.past_key_values.4.key + past_key_values.4.value: initializer.past_key_values.4.value + past_key_values.40.key: initializer.past_key_values.40.key + past_key_values.40.value: initializer.past_key_values.40.value + past_key_values.41.key: initializer.past_key_values.41.key + past_key_values.41.value: initializer.past_key_values.41.value + past_key_values.42.key: initializer.past_key_values.42.key + past_key_values.42.value: initializer.past_key_values.42.value + past_key_values.43.key: initializer.past_key_values.43.key + past_key_values.43.value: initializer.past_key_values.43.value + past_key_values.44.key: initializer.past_key_values.44.key + past_key_values.44.value: initializer.past_key_values.44.value + past_key_values.45.key: initializer.past_key_values.45.key + past_key_values.45.value: initializer.past_key_values.45.value + past_key_values.46.key: initializer.past_key_values.46.key + past_key_values.46.value: initializer.past_key_values.46.value + past_key_values.47.key: initializer.past_key_values.47.key + past_key_values.47.value: initializer.past_key_values.47.value + past_key_values.5.key: initializer.past_key_values.5.key + past_key_values.5.value: initializer.past_key_values.5.value + past_key_values.6.key: initializer.past_key_values.6.key + past_key_values.6.value: initializer.past_key_values.6.value + past_key_values.7.key: initializer.past_key_values.7.key + past_key_values.7.value: initializer.past_key_values.7.value + past_key_values.8.key: initializer.past_key_values.8.key + past_key_values.8.value: initializer.past_key_values.8.value + past_key_values.9.key: initializer.past_key_values.9.key + past_key_values.9.value: initializer.past_key_values.9.value + # Invoke component `model` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: model + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + input_ids: request.input_ids + past_key_values.0.key: initializer.past_key_values.0.key + past_key_values.0.value: initializer.past_key_values.0.value + past_key_values.1.key: initializer.past_key_values.1.key + past_key_values.1.value: initializer.past_key_values.1.value + past_key_values.2.key: initializer.past_key_values.2.key + past_key_values.2.value: initializer.past_key_values.2.value + past_key_values.3.key: initializer.past_key_values.3.key + past_key_values.3.value: initializer.past_key_values.3.value + past_key_values.4.key: initializer.past_key_values.4.key + past_key_values.4.value: initializer.past_key_values.4.value + past_key_values.5.key: initializer.past_key_values.5.key + past_key_values.5.value: initializer.past_key_values.5.value + past_key_values.6.key: initializer.past_key_values.6.key + past_key_values.6.value: initializer.past_key_values.6.value + past_key_values.7.key: initializer.past_key_values.7.key + past_key_values.7.value: initializer.past_key_values.7.value + past_key_values.8.key: initializer.past_key_values.8.key + past_key_values.8.value: initializer.past_key_values.8.value + past_key_values.9.key: initializer.past_key_values.9.key + past_key_values.9.value: initializer.past_key_values.9.value + past_key_values.10.key: initializer.past_key_values.10.key + past_key_values.10.value: initializer.past_key_values.10.value + past_key_values.11.key: initializer.past_key_values.11.key + past_key_values.11.value: initializer.past_key_values.11.value + past_key_values.12.key: initializer.past_key_values.12.key + past_key_values.12.value: initializer.past_key_values.12.value + past_key_values.13.key: initializer.past_key_values.13.key + past_key_values.13.value: initializer.past_key_values.13.value + past_key_values.14.key: initializer.past_key_values.14.key + past_key_values.14.value: initializer.past_key_values.14.value + past_key_values.15.key: initializer.past_key_values.15.key + past_key_values.15.value: initializer.past_key_values.15.value + past_key_values.16.key: initializer.past_key_values.16.key + past_key_values.16.value: initializer.past_key_values.16.value + past_key_values.17.key: initializer.past_key_values.17.key + past_key_values.17.value: initializer.past_key_values.17.value + past_key_values.18.key: initializer.past_key_values.18.key + past_key_values.18.value: initializer.past_key_values.18.value + past_key_values.19.key: initializer.past_key_values.19.key + past_key_values.19.value: initializer.past_key_values.19.value + past_key_values.20.key: initializer.past_key_values.20.key + past_key_values.20.value: initializer.past_key_values.20.value + past_key_values.21.key: initializer.past_key_values.21.key + past_key_values.21.value: initializer.past_key_values.21.value + past_key_values.22.key: initializer.past_key_values.22.key + past_key_values.22.value: initializer.past_key_values.22.value + past_key_values.23.key: initializer.past_key_values.23.key + past_key_values.23.value: initializer.past_key_values.23.value + past_key_values.24.key: initializer.past_key_values.24.key + past_key_values.24.value: initializer.past_key_values.24.value + past_key_values.25.key: initializer.past_key_values.25.key + past_key_values.25.value: initializer.past_key_values.25.value + past_key_values.26.key: initializer.past_key_values.26.key + past_key_values.26.value: initializer.past_key_values.26.value + past_key_values.27.key: initializer.past_key_values.27.key + past_key_values.27.value: initializer.past_key_values.27.value + past_key_values.28.key: initializer.past_key_values.28.key + past_key_values.28.value: initializer.past_key_values.28.value + past_key_values.29.key: initializer.past_key_values.29.key + past_key_values.29.value: initializer.past_key_values.29.value + past_key_values.30.key: initializer.past_key_values.30.key + past_key_values.30.value: initializer.past_key_values.30.value + past_key_values.31.key: initializer.past_key_values.31.key + past_key_values.31.value: initializer.past_key_values.31.value + past_key_values.32.key: initializer.past_key_values.32.key + past_key_values.32.value: initializer.past_key_values.32.value + past_key_values.33.key: initializer.past_key_values.33.key + past_key_values.33.value: initializer.past_key_values.33.value + past_key_values.34.key: initializer.past_key_values.34.key + past_key_values.34.value: initializer.past_key_values.34.value + past_key_values.35.key: initializer.past_key_values.35.key + past_key_values.35.value: initializer.past_key_values.35.value + past_key_values.36.key: initializer.past_key_values.36.key + past_key_values.36.value: initializer.past_key_values.36.value + past_key_values.37.key: initializer.past_key_values.37.key + past_key_values.37.value: initializer.past_key_values.37.value + past_key_values.38.key: initializer.past_key_values.38.key + past_key_values.38.value: initializer.past_key_values.38.value + past_key_values.39.key: initializer.past_key_values.39.key + past_key_values.39.value: initializer.past_key_values.39.value + past_key_values.40.key: initializer.past_key_values.40.key + past_key_values.40.value: initializer.past_key_values.40.value + past_key_values.41.key: initializer.past_key_values.41.key + past_key_values.41.value: initializer.past_key_values.41.value + past_key_values.42.key: initializer.past_key_values.42.key + past_key_values.42.value: initializer.past_key_values.42.value + past_key_values.43.key: initializer.past_key_values.43.key + past_key_values.43.value: initializer.past_key_values.43.value + past_key_values.44.key: initializer.past_key_values.44.key + past_key_values.44.value: initializer.past_key_values.44.value + past_key_values.45.key: initializer.past_key_values.45.key + past_key_values.45.value: initializer.past_key_values.45.value + past_key_values.46.key: initializer.past_key_values.46.key + past_key_values.46.value: initializer.past_key_values.46.value + past_key_values.47.key: initializer.past_key_values.47.key + past_key_values.47.value: initializer.past_key_values.47.value + attention_mask: initializer.attention_mask + # SSA output names become the only values subsequent steps may consume. + outputs: + logits: decoder.setup.logits + present.0.key: decoder.setup.present.0.key + present.0.value: decoder.setup.present.0.value + present.1.key: decoder.setup.present.1.key + present.1.value: decoder.setup.present.1.value + present.2.key: decoder.setup.present.2.key + present.2.value: decoder.setup.present.2.value + present.3.key: decoder.setup.present.3.key + present.3.value: decoder.setup.present.3.value + present.4.key: decoder.setup.present.4.key + present.4.value: decoder.setup.present.4.value + present.5.key: decoder.setup.present.5.key + present.5.value: decoder.setup.present.5.value + present.6.key: decoder.setup.present.6.key + present.6.value: decoder.setup.present.6.value + present.7.key: decoder.setup.present.7.key + present.7.value: decoder.setup.present.7.value + present.8.key: decoder.setup.present.8.key + present.8.value: decoder.setup.present.8.value + present.9.key: decoder.setup.present.9.key + present.9.value: decoder.setup.present.9.value + present.10.key: decoder.setup.present.10.key + present.10.value: decoder.setup.present.10.value + present.11.key: decoder.setup.present.11.key + present.11.value: decoder.setup.present.11.value + present.12.key: decoder.setup.present.12.key + present.12.value: decoder.setup.present.12.value + present.13.key: decoder.setup.present.13.key + present.13.value: decoder.setup.present.13.value + present.14.key: decoder.setup.present.14.key + present.14.value: decoder.setup.present.14.value + present.15.key: decoder.setup.present.15.key + present.15.value: decoder.setup.present.15.value + present.16.key: decoder.setup.present.16.key + present.16.value: decoder.setup.present.16.value + present.17.key: decoder.setup.present.17.key + present.17.value: decoder.setup.present.17.value + present.18.key: decoder.setup.present.18.key + present.18.value: decoder.setup.present.18.value + present.19.key: decoder.setup.present.19.key + present.19.value: decoder.setup.present.19.value + present.20.key: decoder.setup.present.20.key + present.20.value: decoder.setup.present.20.value + present.21.key: decoder.setup.present.21.key + present.21.value: decoder.setup.present.21.value + present.22.key: decoder.setup.present.22.key + present.22.value: decoder.setup.present.22.value + present.23.key: decoder.setup.present.23.key + present.23.value: decoder.setup.present.23.value + present.24.key: decoder.setup.present.24.key + present.24.value: decoder.setup.present.24.value + present.25.key: decoder.setup.present.25.key + present.25.value: decoder.setup.present.25.value + present.26.key: decoder.setup.present.26.key + present.26.value: decoder.setup.present.26.value + present.27.key: decoder.setup.present.27.key + present.27.value: decoder.setup.present.27.value + present.28.key: decoder.setup.present.28.key + present.28.value: decoder.setup.present.28.value + present.29.key: decoder.setup.present.29.key + present.29.value: decoder.setup.present.29.value + present.30.key: decoder.setup.present.30.key + present.30.value: decoder.setup.present.30.value + present.31.key: decoder.setup.present.31.key + present.31.value: decoder.setup.present.31.value + present.32.key: decoder.setup.present.32.key + present.32.value: decoder.setup.present.32.value + present.33.key: decoder.setup.present.33.key + present.33.value: decoder.setup.present.33.value + present.34.key: decoder.setup.present.34.key + present.34.value: decoder.setup.present.34.value + present.35.key: decoder.setup.present.35.key + present.35.value: decoder.setup.present.35.value + present.36.key: decoder.setup.present.36.key + present.36.value: decoder.setup.present.36.value + present.37.key: decoder.setup.present.37.key + present.37.value: decoder.setup.present.37.value + present.38.key: decoder.setup.present.38.key + present.38.value: decoder.setup.present.38.value + present.39.key: decoder.setup.present.39.key + present.39.value: decoder.setup.present.39.value + present.40.key: decoder.setup.present.40.key + present.40.value: decoder.setup.present.40.value + present.41.key: decoder.setup.present.41.key + present.41.value: decoder.setup.present.41.value + present.42.key: decoder.setup.present.42.key + present.42.value: decoder.setup.present.42.value + present.43.key: decoder.setup.present.43.key + present.43.value: decoder.setup.present.43.value + present.44.key: decoder.setup.present.44.key + present.44.value: decoder.setup.present.44.value + present.45.key: decoder.setup.present.45.key + present.45.value: decoder.setup.present.45.value + present.46.key: decoder.setup.present.46.key + present.46.value: decoder.setup.present.46.value + present.47.key: decoder.setup.present.47.key + present.47.value: decoder.setup.present.47.value + # Invoke component `termination_batch_initializer` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: termination_batch_initializer + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + input_eos_ids: request.eos_ids + input_eos_lengths: request.eos_lengths + input_max_iterations: request.row_max_iterations + fallback_max_iterations: request.max_iterations + active: package.active + # SSA output names become the only values subsequent steps may consume. + outputs: + row_eos_ids: termination.eos_ids + eos_lengths: termination.eos_lengths + max_iterations: termination.max_iterations + # Invoke component `last_token_logits` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: last_token_logits + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + logits: decoder.setup.logits + # SSA output names become the only values subsequent steps may consume. + outputs: + last_logits: decoder.setup.last_logits + # Nested steps remain part of the typed executable workflow. + steps: + # Invoke component `token_sampler` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: token_sampler + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + logits: logits + temperature: request.temperature + top_k: request.top_k + top_p: request.top_p + min_p: request.min_p + seed: request.seed + counter: rng_counter + active: active + done: done + # SSA output names become the only values subsequent steps may consume. + outputs: + token: sample.body + next_counter: sample.next_counter + # Invoke component `token_to_slot` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: token_to_slot + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + token: sample.body + # SSA output names become the only values subsequent steps may consume. + outputs: + slot: sample.slot + # Invoke component `generated_length_update` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: generated_length_update + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + left: generated_lengths + right: package.one_token + active: active + done: done + # SSA output names become the only values subsequent steps may consume. + outputs: + total: token.next_lengths + # Invoke component `generated_length_update` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: generated_length_update + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + left: package.zero_batch + right: package.one_token + active: active + done: done + # SSA output names become the only values subsequent steps may consume. + outputs: + total: token.emitted_length + # Invoke component `token_state_update` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: token_state_update + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + current: token + update: sample.slot + active: active + done: done + # SSA output names become the only values subsequent steps may consume. + outputs: + next: token.body + # Invoke component `termination` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: termination + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + tokens: sample.body + eos_ids: termination.eos_ids + eos_lengths: termination.eos_lengths + iteration: loop.iteration + max_iterations: termination.max_iterations + active: active + # SSA output names become the only values subsequent steps may consume. + outputs: + done: loop.done + continue: loop.continue + next_active: loop.next_active + # Invoke component `cache_length_update` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: cache_length_update + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + left: cache_lengths + right: package.one_token + active: active + done: done + # SSA output names become the only values subsequent steps may consume. + outputs: + total: cache_lengths.next + # Invoke component `cache_length_update` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: cache_length_update + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + left: package.zero_batch + right: package.one_token + active: active + done: done + # SSA output names become the only values subsequent steps may consume. + outputs: + total: accepted_len.next + # Publish `token.body` as output `tokens` using append semantics. + - kind: emit + value: token.body + output: tokens + # Emit/update mode is semantic replacement/append behavior, not a buffering optimization. + mode: append + # Valid length prevents padded/unused elements from becoming observable output. + valid_length: token.emitted_length + when: active + # Invoke component `decoder_step_update` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: decoder_step_update + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + attention_mask: attention_mask + logical_length: cache_lengths + # SSA output names become the only values subsequent steps may consume. + outputs: + next_attention_mask: decoder_step.body_attention_mask + # Invoke component `model` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: model + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + input_ids: token.body + past_key_values.0.key: cache_0 + past_key_values.0.value: cache_1 + past_key_values.1.key: cache_2 + past_key_values.1.value: cache_3 + past_key_values.2.key: cache_4 + past_key_values.2.value: cache_5 + past_key_values.3.key: cache_6 + past_key_values.3.value: cache_7 + past_key_values.4.key: cache_8 + past_key_values.4.value: cache_9 + past_key_values.5.key: cache_10 + past_key_values.5.value: cache_11 + past_key_values.6.key: cache_12 + past_key_values.6.value: cache_13 + past_key_values.7.key: cache_14 + past_key_values.7.value: cache_15 + past_key_values.8.key: cache_16 + past_key_values.8.value: cache_17 + past_key_values.9.key: cache_18 + past_key_values.9.value: cache_19 + past_key_values.10.key: cache_20 + past_key_values.10.value: cache_21 + past_key_values.11.key: cache_22 + past_key_values.11.value: cache_23 + past_key_values.12.key: cache_24 + past_key_values.12.value: cache_25 + past_key_values.13.key: cache_26 + past_key_values.13.value: cache_27 + past_key_values.14.key: cache_28 + past_key_values.14.value: cache_29 + past_key_values.15.key: cache_30 + past_key_values.15.value: cache_31 + past_key_values.16.key: cache_32 + past_key_values.16.value: cache_33 + past_key_values.17.key: cache_34 + past_key_values.17.value: cache_35 + past_key_values.18.key: cache_36 + past_key_values.18.value: cache_37 + past_key_values.19.key: cache_38 + past_key_values.19.value: cache_39 + past_key_values.20.key: cache_40 + past_key_values.20.value: cache_41 + past_key_values.21.key: cache_42 + past_key_values.21.value: cache_43 + past_key_values.22.key: cache_44 + past_key_values.22.value: cache_45 + past_key_values.23.key: cache_46 + past_key_values.23.value: cache_47 + past_key_values.24.key: cache_48 + past_key_values.24.value: cache_49 + past_key_values.25.key: cache_50 + past_key_values.25.value: cache_51 + past_key_values.26.key: cache_52 + past_key_values.26.value: cache_53 + past_key_values.27.key: cache_54 + past_key_values.27.value: cache_55 + past_key_values.28.key: cache_56 + past_key_values.28.value: cache_57 + past_key_values.29.key: cache_58 + past_key_values.29.value: cache_59 + past_key_values.30.key: cache_60 + past_key_values.30.value: cache_61 + past_key_values.31.key: cache_62 + past_key_values.31.value: cache_63 + past_key_values.32.key: cache_64 + past_key_values.32.value: cache_65 + past_key_values.33.key: cache_66 + past_key_values.33.value: cache_67 + past_key_values.34.key: cache_68 + past_key_values.34.value: cache_69 + past_key_values.35.key: cache_70 + past_key_values.35.value: cache_71 + past_key_values.36.key: cache_72 + past_key_values.36.value: cache_73 + past_key_values.37.key: cache_74 + past_key_values.37.value: cache_75 + past_key_values.38.key: cache_76 + past_key_values.38.value: cache_77 + past_key_values.39.key: cache_78 + past_key_values.39.value: cache_79 + past_key_values.40.key: cache_80 + past_key_values.40.value: cache_81 + past_key_values.41.key: cache_82 + past_key_values.41.value: cache_83 + past_key_values.42.key: cache_84 + past_key_values.42.value: cache_85 + past_key_values.43.key: cache_86 + past_key_values.43.value: cache_87 + past_key_values.44.key: cache_88 + past_key_values.44.value: cache_89 + past_key_values.45.key: cache_90 + past_key_values.45.value: cache_91 + past_key_values.46.key: cache_92 + past_key_values.46.value: cache_93 + past_key_values.47.key: cache_94 + past_key_values.47.value: cache_95 + attention_mask: decoder_step.body_attention_mask + # SSA output names become the only values subsequent steps may consume. + outputs: + logits: decoder.body.logits + present.0.key: decoder.body.present.0.key + present.0.value: decoder.body.present.0.value + present.1.key: decoder.body.present.1.key + present.1.value: decoder.body.present.1.value + present.2.key: decoder.body.present.2.key + present.2.value: decoder.body.present.2.value + present.3.key: decoder.body.present.3.key + present.3.value: decoder.body.present.3.value + present.4.key: decoder.body.present.4.key + present.4.value: decoder.body.present.4.value + present.5.key: decoder.body.present.5.key + present.5.value: decoder.body.present.5.value + present.6.key: decoder.body.present.6.key + present.6.value: decoder.body.present.6.value + present.7.key: decoder.body.present.7.key + present.7.value: decoder.body.present.7.value + present.8.key: decoder.body.present.8.key + present.8.value: decoder.body.present.8.value + present.9.key: decoder.body.present.9.key + present.9.value: decoder.body.present.9.value + present.10.key: decoder.body.present.10.key + present.10.value: decoder.body.present.10.value + present.11.key: decoder.body.present.11.key + present.11.value: decoder.body.present.11.value + present.12.key: decoder.body.present.12.key + present.12.value: decoder.body.present.12.value + present.13.key: decoder.body.present.13.key + present.13.value: decoder.body.present.13.value + present.14.key: decoder.body.present.14.key + present.14.value: decoder.body.present.14.value + present.15.key: decoder.body.present.15.key + present.15.value: decoder.body.present.15.value + present.16.key: decoder.body.present.16.key + present.16.value: decoder.body.present.16.value + present.17.key: decoder.body.present.17.key + present.17.value: decoder.body.present.17.value + present.18.key: decoder.body.present.18.key + present.18.value: decoder.body.present.18.value + present.19.key: decoder.body.present.19.key + present.19.value: decoder.body.present.19.value + present.20.key: decoder.body.present.20.key + present.20.value: decoder.body.present.20.value + present.21.key: decoder.body.present.21.key + present.21.value: decoder.body.present.21.value + present.22.key: decoder.body.present.22.key + present.22.value: decoder.body.present.22.value + present.23.key: decoder.body.present.23.key + present.23.value: decoder.body.present.23.value + present.24.key: decoder.body.present.24.key + present.24.value: decoder.body.present.24.value + present.25.key: decoder.body.present.25.key + present.25.value: decoder.body.present.25.value + present.26.key: decoder.body.present.26.key + present.26.value: decoder.body.present.26.value + present.27.key: decoder.body.present.27.key + present.27.value: decoder.body.present.27.value + present.28.key: decoder.body.present.28.key + present.28.value: decoder.body.present.28.value + present.29.key: decoder.body.present.29.key + present.29.value: decoder.body.present.29.value + present.30.key: decoder.body.present.30.key + present.30.value: decoder.body.present.30.value + present.31.key: decoder.body.present.31.key + present.31.value: decoder.body.present.31.value + present.32.key: decoder.body.present.32.key + present.32.value: decoder.body.present.32.value + present.33.key: decoder.body.present.33.key + present.33.value: decoder.body.present.33.value + present.34.key: decoder.body.present.34.key + present.34.value: decoder.body.present.34.value + present.35.key: decoder.body.present.35.key + present.35.value: decoder.body.present.35.value + present.36.key: decoder.body.present.36.key + present.36.value: decoder.body.present.36.value + present.37.key: decoder.body.present.37.key + present.37.value: decoder.body.present.37.value + present.38.key: decoder.body.present.38.key + present.38.value: decoder.body.present.38.value + present.39.key: decoder.body.present.39.key + present.39.value: decoder.body.present.39.value + present.40.key: decoder.body.present.40.key + present.40.value: decoder.body.present.40.value + present.41.key: decoder.body.present.41.key + present.41.value: decoder.body.present.41.value + present.42.key: decoder.body.present.42.key + present.42.value: decoder.body.present.42.value + present.43.key: decoder.body.present.43.key + present.43.value: decoder.body.present.43.value + present.44.key: decoder.body.present.44.key + present.44.value: decoder.body.present.44.value + present.45.key: decoder.body.present.45.key + present.45.value: decoder.body.present.45.value + present.46.key: decoder.body.present.46.key + present.46.value: decoder.body.present.46.value + present.47.key: decoder.body.present.47.key + present.47.value: decoder.body.present.47.value + # Invoke component `last_token_logits` with explicit SSA input/output bindings; this is executable tensor computation, not a host-side capability flag. + - kind: invoke + component: last_token_logits + # SSA bindings supply exact values to this step; no implicit tensor discovery occurs. + inputs: + logits: decoder.body.logits + # SSA output names become the only values subsequent steps may consume. + outputs: + last_logits: decoder.body.last_logits + continue_when: active + max_iterations: request.max_iterations + carried: + - cell: token + next: token.body + - cell: logits + next: decoder.body.last_logits + - cell: generated_lengths + next: token.next_lengths + - cell: active + next: loop.next_active + - cell: done + next: loop.done + - cell: cache_lengths + next: cache_lengths.next + - cell: accepted_len + next: accepted_len.next + - cell: rng_counter + next: sample.next_counter + - cell: attention_mask + next: decoder_step.body_attention_mask + - cell: cache_0 + next: decoder.body.present.0.key + - cell: cache_1 + next: decoder.body.present.0.value + - cell: cache_2 + next: decoder.body.present.1.key + - cell: cache_3 + next: decoder.body.present.1.value + - cell: cache_4 + next: decoder.body.present.2.key + - cell: cache_5 + next: decoder.body.present.2.value + - cell: cache_6 + next: decoder.body.present.3.key + - cell: cache_7 + next: decoder.body.present.3.value + - cell: cache_8 + next: decoder.body.present.4.key + - cell: cache_9 + next: decoder.body.present.4.value + - cell: cache_10 + next: decoder.body.present.5.key + - cell: cache_11 + next: decoder.body.present.5.value + - cell: cache_12 + next: decoder.body.present.6.key + - cell: cache_13 + next: decoder.body.present.6.value + - cell: cache_14 + next: decoder.body.present.7.key + - cell: cache_15 + next: decoder.body.present.7.value + - cell: cache_16 + next: decoder.body.present.8.key + - cell: cache_17 + next: decoder.body.present.8.value + - cell: cache_18 + next: decoder.body.present.9.key + - cell: cache_19 + next: decoder.body.present.9.value + - cell: cache_20 + next: decoder.body.present.10.key + - cell: cache_21 + next: decoder.body.present.10.value + - cell: cache_22 + next: decoder.body.present.11.key + - cell: cache_23 + next: decoder.body.present.11.value + - cell: cache_24 + next: decoder.body.present.12.key + - cell: cache_25 + next: decoder.body.present.12.value + - cell: cache_26 + next: decoder.body.present.13.key + - cell: cache_27 + next: decoder.body.present.13.value + - cell: cache_28 + next: decoder.body.present.14.key + - cell: cache_29 + next: decoder.body.present.14.value + - cell: cache_30 + next: decoder.body.present.15.key + - cell: cache_31 + next: decoder.body.present.15.value + - cell: cache_32 + next: decoder.body.present.16.key + - cell: cache_33 + next: decoder.body.present.16.value + - cell: cache_34 + next: decoder.body.present.17.key + - cell: cache_35 + next: decoder.body.present.17.value + - cell: cache_36 + next: decoder.body.present.18.key + - cell: cache_37 + next: decoder.body.present.18.value + - cell: cache_38 + next: decoder.body.present.19.key + - cell: cache_39 + next: decoder.body.present.19.value + - cell: cache_40 + next: decoder.body.present.20.key + - cell: cache_41 + next: decoder.body.present.20.value + - cell: cache_42 + next: decoder.body.present.21.key + - cell: cache_43 + next: decoder.body.present.21.value + - cell: cache_44 + next: decoder.body.present.22.key + - cell: cache_45 + next: decoder.body.present.22.value + - cell: cache_46 + next: decoder.body.present.23.key + - cell: cache_47 + next: decoder.body.present.23.value + - cell: cache_48 + next: decoder.body.present.24.key + - cell: cache_49 + next: decoder.body.present.24.value + - cell: cache_50 + next: decoder.body.present.25.key + - cell: cache_51 + next: decoder.body.present.25.value + - cell: cache_52 + next: decoder.body.present.26.key + - cell: cache_53 + next: decoder.body.present.26.value + - cell: cache_54 + next: decoder.body.present.27.key + - cell: cache_55 + next: decoder.body.present.27.value + - cell: cache_56 + next: decoder.body.present.28.key + - cell: cache_57 + next: decoder.body.present.28.value + - cell: cache_58 + next: decoder.body.present.29.key + - cell: cache_59 + next: decoder.body.present.29.value + - cell: cache_60 + next: decoder.body.present.30.key + - cell: cache_61 + next: decoder.body.present.30.value + - cell: cache_62 + next: decoder.body.present.31.key + - cell: cache_63 + next: decoder.body.present.31.value + - cell: cache_64 + next: decoder.body.present.32.key + - cell: cache_65 + next: decoder.body.present.32.value + - cell: cache_66 + next: decoder.body.present.33.key + - cell: cache_67 + next: decoder.body.present.33.value + - cell: cache_68 + next: decoder.body.present.34.key + - cell: cache_69 + next: decoder.body.present.34.value + - cell: cache_70 + next: decoder.body.present.35.key + - cell: cache_71 + next: decoder.body.present.35.value + - cell: cache_72 + next: decoder.body.present.36.key + - cell: cache_73 + next: decoder.body.present.36.value + - cell: cache_74 + next: decoder.body.present.37.key + - cell: cache_75 + next: decoder.body.present.37.value + - cell: cache_76 + next: decoder.body.present.38.key + - cell: cache_77 + next: decoder.body.present.38.value + - cell: cache_78 + next: decoder.body.present.39.key + - cell: cache_79 + next: decoder.body.present.39.value + - cell: cache_80 + next: decoder.body.present.40.key + - cell: cache_81 + next: decoder.body.present.40.value + - cell: cache_82 + next: decoder.body.present.41.key + - cell: cache_83 + next: decoder.body.present.41.value + - cell: cache_84 + next: decoder.body.present.42.key + - cell: cache_85 + next: decoder.body.present.42.value + - cell: cache_86 + next: decoder.body.present.43.key + - cell: cache_87 + next: decoder.body.present.43.value + - cell: cache_88 + next: decoder.body.present.44.key + - cell: cache_89 + next: decoder.body.present.44.value + - cell: cache_90 + next: decoder.body.present.45.key + - cell: cache_91 + next: decoder.body.present.45.value + - cell: cache_92 + next: decoder.body.present.46.key + - cell: cache_93 + next: decoder.body.present.46.value + - cell: cache_94 + next: decoder.body.present.47.key + - cell: cache_95 + next: decoder.body.present.47.value + termination: generation_eos + iteration: + value: loop.iteration + contract: + dtype: int64 + rank: 1 + shape: + - 1