Restructure: 6 variants (f16+bf16 x default+cuda+onnx-standard)

bf16/cuda/audio_encoder/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b91d094783cb3847141ea5f13529c2348c226a0024a4ed4433754611552a3b7
+size 1239093

bf16/cuda/audio_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbbb99ee9ece21e21b59d81c3d2a60b04d05dc4511951a2a8ab5aa48af4e3c0d
+size 614725024

bf16/cuda/audio_feature_extraction.json

@@ -0,0 +1,30 @@
+{
+    "feature_extraction": {
+        "sequence": [
+            {
+                "operation": {
+                    "name": "audio_decoder",
+                    "type": "AudioDecoder"
+                }
+            },
+            {
+                "operation": {
+                    "name": "gemma4_log_mel",
+                    "type": "Gemma4LogMel",
+                    "attrs": {
+                        "feature_size": 128,
+                        "sampling_rate": 16000,
+                        "frame_length_ms": 20.0,
+                        "hop_length_ms": 10.0,
+                        "min_frequency": 0.0,
+                        "max_frequency": 8000.0,
+                        "preemphasis": 0.0,
+                        "preemphasis_htk_flavor": 1,
+                        "fft_overdrive": 0,
+                        "mel_floor": 0.001
+                    }
+                }
+            }
+        ]
+    }
+}

bf16/cuda/chat_template.jinja

@@ -0,0 +1,351 @@
+{%- macro format_parameters(properties, required, filter_keys=false) -%}
+    {%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}
+    {%- set ns = namespace(found_first=false) -%}
+    {%- for key, value in properties | dictsort -%}
+        {%- set add_comma = false -%}
+        {%- if not filter_keys or key not in standard_keys -%}
+            {%- if ns.found_first %},{% endif -%}
+            {%- set ns.found_first = true -%}
+            {{ key }}:{
+            {%- if value['description'] -%}
+                description:<|"|>{{ value['description'] }}<|"|>
+                {%- set add_comma = true -%}
+            {%- endif -%}
+            {%- if value['type'] | upper == 'STRING' -%}
+                {%- if value['enum'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    enum:{{ format_argument(value['enum']) }}
+                {%- endif -%}
+            {%- elif value['type'] | upper == 'ARRAY' -%}
+                {%- if value['items'] is mapping and value['items'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    items:{
+                    {%- set ns_items = namespace(found_first=false) -%}
+                    {%- for item_key, item_value in value['items'] | dictsort -%}
+                        {%- if item_value is not none -%}
+                            {%- if ns_items.found_first %},{% endif -%}
+                            {%- set ns_items.found_first = true -%}
+                            {%- if item_key == 'properties' -%}
+                                properties:{
+                                {%- if item_value is mapping -%}
+                                    {{- format_parameters(item_value, value['items']['required'] | default([])) -}}
+                                {%- endif -%}
+                                }
+                            {%- elif item_key == 'required' -%}
+                                required:[
+                                {%- for req_item in item_value -%}
+                                    <|"|>{{- req_item -}}<|"|>
+                                    {%- if not loop.last %},{% endif -%}
+                                {%- endfor -%}
+                                ]
+                            {%- elif item_key == 'type' -%}
+                                {%- if item_value is string -%}
+                                    type:{{ format_argument(item_value | upper) }}
+                                {%- else -%}
+                                    type:{{ format_argument(item_value | map('upper') | list) }}
+                                {%- endif -%}
+                            {%- else -%}
+                                {{ item_key }}:{{ format_argument(item_value) }}
+                            {%- endif -%}
+                        {%- endif -%}
+                    {%- endfor -%}
+                    }
+                {%- endif -%}
+            {%- endif -%}
+            {%- if value['nullable'] %}
+                {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                nullable:true
+            {%- endif -%}
+            {%- if value['type'] | upper == 'OBJECT' -%}
+                {%- if value['properties'] is defined and value['properties'] is mapping -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    properties:{
+                    {{- format_parameters(value['properties'], value['required'] | default([])) -}}
+                    }
+                {%- elif value is mapping -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    properties:{
+                    {{- format_parameters(value, value['required'] | default([]), filter_keys=true) -}}
+                    }
+                {%- endif -%}
+                {%- if value['required'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    required:[
+                    {%- for item in value['required'] | default([]) -%}
+                        <|"|>{{- item -}}<|"|>
+                        {%- if not loop.last %},{% endif -%}
+                    {%- endfor -%}
+                    ]
+                {%- endif -%}
+            {%- endif -%}
+            {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+            type:<|"|>{{ value['type'] | upper }}<|"|>}
+        {%- endif -%}
+    {%- endfor -%}
+{%- endmacro -%}
+{%- macro format_function_declaration(tool_data) -%}
+    declaration:{{- tool_data['function']['name'] -}}{description:<|"|>{{- tool_data['function']['description'] -}}<|"|>
+    {%- set params = tool_data['function']['parameters'] -%}
+    {%- if params -%}
+        ,parameters:{
+        {%- if params['properties'] -%}
+            properties:{ {{- format_parameters(params['properties'], params['required']) -}} },
+        {%- endif -%}
+        {%- if params['required'] -%}
+            required:[
+            {%- for item in params['required'] -%}
+                <|"|>{{- item -}}<|"|>
+                {{- ',' if not loop.last -}}
+            {%- endfor -%}
+            ],
+        {%- endif -%}
+        {%- if params['type'] -%}
+            type:<|"|>{{- params['type'] | upper -}}<|"|>}
+        {%- endif -%}
+    {%- endif -%}
+    {%- if 'response' in tool_data['function'] -%}
+        {%- set response_declaration = tool_data['function']['response'] -%}
+        ,response:{
+        {%- if response_declaration['description'] -%}
+            description:<|"|>{{- response_declaration['description'] -}}<|"|>,
+        {%- endif -%}
+        {%- if response_declaration['type'] | upper == 'OBJECT' -%}
+            type:<|"|>{{- response_declaration['type'] | upper -}}<|"|>}
+        {%- endif -%}
+    {%- endif -%}
+    }
+{%- endmacro -%}
+{%- macro format_argument(argument, escape_keys=True) -%}
+    {%- if argument is string -%}
+        {{- '<|"|>' + argument + '<|"|>' -}}
+    {%- elif argument is boolean -%}
+        {{- 'true' if argument else 'false' -}}
+    {%- elif argument is mapping -%}
+        {{- '{' -}}
+        {%- set ns = namespace(found_first=false) -%}
+        {%- for key, value in argument | dictsort -%}
+            {%- if ns.found_first %},{% endif -%}
+            {%- set ns.found_first = true -%}
+            {%- if escape_keys -%}
+                {{- '<|"|>' + key + '<|"|>' -}}
+            {%- else -%}
+                {{- key -}}
+            {%- endif -%}
+            :{{- format_argument(value, escape_keys=escape_keys) -}}
+        {%- endfor -%}
+        {{- '}' -}}
+    {%- elif argument is sequence -%}
+        {{- '[' -}}
+        {%- for item in argument -%}
+            {{- format_argument(item, escape_keys=escape_keys) -}}
+            {%- if not loop.last %},{% endif -%}
+        {%- endfor -%}
+        {{- ']' -}}
+    {%- else -%}
+        {{- argument -}}
+    {%- endif -%}
+{%- endmacro -%}
+{%- macro strip_thinking(text) -%}
+    {%- set ns = namespace(result='') -%}
+    {%- for part in text.split('<channel|>') -%}
+        {%- if '<|channel>' in part -%}
+            {%- set ns.result = ns.result + part.split('<|channel>')[0] -%}
+        {%- else -%}
+            {%- set ns.result = ns.result + part -%}
+        {%- endif -%}
+    {%- endfor -%}
+    {{- ns.result | trim -}}
+{%- endmacro -%}
+
+{%- macro format_tool_response_block(tool_name, response) -%}
+    {{- '<|tool_response>' -}}
+    {%- if response is mapping -%}
+        {{- 'response:' + tool_name + '{' -}}
+        {%- for key, value in response | dictsort -%}
+            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
+            {%- if not loop.last %},{% endif -%}
+        {%- endfor -%}
+        {{- '}' -}}
+    {%- else -%}
+        {{- 'response:' + tool_name + '{value:' + format_argument(response, escape_keys=False) + '}' -}}
+    {%- endif -%}
+    {{- '<tool_response|>' -}}
+{%- endmacro -%}
+
+{%- set ns = namespace(prev_message_type=None) -%}
+{%- set loop_messages = messages -%}
+{{- bos_token -}}
+{#- Handle System/Tool Definitions Block -#}
+{%- if (enable_thinking is defined and enable_thinking) or tools or messages[0]['role'] in ['system', 'developer'] -%}
+    {{- '<|turn>system\n' -}}
+    {#- Inject Thinking token at the very top of the FIRST system turn -#}
+    {%- if enable_thinking is defined and enable_thinking -%}
+        {{- '<|think|>\n' -}}
+        {%- set ns.prev_message_type = 'think' -%}
+    {%- endif -%}
+    {%- if messages[0]['role'] in ['system', 'developer'] -%}
+        {%- if messages[0]['content'] is string -%}
+            {{- messages[0]['content'] | trim -}}
+        {%- elif messages[0]['content'] is sequence -%}
+            {%- for item in messages[0]['content'] -%}
+                {{- item['text'] | trim + ' '-}}
+            {%- endfor -%}
+        {%- endif -%}
+        {%- set loop_messages = messages[1:] -%}
+    {%- endif -%}
+    {%- if tools -%}
+        {%- for tool in tools %}
+            {{- '<|tool>' -}}
+            {{- format_function_declaration(tool) | trim -}}
+            {{- '<tool|>' -}}
+        {%- endfor %}
+        {%- set ns.prev_message_type = 'tool' -%}
+    {%- endif -%}
+    {{- '<turn|>\n' -}}
+{%- endif %}
+
+{#- Pre-scan: find last user message index for reasoning guard -#}
+{%- set ns_turn = namespace(last_user_idx=-1) -%}
+{%- for i in range(loop_messages | length) -%}
+    {%- if loop_messages[i]['role'] == 'user' -%}
+        {%- set ns_turn.last_user_idx = i -%}
+    {%- endif -%}
+{%- endfor -%}
+
+{#- Loop through messages -#}
+{%- for message in loop_messages -%}
+    {%- if message['role'] != 'tool' -%}
+    {%- set ns.prev_message_type = None -%}
+    {%- set role = 'model' if message['role'] == 'assistant' else message['role'] -%}
+    {#- Detect continuation: suppress duplicate <|turn>model when previous non-tool message was also assistant -#}
+    {%- set prev_nt = namespace(role=None, found=false) -%}
+    {%- if loop.index0 > 0 -%}
+        {%- for j in range(loop.index0 - 1, -1, -1) -%}
+            {%- if not prev_nt.found -%}
+                {%- if loop_messages[j]['role'] != 'tool' -%}
+                    {%- set prev_nt.role = loop_messages[j]['role'] -%}
+                    {%- set prev_nt.found = true -%}
+                {%- endif -%}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- endif -%}
+    {%- set continue_same_model_turn = (role == 'model' and prev_nt.role == 'assistant') -%}
+    {%- if not continue_same_model_turn -%}
+        {{- '<|turn>' + role + '\n' }}
+    {%- endif -%}
+
+    {#- Render reasoning/reasoning_content as thinking channel -#}
+    {%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}
+    {%- if thinking_text and loop.index0 > ns_turn.last_user_idx and message.get('tool_calls') -%}
+        {{- '<|channel>thought\n' + thinking_text + '\n<channel|>' -}}
+    {%- endif -%}
+
+            {%- if message['tool_calls'] -%}
+                {%- for tool_call in message['tool_calls'] -%}
+                    {%- set function = tool_call['function'] -%}
+                    {{- '<|tool_call>call:' + function['name'] + '{' -}}
+                    {%- if function['arguments'] is mapping -%}
+                        {%- set ns_args = namespace(found_first=false) -%}
+                        {%- for key, value in function['arguments'] | dictsort -%}
+                            {%- if ns_args.found_first %},{% endif -%}
+                            {%- set ns_args.found_first = true -%}
+                            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
+                        {%- endfor -%}
+                    {%- elif function['arguments'] is string -%}
+                        {{- function['arguments'] -}}
+                    {%- endif -%}
+                    {{- '}<tool_call|>' -}}
+                {%- endfor -%}
+                {%- set ns.prev_message_type = 'tool_call' -%}
+            {%- endif -%}
+
+            {%- set ns_tr_out = namespace(flag=false) -%}
+            {%- if message.get('tool_responses') -%}
+                {#- Legacy: tool_responses embedded on the assistant message (Google/Gemma native) -#}
+                {%- for tool_response in message['tool_responses'] -%}
+                    {{- format_tool_response_block(tool_response['name'] | default('unknown'), tool_response['response']) -}}
+                    {%- set ns_tr_out.flag = true -%}
+                    {%- set ns.prev_message_type = 'tool_response' -%}
+                {%- endfor -%}
+            {%- elif message.get('tool_calls') -%}
+                {#- OpenAI Chat Completions: forward-scan consecutive role:tool messages -#}
+                {%- set ns_tool_scan = namespace(stopped=false) -%}
+                {%- for k in range(loop.index0 + 1, loop_messages | length) -%}
+                    {%- if ns_tool_scan.stopped -%}
+                    {%- elif loop_messages[k]['role'] != 'tool' -%}
+                        {%- set ns_tool_scan.stopped = true -%}
+                    {%- else -%}
+                        {%- set follow = loop_messages[k] -%}
+                        {#- Resolve tool_call_id to function name -#}
+                        {%- set ns_tname = namespace(name=follow.get('name') | default('unknown')) -%}
+                        {%- for tc in message['tool_calls'] -%}
+                            {%- if tc.get('id') == follow.get('tool_call_id') -%}
+                                {%- set ns_tname.name = tc['function']['name'] -%}
+                            {%- endif -%}
+                        {%- endfor -%}
+                        {#- Handle content as string or content-parts array -#}
+                        {%- set tool_body = follow.get('content') -%}
+                        {%- if tool_body is string -%}
+                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}
+                        {%- elif tool_body is sequence and tool_body is not string -%}
+                            {%- set ns_txt = namespace(s='') -%}
+                            {%- for part in tool_body -%}
+                                {%- if part.get('type') == 'text' -%}
+                                    {%- set ns_txt.s = ns_txt.s + (part.get('text') | default('')) -%}
+                                {%- endif -%}
+                            {%- endfor -%}
+                            {{- format_tool_response_block(ns_tname.name, ns_txt.s) -}}
+                        {%- else -%}
+                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}
+                        {%- endif -%}
+                        {%- set ns_tr_out.flag = true -%}
+                        {%- set ns.prev_message_type = 'tool_response' -%}
+                    {%- endif -%}
+                {%- endfor -%}
+            {%- endif -%}
+
+            {%- set captured_content -%}
+            {%- if message['content'] is string -%}
+                {%- if role == 'model' -%}
+                    {{- strip_thinking(message['content']) -}}
+                {%- else -%}
+                    {{- message['content'] | trim -}}
+                {%- endif -%}
+            {%- elif message['content'] is sequence -%}
+                {%- for item in message['content'] -%}
+                    {%- if item['type'] == 'text' -%}
+                        {%- if role == 'model' -%}
+                            {{- strip_thinking(item['text']) -}}
+                        {%- else -%}
+                            {{- item['text'] | trim -}}
+                        {%- endif -%}
+                    {%- elif item['type'] == 'image' -%}
+                        {{- '<|image|>' -}}
+                        {%- set ns.prev_message_type = 'image' -%}
+                    {%- elif item['type'] == 'audio' -%}
+                        {{- '<|audio|>' -}}
+                        {%- set ns.prev_message_type = 'audio' -%}
+                    {%- elif item['type'] == 'video' -%}
+                        {{- '<|video|>' -}}
+                        {%- set ns.prev_message_type = 'video' -%}
+                    {%- endif -%}
+                {%- endfor -%}
+            {%- endif -%}
+            {%- endset -%}
+
+            {{- captured_content -}}
+            {%- set has_content = captured_content | trim | length > 0 -%}
+
+        {%- if ns.prev_message_type == 'tool_call' and not ns_tr_out.flag -%}
+            {{- '<|tool_response>' -}}
+        {%- elif not (ns_tr_out.flag and not has_content) -%}
+            {{- '<turn|>\n' -}}
+        {%- endif -%}
+    {%- endif -%}
+{%- endfor -%}
+
+{%- if add_generation_prompt -%}
+    {%- if ns.prev_message_type != 'tool_response' and ns.prev_message_type != 'tool_call' -%}
+        {{- '<|turn>model\n' -}}
+    {%- endif -%}
+{%- endif -%}

bf16/cuda/decoder/model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e554c5c22e59f60d8d5eee7505bac60e74da7a8d4be57666b3e9a4c2e2fba046
+size 4999430520

bf16/cuda/decoder/model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e3a718a5ed0fbfe379dbd95d01ccb26a59522db7d6490645a83e6c20a63dd6a
+size 4459103336

bf16/cuda/decoder/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4137153e85ae7578d1b0539b99ffcdaaa672c8192899f0b2f8bedf53d66cc855
+size 912580

bf16/cuda/decoder/model.safetensors.index.json

@@ -0,0 +1,550 @@
+{
+  "metadata": {
+    "total_size": 9458465280
+  },
+  "weight_map": {
+    "decoder.model.per_layer_projection_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.0.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.1.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.2.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.3.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.4.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.5.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.6.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.7.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.8.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.9.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.10.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.11.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.12.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.13.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.14.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.15.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.16.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.17.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.18.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.19.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.20.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.21.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.22.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.23.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.24.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.25.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.26.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.27.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.28.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.29.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.30.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.31.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.32.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.33.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.34.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_local.cos_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_local.sin_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_global.cos_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_global.sin_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.per_layer_model_projection.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.q_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.k_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.v_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.o_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.mlp.gate_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.mlp.up_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.mlp.down_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.per_layer_input_gate.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.per_layer_projection.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.lm_head.weight_t": "model-00002-of-00002.safetensors"
+  }
+}

bf16/cuda/embedding/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47d105fc7ae637554a6552b90e4335caa175427b8770d1891e3602cb869776e0
+size 8656

bf16/cuda/embedding/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:741c3297844e13c22a8eae3db118a94ae7d296e490d25e12e1a90c3d143f8681
+size 805312824

bf16/cuda/genai_config.json

@@ -0,0 +1,128 @@
+{
+    "model": {
+        "type": "gemma4",
+        "vocab_size": 262144,
+        "context_length": 131072,
+        "decoder": {
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": [
+                    {
+                        "CUDAExecutionProvider": {
+                            "enable_cuda_graph": "0",
+                            "enable_skip_layer_norm_strict_mode": "1"
+                        }
+                    }
+                ]
+            },
+            "filename": "decoder/model.onnx",
+            "head_size": 256,
+            "hidden_size": 1536,
+            "inputs": {
+                "inputs_embeds": "inputs_embeds",
+                "attention_mask": "attention_mask",
+                "position_ids": "position_ids",
+                "input_ids": "input_ids",
+                "past_key_names": "past_key_values.%d.key",
+                "past_value_names": "past_key_values.%d.value"
+            },
+            "outputs": {
+                "logits": "logits",
+                "present_key_names": "present.%d.key",
+                "present_value_names": "present.%d.value"
+            },
+            "num_attention_heads": 8,
+            "num_hidden_layers": 35,
+            "num_key_value_heads": 1
+        },
+        "bos_token_id": 2,
+        "eos_token_id": [
+            1,
+            106
+        ],
+        "pad_token_id": 0,
+        "vision": {
+            "filename": "vision_encoder/model.onnx",
+            "config_filename": "image_processor.json",
+            "inputs": {
+                "pixel_values": "pixel_values",
+                "pixel_position_ids": "pixel_position_ids"
+            },
+            "outputs": {
+                "image_features": "image_features"
+            },
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": [
+                    {
+                        "CUDAExecutionProvider": {
+                            "enable_cuda_graph": "0",
+                            "enable_skip_layer_norm_strict_mode": "1"
+                        }
+                    }
+                ]
+            },
+            "spatial_merge_size": 2
+        },
+        "embedding": {
+            "filename": "embedding/model.onnx",
+            "inputs": {
+                "input_ids": "input_ids",
+                "image_features": "image_features",
+                "audio_features": "audio_features"
+            },
+            "outputs": {
+                "inputs_embeds": "inputs_embeds"
+            },
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": [
+                    {
+                        "CUDAExecutionProvider": {
+                            "enable_cuda_graph": "0",
+                            "enable_skip_layer_norm_strict_mode": "1"
+                        }
+                    }
+                ]
+            }
+        },
+        "speech": {
+            "filename": "audio_encoder/model.onnx",
+            "config_filename": "audio_feature_extraction.json",
+            "inputs": {
+                "audio_embeds": "input_features",
+                "attention_mask": "input_features_mask"
+            },
+            "outputs": {
+                "audio_features": "audio_features"
+            },
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": [
+                    {
+                        "CUDAExecutionProvider": {
+                            "enable_cuda_graph": "0",
+                            "enable_skip_layer_norm_strict_mode": "1"
+                        }
+                    }
+                ]
+            }
+        },
+        "image_token_id": 258880,
+        "audio_token_id": 258881,
+        "boa_token_id": 256000
+    },
+    "search": {
+        "do_sample": true,
+        "early_stopping": true,
+        "max_length": 131072,
+        "min_length": 0,
+        "num_beams": 1,
+        "num_return_sequences": 1,
+        "past_present_share_buffer": false,
+        "repetition_penalty": 1.0,
+        "temperature": 1.0,
+        "top_k": 1,
+        "top_p": 1.0
+    }
+}

bf16/cuda/image_processor.json

@@ -0,0 +1,27 @@
+{
+    "processor": {
+        "name": "gemma_4_image_processing",
+        "transforms": [
+            {
+                "operation": {
+                    "name": "decode_image",
+                    "type": "DecodeImage",
+                    "attrs": {
+                        "color_space": "RGB"
+                    }
+                }
+            },
+            {
+                "operation": {
+                    "name": "gemma4_image_transform",
+                    "type": "Gemma4ImageTransform",
+                    "attrs": {
+                        "patch_size": 16,
+                        "max_soft_tokens": 280,
+                        "pooling_kernel_size": 3
+                    }
+                }
+            }
+        ]
+    }
+}

bf16/cuda/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc8d3a0ce36466ccc1278bf987df5f71db1719b9ca6b4118264f45cb627bfe0f
+size 32169626

bf16/cuda/tokenizer_config.json

@@ -0,0 +1,75 @@
+{
+  "audio_token": "<|audio|>",
+  "backend": "tokenizers",
+  "boa_token": "<|audio>",
+  "boi_token": "<|image>",
+  "bos_token": "<bos>",
+  "eoa_token": "<audio|>",
+  "eoc_token": "<channel|>",
+  "eoi_token": "<image|>",
+  "eos_token": "<eos>",
+  "eot_token": "<turn|>",
+  "escape_token": "<|\"|>",
+  "etc_token": "<tool_call|>",
+  "etd_token": "<tool|>",
+  "etr_token": "<tool_response|>",
+  "extra_special_tokens": [
+    "<|video|>"
+  ],
+  "image_token": "<|image|>",
+  "mask_token": "<mask>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "padding_side": "left",
+  "processor_class": "Gemma4Processor",
+  "response_schema": {
+    "type": "object",
+    "properties": {
+      "role": {
+        "const": "assistant"
+      },
+      "thinking": {
+        "type": "string"
+      },
+      "content": {
+        "type": "string"
+      },
+      "tool_calls": {
+        "x-regex-iterator": "<\\|tool_call>(.*?)<tool_call\\|>",
+        "type": "array",
+        "items": {
+          "type": "object",
+          "properties": {
+            "type": {
+              "const": "function"
+            },
+            "function": {
+              "type": "object",
+              "x-regex": "call\\:(?P<name>\\w+)(?P<arguments>\\{.*\\})",
+              "properties": {
+                "name": {
+                  "type": "string"
+                },
+                "arguments": {
+                  "type": "object",
+                  "x-parser": "gemma4-tool-call",
+                  "additionalProperties": {}
+                }
+              }
+            }
+          }
+        }
+      }
+    },
+    "x-regex": "(\\<\\|channel\\>thought\\n(?P<thinking>.*?)\\<channel\\|\\>)?(?P<tool_calls>\\<\\|tool_call\\>.*\\<tool_call\\|\\>)?(?P<content>(?:(?!\\<turn\\|\\>)(?!\\<\\|tool_response\\>).)+)?(?:\\<turn\\|\\>|\\<\\|tool_response\\>)?"
+  },
+  "soc_token": "<|channel>",
+  "sot_token": "<|turn>",
+  "stc_token": "<|tool_call>",
+  "std_token": "<|tool>",
+  "str_token": "<|tool_response>",
+  "think_token": "<|think|>",
+  "tokenizer_class": "GemmaTokenizer",
+  "unk_token": "<unk>",
+  "chat_template": "{%- macro format_parameters(properties, required, filter_keys=false) -%}\n    {%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}\n    {%- set ns = namespace(found_first=false) -%}\n    {%- for key, value in properties | dictsort -%}\n        {%- set add_comma = false -%}\n        {%- if not filter_keys or key not in standard_keys -%}\n            {%- if ns.found_first %},{% endif -%}\n            {%- set ns.found_first = true -%}\n            {{ key }}:{\n            {%- if value['description'] -%}\n                description:<|\"|>{{ value['description'] }}<|\"|>\n                {%- set add_comma = true -%}\n            {%- endif -%}\n            {%- if value['type'] | upper == 'STRING' -%}\n                {%- if value['enum'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    enum:{{ format_argument(value['enum']) }}\n                {%- endif -%}\n            {%- elif value['type'] | upper == 'ARRAY' -%}\n                {%- if value['items'] is mapping and value['items'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    items:{\n                    {%- set ns_items = namespace(found_first=false) -%}\n                    {%- for item_key, item_value in value['items'] | dictsort -%}\n                        {%- if item_value is not none -%}\n                            {%- if ns_items.found_first %},{% endif -%}\n                            {%- set ns_items.found_first = true -%}\n                            {%- if item_key == 'properties' -%}\n                                properties:{\n                                {%- if item_value is mapping -%}\n                                    {{- format_parameters(item_value, value['items']['required'] | default([])) -}}\n                                {%- endif -%}\n                                }\n                            {%- elif item_key == 'required' -%}\n                                required:[\n                                {%- for req_item in item_value -%}\n                                    <|\"|>{{- req_item -}}<|\"|>\n                                    {%- if not loop.last %},{% endif -%}\n                                {%- endfor -%}\n                                ]\n                            {%- elif item_key == 'type' -%}\n                                {%- if item_value is string -%}\n                                    type:{{ format_argument(item_value | upper) }}\n                                {%- else -%}\n                                    type:{{ format_argument(item_value | map('upper') | list) }}\n                                {%- endif -%}\n                            {%- else -%}\n                                {{ item_key }}:{{ format_argument(item_value) }}\n                            {%- endif -%}\n                        {%- endif -%}\n                    {%- endfor -%}\n                    }\n                {%- endif -%}\n            {%- endif -%}\n            {%- if value['nullable'] %}\n                {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                nullable:true\n            {%- endif -%}\n            {%- if value['type'] | upper == 'OBJECT' -%}\n                {%- if value['properties'] is defined and value['properties'] is mapping -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    properties:{\n                    {{- format_parameters(value['properties'], value['required'] | default([])) -}}\n                    }\n                {%- elif value is mapping -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    properties:{\n                    {{- format_parameters(value, value['required'] | default([]), filter_keys=true) -}}\n                    }\n                {%- endif -%}\n                {%- if value['required'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    required:[\n                    {%- for item in value['required'] | default([]) -%}\n                        <|\"|>{{- item -}}<|\"|>\n                        {%- if not loop.last %},{% endif -%}\n                    {%- endfor -%}\n                    ]\n                {%- endif -%}\n            {%- endif -%}\n            {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n            type:<|\"|>{{ value['type'] | upper }}<|\"|>}\n        {%- endif -%}\n    {%- endfor -%}\n{%- endmacro -%}\n{%- macro format_function_declaration(tool_data) -%}\n    declaration:{{- tool_data['function']['name'] -}}{description:<|\"|>{{- tool_data['function']['description'] -}}<|\"|>\n    {%- set params = tool_data['function']['parameters'] -%}\n    {%- if params -%}\n        ,parameters:{\n        {%- if params['properties'] -%}\n            properties:{ {{- format_parameters(params['properties'], params['required']) -}} },\n        {%- endif -%}\n        {%- if params['required'] -%}\n            required:[\n            {%- for item in params['required'] -%}\n                <|\"|>{{- item -}}<|\"|>\n                {{- ',' if not loop.last -}}\n            {%- endfor -%}\n            ],\n        {%- endif -%}\n        {%- if params['type'] -%}\n            type:<|\"|>{{- params['type'] | upper -}}<|\"|>}\n        {%- endif -%}\n    {%- endif -%}\n    {%- if 'response' in tool_data['function'] -%}\n        {%- set response_declaration = tool_data['function']['response'] -%}\n        ,response:{\n        {%- if response_declaration['description'] -%}\n            description:<|\"|>{{- response_declaration['description'] -}}<|\"|>,\n        {%- endif -%}\n        {%- if response_declaration['type'] | upper == 'OBJECT' -%}\n            type:<|\"|>{{- response_declaration['type'] | upper -}}<|\"|>}\n        {%- endif -%}\n    {%- endif -%}\n    }\n{%- endmacro -%}\n{%- macro format_argument(argument, escape_keys=True) -%}\n    {%- if argument is string -%}\n        {{- '<|\"|>' + argument + '<|\"|>' -}}\n    {%- elif argument is boolean -%}\n        {{- 'true' if argument else 'false' -}}\n    {%- elif argument is mapping -%}\n        {{- '{' -}}\n        {%- set ns = namespace(found_first=false) -%}\n        {%- for key, value in argument | dictsort -%}\n            {%- if ns.found_first %},{% endif -%}\n            {%- set ns.found_first = true -%}\n            {%- if escape_keys -%}\n                {{- '<|\"|>' + key + '<|\"|>' -}}\n            {%- else -%}\n                {{- key -}}\n            {%- endif -%}\n            :{{- format_argument(value, escape_keys=escape_keys) -}}\n        {%- endfor -%}\n        {{- '}' -}}\n    {%- elif argument is sequence -%}\n        {{- '[' -}}\n        {%- for item in argument -%}\n            {{- format_argument(item, escape_keys=escape_keys) -}}\n            {%- if not loop.last %},{% endif -%}\n        {%- endfor -%}\n        {{- ']' -}}\n    {%- else -%}\n        {{- argument -}}\n    {%- endif -%}\n{%- endmacro -%}\n{%- macro strip_thinking(text) -%}\n    {%- set ns = namespace(result='') -%}\n    {%- for part in text.split('<channel|>') -%}\n        {%- if '<|channel>' in part -%}\n            {%- set ns.result = ns.result + part.split('<|channel>')[0] -%}\n        {%- else -%}\n            {%- set ns.result = ns.result + part -%}\n        {%- endif -%}\n    {%- endfor -%}\n    {{- ns.result | trim -}}\n{%- endmacro -%}\n\n{%- macro format_tool_response_block(tool_name, response) -%}\n    {{- '<|tool_response>' -}}\n    {%- if response is mapping -%}\n        {{- 'response:' + tool_name + '{' -}}\n        {%- for key, value in response | dictsort -%}\n            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}\n            {%- if not loop.last %},{% endif -%}\n        {%- endfor -%}\n        {{- '}' -}}\n    {%- else -%}\n        {{- 'response:' + tool_name + '{value:' + format_argument(response, escape_keys=False) + '}' -}}\n    {%- endif -%}\n    {{- '<tool_response|>' -}}\n{%- endmacro -%}\n\n{%- set ns = namespace(prev_message_type=None) -%}\n{%- set loop_messages = messages -%}\n{{- bos_token -}}\n{#- Handle System/Tool Definitions Block -#}\n{%- if (enable_thinking is defined and enable_thinking) or tools or messages[0]['role'] in ['system', 'developer'] -%}\n    {{- '<|turn>system\\n' -}}\n    {#- Inject Thinking token at the very top of the FIRST system turn -#}\n    {%- if enable_thinking is defined and enable_thinking -%}\n        {{- '<|think|>\\n' -}}\n        {%- set ns.prev_message_type = 'think' -%}\n    {%- endif -%}\n    {%- if messages[0]['role'] in ['system', 'developer'] -%}\n        {%- if messages[0]['content'] is string -%}\n            {{- messages[0]['content'] | trim -}}\n        {%- elif messages[0]['content'] is sequence -%}\n            {%- for item in messages[0]['content'] -%}\n                {{- item['text'] | trim + ' '-}}\n            {%- endfor -%}\n        {%- endif -%}\n        {%- set loop_messages = messages[1:] -%}\n    {%- endif -%}\n    {%- if tools -%}\n        {%- for tool in tools %}\n            {{- '<|tool>' -}}\n            {{- format_function_declaration(tool) | trim -}}\n            {{- '<tool|>' -}}\n        {%- endfor %}\n        {%- set ns.prev_message_type = 'tool' -%}\n    {%- endif -%}\n    {{- '<turn|>\\n' -}}\n{%- endif %}\n\n{#- Pre-scan: find last user message index for reasoning guard -#}\n{%- set ns_turn = namespace(last_user_idx=-1) -%}\n{%- for i in range(loop_messages | length) -%}\n    {%- if loop_messages[i]['role'] == 'user' -%}\n        {%- set ns_turn.last_user_idx = i -%}\n    {%- endif -%}\n{%- endfor -%}\n\n{#- Loop through messages -#}\n{%- for message in loop_messages -%}\n    {%- if message['role'] != 'tool' -%}\n    {%- set ns.prev_message_type = None -%}\n    {%- set role = 'model' if message['role'] == 'assistant' else message['role'] -%}\n    {#- Detect continuation: suppress duplicate <|turn>model when previous non-tool message was also assistant -#}\n    {%- set prev_nt = namespace(role=None, found=false) -%}\n    {%- if loop.index0 > 0 -%}\n        {%- for j in range(loop.index0 - 1, -1, -1) -%}\n            {%- if not prev_nt.found -%}\n                {%- if loop_messages[j]['role'] != 'tool' -%}\n                    {%- set prev_nt.role = loop_messages[j]['role'] -%}\n                    {%- set prev_nt.found = true -%}\n                {%- endif -%}\n            {%- endif -%}\n        {%- endfor -%}\n    {%- endif -%}\n    {%- set continue_same_model_turn = (role == 'model' and prev_nt.role == 'assistant') -%}\n    {%- if not continue_same_model_turn -%}\n        {{- '<|turn>' + role + '\\n' }}\n    {%- endif -%}\n\n    {#- Render reasoning/reasoning_content as thinking channel -#}\n    {%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}\n    {%- if thinking_text and loop.index0 > ns_turn.last_user_idx and message.get('tool_calls') -%}\n        {{- '<|channel>thought\\n' + thinking_text + '\\n<channel|>' -}}\n    {%- endif -%}\n\n            {%- if message['tool_calls'] -%}\n                {%- for tool_call in message['tool_calls'] -%}\n                    {%- set function = tool_call['function'] -%}\n                    {{- '<|tool_call>call:' + function['name'] + '{' -}}\n                    {%- if function['arguments'] is mapping -%}\n                        {%- set ns_args = namespace(found_first=false) -%}\n                        {%- for key, value in function['arguments'] | dictsort -%}\n                            {%- if ns_args.found_first %},{% endif -%}\n                            {%- set ns_args.found_first = true -%}\n                            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}\n                        {%- endfor -%}\n                    {%- elif function['arguments'] is string -%}\n                        {{- function['arguments'] -}}\n                    {%- endif -%}\n                    {{- '}<tool_call|>' -}}\n                {%- endfor -%}\n                {%- set ns.prev_message_type = 'tool_call' -%}\n            {%- endif -%}\n\n            {%- set ns_tr_out = namespace(flag=false) -%}\n            {%- if message.get('tool_responses') -%}\n                {#- Legacy: tool_responses embedded on the assistant message (Google/Gemma native) -#}\n                {%- for tool_response in message['tool_responses'] -%}\n                    {{- format_tool_response_block(tool_response['name'] | default('unknown'), tool_response['response']) -}}\n                    {%- set ns_tr_out.flag = true -%}\n                    {%- set ns.prev_message_type = 'tool_response' -%}\n                {%- endfor -%}\n            {%- elif message.get('tool_calls') -%}\n                {#- OpenAI Chat Completions: forward-scan consecutive role:tool messages -#}\n                {%- set ns_tool_scan = namespace(stopped=false) -%}\n                {%- for k in range(loop.index0 + 1, loop_messages | length) -%}\n                    {%- if ns_tool_scan.stopped -%}\n                    {%- elif loop_messages[k]['role'] != 'tool' -%}\n                        {%- set ns_tool_scan.stopped = true -%}\n                    {%- else -%}\n                        {%- set follow = loop_messages[k] -%}\n                        {#- Resolve tool_call_id to function name -#}\n                        {%- set ns_tname = namespace(name=follow.get('name') | default('unknown')) -%}\n                        {%- for tc in message['tool_calls'] -%}\n                            {%- if tc.get('id') == follow.get('tool_call_id') -%}\n                                {%- set ns_tname.name = tc['function']['name'] -%}\n                            {%- endif -%}\n                        {%- endfor -%}\n                        {#- Handle content as string or content-parts array -#}\n                        {%- set tool_body = follow.get('content') -%}\n                        {%- if tool_body is string -%}\n                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}\n                        {%- elif tool_body is sequence and tool_body is not string -%}\n                            {%- set ns_txt = namespace(s='') -%}\n                            {%- for part in tool_body -%}\n                                {%- if part.get('type') == 'text' -%}\n                                    {%- set ns_txt.s = ns_txt.s + (part.get('text') | default('')) -%}\n                                {%- endif -%}\n                            {%- endfor -%}\n                            {{- format_tool_response_block(ns_tname.name, ns_txt.s) -}}\n                        {%- else -%}\n                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}\n                        {%- endif -%}\n                        {%- set ns_tr_out.flag = true -%}\n                        {%- set ns.prev_message_type = 'tool_response' -%}\n                    {%- endif -%}\n                {%- endfor -%}\n            {%- endif -%}\n\n            {%- set captured_content -%}\n            {%- if message['content'] is string -%}\n                {%- if role == 'model' -%}\n                    {{- strip_thinking(message['content']) -}}\n                {%- else -%}\n                    {{- message['content'] | trim -}}\n                {%- endif -%}\n            {%- elif message['content'] is sequence -%}\n                {%- for item in message['content'] -%}\n                    {%- if item['type'] == 'text' -%}\n                        {%- if role == 'model' -%}\n                            {{- strip_thinking(item['text']) -}}\n                        {%- else -%}\n                            {{- item['text'] | trim -}}\n                        {%- endif -%}\n                    {%- elif item['type'] == 'image' -%}\n                        {{- '<|image|>' -}}\n                        {%- set ns.prev_message_type = 'image' -%}\n                    {%- elif item['type'] == 'audio' -%}\n                        {{- '<|audio|>' -}}\n                        {%- set ns.prev_message_type = 'audio' -%}\n                    {%- elif item['type'] == 'video' -%}\n                        {{- '<|video|>' -}}\n                        {%- set ns.prev_message_type = 'video' -%}\n                    {%- endif -%}\n                {%- endfor -%}\n            {%- endif -%}\n            {%- endset -%}\n\n            {{- captured_content -}}\n            {%- set has_content = captured_content | trim | length > 0 -%}\n\n        {%- if ns.prev_message_type == 'tool_call' and not ns_tr_out.flag -%}\n            {{- '<|tool_response>' -}}\n        {%- elif not (ns_tr_out.flag and not has_content) -%}\n            {{- '<turn|>\\n' -}}\n        {%- endif -%}\n    {%- endif -%}\n{%- endfor -%}\n\n{%- if add_generation_prompt -%}\n    {%- if ns.prev_message_type != 'tool_response' and ns.prev_message_type != 'tool_call' -%}\n        {{- '<|turn>model\\n' -}}\n    {%- endif -%}\n{%- endif -%}"
+}

bf16/cuda/vision_encoder/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b1aa8d01379401cce914cd2bed70e776238d174ee6b08c016ae94c6b3617c48
+size 1239595

bf16/cuda/vision_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33fd860d1ce013444d2f16667e4625045971ad0ca905d1cb3b62feae2e937ba9
+size 358083464

bf16/default/audio_encoder/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b91d094783cb3847141ea5f13529c2348c226a0024a4ed4433754611552a3b7
+size 1239093

bf16/default/audio_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbbb99ee9ece21e21b59d81c3d2a60b04d05dc4511951a2a8ab5aa48af4e3c0d
+size 614725024

bf16/default/audio_feature_extraction.json

@@ -0,0 +1,30 @@
+{
+    "feature_extraction": {
+        "sequence": [
+            {
+                "operation": {
+                    "name": "audio_decoder",
+                    "type": "AudioDecoder"
+                }
+            },
+            {
+                "operation": {
+                    "name": "gemma4_log_mel",
+                    "type": "Gemma4LogMel",
+                    "attrs": {
+                        "feature_size": 128,
+                        "sampling_rate": 16000,
+                        "frame_length_ms": 20.0,
+                        "hop_length_ms": 10.0,
+                        "min_frequency": 0.0,
+                        "max_frequency": 8000.0,
+                        "preemphasis": 0.0,
+                        "preemphasis_htk_flavor": 1,
+                        "fft_overdrive": 0,
+                        "mel_floor": 0.001
+                    }
+                }
+            }
+        ]
+    }
+}

bf16/default/chat_template.jinja

@@ -0,0 +1,351 @@
+{%- macro format_parameters(properties, required, filter_keys=false) -%}
+    {%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}
+    {%- set ns = namespace(found_first=false) -%}
+    {%- for key, value in properties | dictsort -%}
+        {%- set add_comma = false -%}
+        {%- if not filter_keys or key not in standard_keys -%}
+            {%- if ns.found_first %},{% endif -%}
+            {%- set ns.found_first = true -%}
+            {{ key }}:{
+            {%- if value['description'] -%}
+                description:<|"|>{{ value['description'] }}<|"|>
+                {%- set add_comma = true -%}
+            {%- endif -%}
+            {%- if value['type'] | upper == 'STRING' -%}
+                {%- if value['enum'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    enum:{{ format_argument(value['enum']) }}
+                {%- endif -%}
+            {%- elif value['type'] | upper == 'ARRAY' -%}
+                {%- if value['items'] is mapping and value['items'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    items:{
+                    {%- set ns_items = namespace(found_first=false) -%}
+                    {%- for item_key, item_value in value['items'] | dictsort -%}
+                        {%- if item_value is not none -%}
+                            {%- if ns_items.found_first %},{% endif -%}
+                            {%- set ns_items.found_first = true -%}
+                            {%- if item_key == 'properties' -%}
+                                properties:{
+                                {%- if item_value is mapping -%}
+                                    {{- format_parameters(item_value, value['items']['required'] | default([])) -}}
+                                {%- endif -%}
+                                }
+                            {%- elif item_key == 'required' -%}
+                                required:[
+                                {%- for req_item in item_value -%}
+                                    <|"|>{{- req_item -}}<|"|>
+                                    {%- if not loop.last %},{% endif -%}
+                                {%- endfor -%}
+                                ]
+                            {%- elif item_key == 'type' -%}
+                                {%- if item_value is string -%}
+                                    type:{{ format_argument(item_value | upper) }}
+                                {%- else -%}
+                                    type:{{ format_argument(item_value | map('upper') | list) }}
+                                {%- endif -%}
+                            {%- else -%}
+                                {{ item_key }}:{{ format_argument(item_value) }}
+                            {%- endif -%}
+                        {%- endif -%}
+                    {%- endfor -%}
+                    }
+                {%- endif -%}
+            {%- endif -%}
+            {%- if value['nullable'] %}
+                {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                nullable:true
+            {%- endif -%}
+            {%- if value['type'] | upper == 'OBJECT' -%}
+                {%- if value['properties'] is defined and value['properties'] is mapping -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    properties:{
+                    {{- format_parameters(value['properties'], value['required'] | default([])) -}}
+                    }
+                {%- elif value is mapping -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    properties:{
+                    {{- format_parameters(value, value['required'] | default([]), filter_keys=true) -}}
+                    }
+                {%- endif -%}
+                {%- if value['required'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    required:[
+                    {%- for item in value['required'] | default([]) -%}
+                        <|"|>{{- item -}}<|"|>
+                        {%- if not loop.last %},{% endif -%}
+                    {%- endfor -%}
+                    ]
+                {%- endif -%}
+            {%- endif -%}
+            {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+            type:<|"|>{{ value['type'] | upper }}<|"|>}
+        {%- endif -%}
+    {%- endfor -%}
+{%- endmacro -%}
+{%- macro format_function_declaration(tool_data) -%}
+    declaration:{{- tool_data['function']['name'] -}}{description:<|"|>{{- tool_data['function']['description'] -}}<|"|>
+    {%- set params = tool_data['function']['parameters'] -%}
+    {%- if params -%}
+        ,parameters:{
+        {%- if params['properties'] -%}
+            properties:{ {{- format_parameters(params['properties'], params['required']) -}} },
+        {%- endif -%}
+        {%- if params['required'] -%}
+            required:[
+            {%- for item in params['required'] -%}
+                <|"|>{{- item -}}<|"|>
+                {{- ',' if not loop.last -}}
+            {%- endfor -%}
+            ],
+        {%- endif -%}
+        {%- if params['type'] -%}
+            type:<|"|>{{- params['type'] | upper -}}<|"|>}
+        {%- endif -%}
+    {%- endif -%}
+    {%- if 'response' in tool_data['function'] -%}
+        {%- set response_declaration = tool_data['function']['response'] -%}
+        ,response:{
+        {%- if response_declaration['description'] -%}
+            description:<|"|>{{- response_declaration['description'] -}}<|"|>,
+        {%- endif -%}
+        {%- if response_declaration['type'] | upper == 'OBJECT' -%}
+            type:<|"|>{{- response_declaration['type'] | upper -}}<|"|>}
+        {%- endif -%}
+    {%- endif -%}
+    }
+{%- endmacro -%}
+{%- macro format_argument(argument, escape_keys=True) -%}
+    {%- if argument is string -%}
+        {{- '<|"|>' + argument + '<|"|>' -}}
+    {%- elif argument is boolean -%}
+        {{- 'true' if argument else 'false' -}}
+    {%- elif argument is mapping -%}
+        {{- '{' -}}
+        {%- set ns = namespace(found_first=false) -%}
+        {%- for key, value in argument | dictsort -%}
+            {%- if ns.found_first %},{% endif -%}
+            {%- set ns.found_first = true -%}
+            {%- if escape_keys -%}
+                {{- '<|"|>' + key + '<|"|>' -}}
+            {%- else -%}
+                {{- key -}}
+            {%- endif -%}
+            :{{- format_argument(value, escape_keys=escape_keys) -}}
+        {%- endfor -%}
+        {{- '}' -}}
+    {%- elif argument is sequence -%}
+        {{- '[' -}}
+        {%- for item in argument -%}
+            {{- format_argument(item, escape_keys=escape_keys) -}}
+            {%- if not loop.last %},{% endif -%}
+        {%- endfor -%}
+        {{- ']' -}}
+    {%- else -%}
+        {{- argument -}}
+    {%- endif -%}
+{%- endmacro -%}
+{%- macro strip_thinking(text) -%}
+    {%- set ns = namespace(result='') -%}
+    {%- for part in text.split('<channel|>') -%}
+        {%- if '<|channel>' in part -%}
+            {%- set ns.result = ns.result + part.split('<|channel>')[0] -%}
+        {%- else -%}
+            {%- set ns.result = ns.result + part -%}
+        {%- endif -%}
+    {%- endfor -%}
+    {{- ns.result | trim -}}
+{%- endmacro -%}
+
+{%- macro format_tool_response_block(tool_name, response) -%}
+    {{- '<|tool_response>' -}}
+    {%- if response is mapping -%}
+        {{- 'response:' + tool_name + '{' -}}
+        {%- for key, value in response | dictsort -%}
+            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
+            {%- if not loop.last %},{% endif -%}
+        {%- endfor -%}
+        {{- '}' -}}
+    {%- else -%}
+        {{- 'response:' + tool_name + '{value:' + format_argument(response, escape_keys=False) + '}' -}}
+    {%- endif -%}
+    {{- '<tool_response|>' -}}
+{%- endmacro -%}
+
+{%- set ns = namespace(prev_message_type=None) -%}
+{%- set loop_messages = messages -%}
+{{- bos_token -}}
+{#- Handle System/Tool Definitions Block -#}
+{%- if (enable_thinking is defined and enable_thinking) or tools or messages[0]['role'] in ['system', 'developer'] -%}
+    {{- '<|turn>system\n' -}}
+    {#- Inject Thinking token at the very top of the FIRST system turn -#}
+    {%- if enable_thinking is defined and enable_thinking -%}
+        {{- '<|think|>\n' -}}
+        {%- set ns.prev_message_type = 'think' -%}
+    {%- endif -%}
+    {%- if messages[0]['role'] in ['system', 'developer'] -%}
+        {%- if messages[0]['content'] is string -%}
+            {{- messages[0]['content'] | trim -}}
+        {%- elif messages[0]['content'] is sequence -%}
+            {%- for item in messages[0]['content'] -%}
+                {{- item['text'] | trim + ' '-}}
+            {%- endfor -%}
+        {%- endif -%}
+        {%- set loop_messages = messages[1:] -%}
+    {%- endif -%}
+    {%- if tools -%}
+        {%- for tool in tools %}
+            {{- '<|tool>' -}}
+            {{- format_function_declaration(tool) | trim -}}
+            {{- '<tool|>' -}}
+        {%- endfor %}
+        {%- set ns.prev_message_type = 'tool' -%}
+    {%- endif -%}
+    {{- '<turn|>\n' -}}
+{%- endif %}
+
+{#- Pre-scan: find last user message index for reasoning guard -#}
+{%- set ns_turn = namespace(last_user_idx=-1) -%}
+{%- for i in range(loop_messages | length) -%}
+    {%- if loop_messages[i]['role'] == 'user' -%}
+        {%- set ns_turn.last_user_idx = i -%}
+    {%- endif -%}
+{%- endfor -%}
+
+{#- Loop through messages -#}
+{%- for message in loop_messages -%}
+    {%- if message['role'] != 'tool' -%}
+    {%- set ns.prev_message_type = None -%}
+    {%- set role = 'model' if message['role'] == 'assistant' else message['role'] -%}
+    {#- Detect continuation: suppress duplicate <|turn>model when previous non-tool message was also assistant -#}
+    {%- set prev_nt = namespace(role=None, found=false) -%}
+    {%- if loop.index0 > 0 -%}
+        {%- for j in range(loop.index0 - 1, -1, -1) -%}
+            {%- if not prev_nt.found -%}
+                {%- if loop_messages[j]['role'] != 'tool' -%}
+                    {%- set prev_nt.role = loop_messages[j]['role'] -%}
+                    {%- set prev_nt.found = true -%}
+                {%- endif -%}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- endif -%}
+    {%- set continue_same_model_turn = (role == 'model' and prev_nt.role == 'assistant') -%}
+    {%- if not continue_same_model_turn -%}
+        {{- '<|turn>' + role + '\n' }}
+    {%- endif -%}
+
+    {#- Render reasoning/reasoning_content as thinking channel -#}
+    {%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}
+    {%- if thinking_text and loop.index0 > ns_turn.last_user_idx and message.get('tool_calls') -%}
+        {{- '<|channel>thought\n' + thinking_text + '\n<channel|>' -}}
+    {%- endif -%}
+
+            {%- if message['tool_calls'] -%}
+                {%- for tool_call in message['tool_calls'] -%}
+                    {%- set function = tool_call['function'] -%}
+                    {{- '<|tool_call>call:' + function['name'] + '{' -}}
+                    {%- if function['arguments'] is mapping -%}
+                        {%- set ns_args = namespace(found_first=false) -%}
+                        {%- for key, value in function['arguments'] | dictsort -%}
+                            {%- if ns_args.found_first %},{% endif -%}
+                            {%- set ns_args.found_first = true -%}
+                            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
+                        {%- endfor -%}
+                    {%- elif function['arguments'] is string -%}
+                        {{- function['arguments'] -}}
+                    {%- endif -%}
+                    {{- '}<tool_call|>' -}}
+                {%- endfor -%}
+                {%- set ns.prev_message_type = 'tool_call' -%}
+            {%- endif -%}
+
+            {%- set ns_tr_out = namespace(flag=false) -%}
+            {%- if message.get('tool_responses') -%}
+                {#- Legacy: tool_responses embedded on the assistant message (Google/Gemma native) -#}
+                {%- for tool_response in message['tool_responses'] -%}
+                    {{- format_tool_response_block(tool_response['name'] | default('unknown'), tool_response['response']) -}}
+                    {%- set ns_tr_out.flag = true -%}
+                    {%- set ns.prev_message_type = 'tool_response' -%}
+                {%- endfor -%}
+            {%- elif message.get('tool_calls') -%}
+                {#- OpenAI Chat Completions: forward-scan consecutive role:tool messages -#}
+                {%- set ns_tool_scan = namespace(stopped=false) -%}
+                {%- for k in range(loop.index0 + 1, loop_messages | length) -%}
+                    {%- if ns_tool_scan.stopped -%}
+                    {%- elif loop_messages[k]['role'] != 'tool' -%}
+                        {%- set ns_tool_scan.stopped = true -%}
+                    {%- else -%}
+                        {%- set follow = loop_messages[k] -%}
+                        {#- Resolve tool_call_id to function name -#}
+                        {%- set ns_tname = namespace(name=follow.get('name') | default('unknown')) -%}
+                        {%- for tc in message['tool_calls'] -%}
+                            {%- if tc.get('id') == follow.get('tool_call_id') -%}
+                                {%- set ns_tname.name = tc['function']['name'] -%}
+                            {%- endif -%}
+                        {%- endfor -%}
+                        {#- Handle content as string or content-parts array -#}
+                        {%- set tool_body = follow.get('content') -%}
+                        {%- if tool_body is string -%}
+                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}
+                        {%- elif tool_body is sequence and tool_body is not string -%}
+                            {%- set ns_txt = namespace(s='') -%}
+                            {%- for part in tool_body -%}
+                                {%- if part.get('type') == 'text' -%}
+                                    {%- set ns_txt.s = ns_txt.s + (part.get('text') | default('')) -%}
+                                {%- endif -%}
+                            {%- endfor -%}
+                            {{- format_tool_response_block(ns_tname.name, ns_txt.s) -}}
+                        {%- else -%}
+                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}
+                        {%- endif -%}
+                        {%- set ns_tr_out.flag = true -%}
+                        {%- set ns.prev_message_type = 'tool_response' -%}
+                    {%- endif -%}
+                {%- endfor -%}
+            {%- endif -%}
+
+            {%- set captured_content -%}
+            {%- if message['content'] is string -%}
+                {%- if role == 'model' -%}
+                    {{- strip_thinking(message['content']) -}}
+                {%- else -%}
+                    {{- message['content'] | trim -}}
+                {%- endif -%}
+            {%- elif message['content'] is sequence -%}
+                {%- for item in message['content'] -%}
+                    {%- if item['type'] == 'text' -%}
+                        {%- if role == 'model' -%}
+                            {{- strip_thinking(item['text']) -}}
+                        {%- else -%}
+                            {{- item['text'] | trim -}}
+                        {%- endif -%}
+                    {%- elif item['type'] == 'image' -%}
+                        {{- '<|image|>' -}}
+                        {%- set ns.prev_message_type = 'image' -%}
+                    {%- elif item['type'] == 'audio' -%}
+                        {{- '<|audio|>' -}}
+                        {%- set ns.prev_message_type = 'audio' -%}
+                    {%- elif item['type'] == 'video' -%}
+                        {{- '<|video|>' -}}
+                        {%- set ns.prev_message_type = 'video' -%}
+                    {%- endif -%}
+                {%- endfor -%}
+            {%- endif -%}
+            {%- endset -%}
+
+            {{- captured_content -}}
+            {%- set has_content = captured_content | trim | length > 0 -%}
+
+        {%- if ns.prev_message_type == 'tool_call' and not ns_tr_out.flag -%}
+            {{- '<|tool_response>' -}}
+        {%- elif not (ns_tr_out.flag and not has_content) -%}
+            {{- '<turn|>\n' -}}
+        {%- endif -%}
+    {%- endif -%}
+{%- endfor -%}
+
+{%- if add_generation_prompt -%}
+    {%- if ns.prev_message_type != 'tool_response' and ns.prev_message_type != 'tool_call' -%}
+        {{- '<|turn>model\n' -}}
+    {%- endif -%}
+{%- endif -%}

bf16/default/decoder/model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e554c5c22e59f60d8d5eee7505bac60e74da7a8d4be57666b3e9a4c2e2fba046
+size 4999430520

bf16/default/decoder/model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e3a718a5ed0fbfe379dbd95d01ccb26a59522db7d6490645a83e6c20a63dd6a
+size 4459103336

bf16/default/decoder/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4599f57e737314b94d50e6d324c9bbb52c3bde7050e5288ea9483cff90c29591
+size 915092

bf16/default/decoder/model.safetensors.index.json

@@ -0,0 +1,550 @@
+{
+  "metadata": {
+    "total_size": 9458465280
+  },
+  "weight_map": {
+    "decoder.model.per_layer_projection_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.0.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.1.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.2.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.3.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.4.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.5.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.6.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.7.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.8.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.9.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.10.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.11.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.12.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.13.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.14.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.15.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.16.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.17.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.18.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.19.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.20.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.21.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.22.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.23.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.24.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.25.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.26.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.27.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.28.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.29.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.30.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.31.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.32.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.33.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.34.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_local.cos_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_local.sin_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_global.cos_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_global.sin_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.per_layer_model_projection.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.q_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.k_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.v_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.o_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.mlp.gate_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.mlp.up_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.mlp.down_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.per_layer_input_gate.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.per_layer_projection.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.lm_head.weight_t": "model-00002-of-00002.safetensors"
+  }
+}

bf16/default/embedding/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47d105fc7ae637554a6552b90e4335caa175427b8770d1891e3602cb869776e0
+size 8656

bf16/default/embedding/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:741c3297844e13c22a8eae3db118a94ae7d296e490d25e12e1a90c3d143f8681
+size 805312824

bf16/default/genai_config.json

@@ -0,0 +1,100 @@
+{
+    "model": {
+        "type": "gemma4",
+        "vocab_size": 262144,
+        "context_length": 131072,
+        "decoder": {
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": []
+            },
+            "filename": "decoder/model.onnx",
+            "head_size": 256,
+            "hidden_size": 1536,
+            "inputs": {
+                "inputs_embeds": "inputs_embeds",
+                "attention_mask": "attention_mask",
+                "position_ids": "position_ids",
+                "input_ids": "input_ids",
+                "past_key_names": "past_key_values.%d.key",
+                "past_value_names": "past_key_values.%d.value"
+            },
+            "outputs": {
+                "logits": "logits",
+                "present_key_names": "present.%d.key",
+                "present_value_names": "present.%d.value"
+            },
+            "num_attention_heads": 8,
+            "num_hidden_layers": 35,
+            "num_key_value_heads": 1
+        },
+        "bos_token_id": 2,
+        "eos_token_id": [
+            1,
+            106
+        ],
+        "pad_token_id": 0,
+        "vision": {
+            "filename": "vision_encoder/model.onnx",
+            "config_filename": "image_processor.json",
+            "inputs": {
+                "pixel_values": "pixel_values",
+                "pixel_position_ids": "pixel_position_ids"
+            },
+            "outputs": {
+                "image_features": "image_features"
+            },
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": []
+            },
+            "spatial_merge_size": 2
+        },
+        "embedding": {
+            "filename": "embedding/model.onnx",
+            "inputs": {
+                "input_ids": "input_ids",
+                "image_features": "image_features",
+                "audio_features": "audio_features"
+            },
+            "outputs": {
+                "inputs_embeds": "inputs_embeds"
+            },
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": []
+            }
+        },
+        "speech": {
+            "filename": "audio_encoder/model.onnx",
+            "config_filename": "audio_feature_extraction.json",
+            "inputs": {
+                "audio_embeds": "input_features",
+                "attention_mask": "input_features_mask"
+            },
+            "outputs": {
+                "audio_features": "audio_features"
+            },
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": []
+            }
+        },
+        "image_token_id": 258880,
+        "audio_token_id": 258881,
+        "boa_token_id": 256000
+    },
+    "search": {
+        "do_sample": true,
+        "early_stopping": true,
+        "max_length": 131072,
+        "min_length": 0,
+        "num_beams": 1,
+        "num_return_sequences": 1,
+        "past_present_share_buffer": false,
+        "repetition_penalty": 1.0,
+        "temperature": 1.0,
+        "top_k": 1,
+        "top_p": 1.0
+    }
+}

bf16/default/image_processor.json

@@ -0,0 +1,27 @@
+{
+    "processor": {
+        "name": "gemma_4_image_processing",
+        "transforms": [
+            {
+                "operation": {
+                    "name": "decode_image",
+                    "type": "DecodeImage",
+                    "attrs": {
+                        "color_space": "RGB"
+                    }
+                }
+            },
+            {
+                "operation": {
+                    "name": "gemma4_image_transform",
+                    "type": "Gemma4ImageTransform",
+                    "attrs": {
+                        "patch_size": 16,
+                        "max_soft_tokens": 280,
+                        "pooling_kernel_size": 3
+                    }
+                }
+            }
+        ]
+    }
+}

bf16/default/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc8d3a0ce36466ccc1278bf987df5f71db1719b9ca6b4118264f45cb627bfe0f
+size 32169626

bf16/default/tokenizer_config.json

@@ -0,0 +1,75 @@
+{
+  "audio_token": "<|audio|>",
+  "backend": "tokenizers",
+  "boa_token": "<|audio>",
+  "boi_token": "<|image>",
+  "bos_token": "<bos>",
+  "eoa_token": "<audio|>",
+  "eoc_token": "<channel|>",
+  "eoi_token": "<image|>",
+  "eos_token": "<eos>",
+  "eot_token": "<turn|>",
+  "escape_token": "<|\"|>",
+  "etc_token": "<tool_call|>",
+  "etd_token": "<tool|>",
+  "etr_token": "<tool_response|>",
+  "extra_special_tokens": [
+    "<|video|>"
+  ],
+  "image_token": "<|image|>",
+  "mask_token": "<mask>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "padding_side": "left",
+  "processor_class": "Gemma4Processor",
+  "response_schema": {
+    "type": "object",
+    "properties": {
+      "role": {
+        "const": "assistant"
+      },
+      "thinking": {
+        "type": "string"
+      },
+      "content": {
+        "type": "string"
+      },
+      "tool_calls": {
+        "x-regex-iterator": "<\\|tool_call>(.*?)<tool_call\\|>",
+        "type": "array",
+        "items": {
+          "type": "object",
+          "properties": {
+            "type": {
+              "const": "function"
+            },
+            "function": {
+              "type": "object",
+              "x-regex": "call\\:(?P<name>\\w+)(?P<arguments>\\{.*\\})",
+              "properties": {
+                "name": {
+                  "type": "string"
+                },
+                "arguments": {
+                  "type": "object",
+                  "x-parser": "gemma4-tool-call",
+                  "additionalProperties": {}
+                }
+              }
+            }
+          }
+        }
+      }
+    },
+    "x-regex": "(\\<\\|channel\\>thought\\n(?P<thinking>.*?)\\<channel\\|\\>)?(?P<tool_calls>\\<\\|tool_call\\>.*\\<tool_call\\|\\>)?(?P<content>(?:(?!\\<turn\\|\\>)(?!\\<\\|tool_response\\>).)+)?(?:\\<turn\\|\\>|\\<\\|tool_response\\>)?"
+  },
+  "soc_token": "<|channel>",
+  "sot_token": "<|turn>",
+  "stc_token": "<|tool_call>",
+  "std_token": "<|tool>",
+  "str_token": "<|tool_response>",
+  "think_token": "<|think|>",
+  "tokenizer_class": "GemmaTokenizer",
+  "unk_token": "<unk>",
+  "chat_template": "{%- macro format_parameters(properties, required, filter_keys=false) -%}\n    {%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}\n    {%- set ns = namespace(found_first=false) -%}\n    {%- for key, value in properties | dictsort -%}\n        {%- set add_comma = false -%}\n        {%- if not filter_keys or key not in standard_keys -%}\n            {%- if ns.found_first %},{% endif -%}\n            {%- set ns.found_first = true -%}\n            {{ key }}:{\n            {%- if value['description'] -%}\n                description:<|\"|>{{ value['description'] }}<|\"|>\n                {%- set add_comma = true -%}\n            {%- endif -%}\n            {%- if value['type'] | upper == 'STRING' -%}\n                {%- if value['enum'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    enum:{{ format_argument(value['enum']) }}\n                {%- endif -%}\n            {%- elif value['type'] | upper == 'ARRAY' -%}\n                {%- if value['items'] is mapping and value['items'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    items:{\n                    {%- set ns_items = namespace(found_first=false) -%}\n                    {%- for item_key, item_value in value['items'] | dictsort -%}\n                        {%- if item_value is not none -%}\n                            {%- if ns_items.found_first %},{% endif -%}\n                            {%- set ns_items.found_first = true -%}\n                            {%- if item_key == 'properties' -%}\n                                properties:{\n                                {%- if item_value is mapping -%}\n                                    {{- format_parameters(item_value, value['items']['required'] | default([])) -}}\n                                {%- endif -%}\n                                }\n                            {%- elif item_key == 'required' -%}\n                                required:[\n                                {%- for req_item in item_value -%}\n                                    <|\"|>{{- req_item -}}<|\"|>\n                                    {%- if not loop.last %},{% endif -%}\n                                {%- endfor -%}\n                                ]\n                            {%- elif item_key == 'type' -%}\n                                {%- if item_value is string -%}\n                                    type:{{ format_argument(item_value | upper) }}\n                                {%- else -%}\n                                    type:{{ format_argument(item_value | map('upper') | list) }}\n                                {%- endif -%}\n                            {%- else -%}\n                                {{ item_key }}:{{ format_argument(item_value) }}\n                            {%- endif -%}\n                        {%- endif -%}\n                    {%- endfor -%}\n                    }\n                {%- endif -%}\n            {%- endif -%}\n            {%- if value['nullable'] %}\n                {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                nullable:true\n            {%- endif -%}\n            {%- if value['type'] | upper == 'OBJECT' -%}\n                {%- if value['properties'] is defined and value['properties'] is mapping -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    properties:{\n                    {{- format_parameters(value['properties'], value['required'] | default([])) -}}\n                    }\n                {%- elif value is mapping -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    properties:{\n                    {{- format_parameters(value, value['required'] | default([]), filter_keys=true) -}}\n                    }\n                {%- endif -%}\n                {%- if value['required'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    required:[\n                    {%- for item in value['required'] | default([]) -%}\n                        <|\"|>{{- item -}}<|\"|>\n                        {%- if not loop.last %},{% endif -%}\n                    {%- endfor -%}\n                    ]\n                {%- endif -%}\n            {%- endif -%}\n            {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n            type:<|\"|>{{ value['type'] | upper }}<|\"|>}\n        {%- endif -%}\n    {%- endfor -%}\n{%- endmacro -%}\n{%- macro format_function_declaration(tool_data) -%}\n    declaration:{{- tool_data['function']['name'] -}}{description:<|\"|>{{- tool_data['function']['description'] -}}<|\"|>\n    {%- set params = tool_data['function']['parameters'] -%}\n    {%- if params -%}\n        ,parameters:{\n        {%- if params['properties'] -%}\n            properties:{ {{- format_parameters(params['properties'], params['required']) -}} },\n        {%- endif -%}\n        {%- if params['required'] -%}\n            required:[\n            {%- for item in params['required'] -%}\n                <|\"|>{{- item -}}<|\"|>\n                {{- ',' if not loop.last -}}\n            {%- endfor -%}\n            ],\n        {%- endif -%}\n        {%- if params['type'] -%}\n            type:<|\"|>{{- params['type'] | upper -}}<|\"|>}\n        {%- endif -%}\n    {%- endif -%}\n    {%- if 'response' in tool_data['function'] -%}\n        {%- set response_declaration = tool_data['function']['response'] -%}\n        ,response:{\n        {%- if response_declaration['description'] -%}\n            description:<|\"|>{{- response_declaration['description'] -}}<|\"|>,\n        {%- endif -%}\n        {%- if response_declaration['type'] | upper == 'OBJECT' -%}\n            type:<|\"|>{{- response_declaration['type'] | upper -}}<|\"|>}\n        {%- endif -%}\n    {%- endif -%}\n    }\n{%- endmacro -%}\n{%- macro format_argument(argument, escape_keys=True) -%}\n    {%- if argument is string -%}\n        {{- '<|\"|>' + argument + '<|\"|>' -}}\n    {%- elif argument is boolean -%}\n        {{- 'true' if argument else 'false' -}}\n    {%- elif argument is mapping -%}\n        {{- '{' -}}\n        {%- set ns = namespace(found_first=false) -%}\n        {%- for key, value in argument | dictsort -%}\n            {%- if ns.found_first %},{% endif -%}\n            {%- set ns.found_first = true -%}\n            {%- if escape_keys -%}\n                {{- '<|\"|>' + key + '<|\"|>' -}}\n            {%- else -%}\n                {{- key -}}\n            {%- endif -%}\n            :{{- format_argument(value, escape_keys=escape_keys) -}}\n        {%- endfor -%}\n        {{- '}' -}}\n    {%- elif argument is sequence -%}\n        {{- '[' -}}\n        {%- for item in argument -%}\n            {{- format_argument(item, escape_keys=escape_keys) -}}\n            {%- if not loop.last %},{% endif -%}\n        {%- endfor -%}\n        {{- ']' -}}\n    {%- else -%}\n        {{- argument -}}\n    {%- endif -%}\n{%- endmacro -%}\n{%- macro strip_thinking(text) -%}\n    {%- set ns = namespace(result='') -%}\n    {%- for part in text.split('<channel|>') -%}\n        {%- if '<|channel>' in part -%}\n            {%- set ns.result = ns.result + part.split('<|channel>')[0] -%}\n        {%- else -%}\n            {%- set ns.result = ns.result + part -%}\n        {%- endif -%}\n    {%- endfor -%}\n    {{- ns.result | trim -}}\n{%- endmacro -%}\n\n{%- macro format_tool_response_block(tool_name, response) -%}\n    {{- '<|tool_response>' -}}\n    {%- if response is mapping -%}\n        {{- 'response:' + tool_name + '{' -}}\n        {%- for key, value in response | dictsort -%}\n            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}\n            {%- if not loop.last %},{% endif -%}\n        {%- endfor -%}\n        {{- '}' -}}\n    {%- else -%}\n        {{- 'response:' + tool_name + '{value:' + format_argument(response, escape_keys=False) + '}' -}}\n    {%- endif -%}\n    {{- '<tool_response|>' -}}\n{%- endmacro -%}\n\n{%- set ns = namespace(prev_message_type=None) -%}\n{%- set loop_messages = messages -%}\n{{- bos_token -}}\n{#- Handle System/Tool Definitions Block -#}\n{%- if (enable_thinking is defined and enable_thinking) or tools or messages[0]['role'] in ['system', 'developer'] -%}\n    {{- '<|turn>system\\n' -}}\n    {#- Inject Thinking token at the very top of the FIRST system turn -#}\n    {%- if enable_thinking is defined and enable_thinking -%}\n        {{- '<|think|>\\n' -}}\n        {%- set ns.prev_message_type = 'think' -%}\n    {%- endif -%}\n    {%- if messages[0]['role'] in ['system', 'developer'] -%}\n        {%- if messages[0]['content'] is string -%}\n            {{- messages[0]['content'] | trim -}}\n        {%- elif messages[0]['content'] is sequence -%}\n            {%- for item in messages[0]['content'] -%}\n                {{- item['text'] | trim + ' '-}}\n            {%- endfor -%}\n        {%- endif -%}\n        {%- set loop_messages = messages[1:] -%}\n    {%- endif -%}\n    {%- if tools -%}\n        {%- for tool in tools %}\n            {{- '<|tool>' -}}\n            {{- format_function_declaration(tool) | trim -}}\n            {{- '<tool|>' -}}\n        {%- endfor %}\n        {%- set ns.prev_message_type = 'tool' -%}\n    {%- endif -%}\n    {{- '<turn|>\\n' -}}\n{%- endif %}\n\n{#- Pre-scan: find last user message index for reasoning guard -#}\n{%- set ns_turn = namespace(last_user_idx=-1) -%}\n{%- for i in range(loop_messages | length) -%}\n    {%- if loop_messages[i]['role'] == 'user' -%}\n        {%- set ns_turn.last_user_idx = i -%}\n    {%- endif -%}\n{%- endfor -%}\n\n{#- Loop through messages -#}\n{%- for message in loop_messages -%}\n    {%- if message['role'] != 'tool' -%}\n    {%- set ns.prev_message_type = None -%}\n    {%- set role = 'model' if message['role'] == 'assistant' else message['role'] -%}\n    {#- Detect continuation: suppress duplicate <|turn>model when previous non-tool message was also assistant -#}\n    {%- set prev_nt = namespace(role=None, found=false) -%}\n    {%- if loop.index0 > 0 -%}\n        {%- for j in range(loop.index0 - 1, -1, -1) -%}\n            {%- if not prev_nt.found -%}\n                {%- if loop_messages[j]['role'] != 'tool' -%}\n                    {%- set prev_nt.role = loop_messages[j]['role'] -%}\n                    {%- set prev_nt.found = true -%}\n                {%- endif -%}\n            {%- endif -%}\n        {%- endfor -%}\n    {%- endif -%}\n    {%- set continue_same_model_turn = (role == 'model' and prev_nt.role == 'assistant') -%}\n    {%- if not continue_same_model_turn -%}\n        {{- '<|turn>' + role + '\\n' }}\n    {%- endif -%}\n\n    {#- Render reasoning/reasoning_content as thinking channel -#}\n    {%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}\n    {%- if thinking_text and loop.index0 > ns_turn.last_user_idx and message.get('tool_calls') -%}\n        {{- '<|channel>thought\\n' + thinking_text + '\\n<channel|>' -}}\n    {%- endif -%}\n\n            {%- if message['tool_calls'] -%}\n                {%- for tool_call in message['tool_calls'] -%}\n                    {%- set function = tool_call['function'] -%}\n                    {{- '<|tool_call>call:' + function['name'] + '{' -}}\n                    {%- if function['arguments'] is mapping -%}\n                        {%- set ns_args = namespace(found_first=false) -%}\n                        {%- for key, value in function['arguments'] | dictsort -%}\n                            {%- if ns_args.found_first %},{% endif -%}\n                            {%- set ns_args.found_first = true -%}\n                            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}\n                        {%- endfor -%}\n                    {%- elif function['arguments'] is string -%}\n                        {{- function['arguments'] -}}\n                    {%- endif -%}\n                    {{- '}<tool_call|>' -}}\n                {%- endfor -%}\n                {%- set ns.prev_message_type = 'tool_call' -%}\n            {%- endif -%}\n\n            {%- set ns_tr_out = namespace(flag=false) -%}\n            {%- if message.get('tool_responses') -%}\n                {#- Legacy: tool_responses embedded on the assistant message (Google/Gemma native) -#}\n                {%- for tool_response in message['tool_responses'] -%}\n                    {{- format_tool_response_block(tool_response['name'] | default('unknown'), tool_response['response']) -}}\n                    {%- set ns_tr_out.flag = true -%}\n                    {%- set ns.prev_message_type = 'tool_response' -%}\n                {%- endfor -%}\n            {%- elif message.get('tool_calls') -%}\n                {#- OpenAI Chat Completions: forward-scan consecutive role:tool messages -#}\n                {%- set ns_tool_scan = namespace(stopped=false) -%}\n                {%- for k in range(loop.index0 + 1, loop_messages | length) -%}\n                    {%- if ns_tool_scan.stopped -%}\n                    {%- elif loop_messages[k]['role'] != 'tool' -%}\n                        {%- set ns_tool_scan.stopped = true -%}\n                    {%- else -%}\n                        {%- set follow = loop_messages[k] -%}\n                        {#- Resolve tool_call_id to function name -#}\n                        {%- set ns_tname = namespace(name=follow.get('name') | default('unknown')) -%}\n                        {%- for tc in message['tool_calls'] -%}\n                            {%- if tc.get('id') == follow.get('tool_call_id') -%}\n                                {%- set ns_tname.name = tc['function']['name'] -%}\n                            {%- endif -%}\n                        {%- endfor -%}\n                        {#- Handle content as string or content-parts array -#}\n                        {%- set tool_body = follow.get('content') -%}\n                        {%- if tool_body is string -%}\n                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}\n                        {%- elif tool_body is sequence and tool_body is not string -%}\n                            {%- set ns_txt = namespace(s='') -%}\n                            {%- for part in tool_body -%}\n                                {%- if part.get('type') == 'text' -%}\n                                    {%- set ns_txt.s = ns_txt.s + (part.get('text') | default('')) -%}\n                                {%- endif -%}\n                            {%- endfor -%}\n                            {{- format_tool_response_block(ns_tname.name, ns_txt.s) -}}\n                        {%- else -%}\n                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}\n                        {%- endif -%}\n                        {%- set ns_tr_out.flag = true -%}\n                        {%- set ns.prev_message_type = 'tool_response' -%}\n                    {%- endif -%}\n                {%- endfor -%}\n            {%- endif -%}\n\n            {%- set captured_content -%}\n            {%- if message['content'] is string -%}\n                {%- if role == 'model' -%}\n                    {{- strip_thinking(message['content']) -}}\n                {%- else -%}\n                    {{- message['content'] | trim -}}\n                {%- endif -%}\n            {%- elif message['content'] is sequence -%}\n                {%- for item in message['content'] -%}\n                    {%- if item['type'] == 'text' -%}\n                        {%- if role == 'model' -%}\n                            {{- strip_thinking(item['text']) -}}\n                        {%- else -%}\n                            {{- item['text'] | trim -}}\n                        {%- endif -%}\n                    {%- elif item['type'] == 'image' -%}\n                        {{- '<|image|>' -}}\n                        {%- set ns.prev_message_type = 'image' -%}\n                    {%- elif item['type'] == 'audio' -%}\n                        {{- '<|audio|>' -}}\n                        {%- set ns.prev_message_type = 'audio' -%}\n                    {%- elif item['type'] == 'video' -%}\n                        {{- '<|video|>' -}}\n                        {%- set ns.prev_message_type = 'video' -%}\n                    {%- endif -%}\n                {%- endfor -%}\n            {%- endif -%}\n            {%- endset -%}\n\n            {{- captured_content -}}\n            {%- set has_content = captured_content | trim | length > 0 -%}\n\n        {%- if ns.prev_message_type == 'tool_call' and not ns_tr_out.flag -%}\n            {{- '<|tool_response>' -}}\n        {%- elif not (ns_tr_out.flag and not has_content) -%}\n            {{- '<turn|>\\n' -}}\n        {%- endif -%}\n    {%- endif -%}\n{%- endfor -%}\n\n{%- if add_generation_prompt -%}\n    {%- if ns.prev_message_type != 'tool_response' and ns.prev_message_type != 'tool_call' -%}\n        {{- '<|turn>model\\n' -}}\n    {%- endif -%}\n{%- endif -%}"
+}

bf16/default/vision_encoder/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b1aa8d01379401cce914cd2bed70e776238d174ee6b08c016ae94c6b3617c48
+size 1239595

bf16/default/vision_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33fd860d1ce013444d2f16667e4625045971ad0ca905d1cb3b62feae2e937ba9
+size 358083464

bf16/onnx-standard/audio_encoder/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d470d65d71c8e052b0cbf28b835cbe2168cb9d2c04747c1e1d1c69f2ef120f22
+size 1238451

bf16/onnx-standard/audio_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbbb99ee9ece21e21b59d81c3d2a60b04d05dc4511951a2a8ab5aa48af4e3c0d
+size 614725024

bf16/onnx-standard/audio_feature_extraction.json

@@ -0,0 +1,30 @@
+{
+    "feature_extraction": {
+        "sequence": [
+            {
+                "operation": {
+                    "name": "audio_decoder",
+                    "type": "AudioDecoder"
+                }
+            },
+            {
+                "operation": {
+                    "name": "gemma4_log_mel",
+                    "type": "Gemma4LogMel",
+                    "attrs": {
+                        "feature_size": 128,
+                        "sampling_rate": 16000,
+                        "frame_length_ms": 20.0,
+                        "hop_length_ms": 10.0,
+                        "min_frequency": 0.0,
+                        "max_frequency": 8000.0,
+                        "preemphasis": 0.0,
+                        "preemphasis_htk_flavor": 1,
+                        "fft_overdrive": 0,
+                        "mel_floor": 0.001
+                    }
+                }
+            }
+        ]
+    }
+}

bf16/onnx-standard/chat_template.jinja

@@ -0,0 +1,351 @@
+{%- macro format_parameters(properties, required, filter_keys=false) -%}
+    {%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}
+    {%- set ns = namespace(found_first=false) -%}
+    {%- for key, value in properties | dictsort -%}
+        {%- set add_comma = false -%}
+        {%- if not filter_keys or key not in standard_keys -%}
+            {%- if ns.found_first %},{% endif -%}
+            {%- set ns.found_first = true -%}
+            {{ key }}:{
+            {%- if value['description'] -%}
+                description:<|"|>{{ value['description'] }}<|"|>
+                {%- set add_comma = true -%}
+            {%- endif -%}
+            {%- if value['type'] | upper == 'STRING' -%}
+                {%- if value['enum'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    enum:{{ format_argument(value['enum']) }}
+                {%- endif -%}
+            {%- elif value['type'] | upper == 'ARRAY' -%}
+                {%- if value['items'] is mapping and value['items'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    items:{
+                    {%- set ns_items = namespace(found_first=false) -%}
+                    {%- for item_key, item_value in value['items'] | dictsort -%}
+                        {%- if item_value is not none -%}
+                            {%- if ns_items.found_first %},{% endif -%}
+                            {%- set ns_items.found_first = true -%}
+                            {%- if item_key == 'properties' -%}
+                                properties:{
+                                {%- if item_value is mapping -%}
+                                    {{- format_parameters(item_value, value['items']['required'] | default([])) -}}
+                                {%- endif -%}
+                                }
+                            {%- elif item_key == 'required' -%}
+                                required:[
+                                {%- for req_item in item_value -%}
+                                    <|"|>{{- req_item -}}<|"|>
+                                    {%- if not loop.last %},{% endif -%}
+                                {%- endfor -%}
+                                ]
+                            {%- elif item_key == 'type' -%}
+                                {%- if item_value is string -%}
+                                    type:{{ format_argument(item_value | upper) }}
+                                {%- else -%}
+                                    type:{{ format_argument(item_value | map('upper') | list) }}
+                                {%- endif -%}
+                            {%- else -%}
+                                {{ item_key }}:{{ format_argument(item_value) }}
+                            {%- endif -%}
+                        {%- endif -%}
+                    {%- endfor -%}
+                    }
+                {%- endif -%}
+            {%- endif -%}
+            {%- if value['nullable'] %}
+                {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                nullable:true
+            {%- endif -%}
+            {%- if value['type'] | upper == 'OBJECT' -%}
+                {%- if value['properties'] is defined and value['properties'] is mapping -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    properties:{
+                    {{- format_parameters(value['properties'], value['required'] | default([])) -}}
+                    }
+                {%- elif value is mapping -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    properties:{
+                    {{- format_parameters(value, value['required'] | default([]), filter_keys=true) -}}
+                    }
+                {%- endif -%}
+                {%- if value['required'] -%}
+                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+                    required:[
+                    {%- for item in value['required'] | default([]) -%}
+                        <|"|>{{- item -}}<|"|>
+                        {%- if not loop.last %},{% endif -%}
+                    {%- endfor -%}
+                    ]
+                {%- endif -%}
+            {%- endif -%}
+            {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}
+            type:<|"|>{{ value['type'] | upper }}<|"|>}
+        {%- endif -%}
+    {%- endfor -%}
+{%- endmacro -%}
+{%- macro format_function_declaration(tool_data) -%}
+    declaration:{{- tool_data['function']['name'] -}}{description:<|"|>{{- tool_data['function']['description'] -}}<|"|>
+    {%- set params = tool_data['function']['parameters'] -%}
+    {%- if params -%}
+        ,parameters:{
+        {%- if params['properties'] -%}
+            properties:{ {{- format_parameters(params['properties'], params['required']) -}} },
+        {%- endif -%}
+        {%- if params['required'] -%}
+            required:[
+            {%- for item in params['required'] -%}
+                <|"|>{{- item -}}<|"|>
+                {{- ',' if not loop.last -}}
+            {%- endfor -%}
+            ],
+        {%- endif -%}
+        {%- if params['type'] -%}
+            type:<|"|>{{- params['type'] | upper -}}<|"|>}
+        {%- endif -%}
+    {%- endif -%}
+    {%- if 'response' in tool_data['function'] -%}
+        {%- set response_declaration = tool_data['function']['response'] -%}
+        ,response:{
+        {%- if response_declaration['description'] -%}
+            description:<|"|>{{- response_declaration['description'] -}}<|"|>,
+        {%- endif -%}
+        {%- if response_declaration['type'] | upper == 'OBJECT' -%}
+            type:<|"|>{{- response_declaration['type'] | upper -}}<|"|>}
+        {%- endif -%}
+    {%- endif -%}
+    }
+{%- endmacro -%}
+{%- macro format_argument(argument, escape_keys=True) -%}
+    {%- if argument is string -%}
+        {{- '<|"|>' + argument + '<|"|>' -}}
+    {%- elif argument is boolean -%}
+        {{- 'true' if argument else 'false' -}}
+    {%- elif argument is mapping -%}
+        {{- '{' -}}
+        {%- set ns = namespace(found_first=false) -%}
+        {%- for key, value in argument | dictsort -%}
+            {%- if ns.found_first %},{% endif -%}
+            {%- set ns.found_first = true -%}
+            {%- if escape_keys -%}
+                {{- '<|"|>' + key + '<|"|>' -}}
+            {%- else -%}
+                {{- key -}}
+            {%- endif -%}
+            :{{- format_argument(value, escape_keys=escape_keys) -}}
+        {%- endfor -%}
+        {{- '}' -}}
+    {%- elif argument is sequence -%}
+        {{- '[' -}}
+        {%- for item in argument -%}
+            {{- format_argument(item, escape_keys=escape_keys) -}}
+            {%- if not loop.last %},{% endif -%}
+        {%- endfor -%}
+        {{- ']' -}}
+    {%- else -%}
+        {{- argument -}}
+    {%- endif -%}
+{%- endmacro -%}
+{%- macro strip_thinking(text) -%}
+    {%- set ns = namespace(result='') -%}
+    {%- for part in text.split('<channel|>') -%}
+        {%- if '<|channel>' in part -%}
+            {%- set ns.result = ns.result + part.split('<|channel>')[0] -%}
+        {%- else -%}
+            {%- set ns.result = ns.result + part -%}
+        {%- endif -%}
+    {%- endfor -%}
+    {{- ns.result | trim -}}
+{%- endmacro -%}
+
+{%- macro format_tool_response_block(tool_name, response) -%}
+    {{- '<|tool_response>' -}}
+    {%- if response is mapping -%}
+        {{- 'response:' + tool_name + '{' -}}
+        {%- for key, value in response | dictsort -%}
+            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
+            {%- if not loop.last %},{% endif -%}
+        {%- endfor -%}
+        {{- '}' -}}
+    {%- else -%}
+        {{- 'response:' + tool_name + '{value:' + format_argument(response, escape_keys=False) + '}' -}}
+    {%- endif -%}
+    {{- '<tool_response|>' -}}
+{%- endmacro -%}
+
+{%- set ns = namespace(prev_message_type=None) -%}
+{%- set loop_messages = messages -%}
+{{- bos_token -}}
+{#- Handle System/Tool Definitions Block -#}
+{%- if (enable_thinking is defined and enable_thinking) or tools or messages[0]['role'] in ['system', 'developer'] -%}
+    {{- '<|turn>system\n' -}}
+    {#- Inject Thinking token at the very top of the FIRST system turn -#}
+    {%- if enable_thinking is defined and enable_thinking -%}
+        {{- '<|think|>\n' -}}
+        {%- set ns.prev_message_type = 'think' -%}
+    {%- endif -%}
+    {%- if messages[0]['role'] in ['system', 'developer'] -%}
+        {%- if messages[0]['content'] is string -%}
+            {{- messages[0]['content'] | trim -}}
+        {%- elif messages[0]['content'] is sequence -%}
+            {%- for item in messages[0]['content'] -%}
+                {{- item['text'] | trim + ' '-}}
+            {%- endfor -%}
+        {%- endif -%}
+        {%- set loop_messages = messages[1:] -%}
+    {%- endif -%}
+    {%- if tools -%}
+        {%- for tool in tools %}
+            {{- '<|tool>' -}}
+            {{- format_function_declaration(tool) | trim -}}
+            {{- '<tool|>' -}}
+        {%- endfor %}
+        {%- set ns.prev_message_type = 'tool' -%}
+    {%- endif -%}
+    {{- '<turn|>\n' -}}
+{%- endif %}
+
+{#- Pre-scan: find last user message index for reasoning guard -#}
+{%- set ns_turn = namespace(last_user_idx=-1) -%}
+{%- for i in range(loop_messages | length) -%}
+    {%- if loop_messages[i]['role'] == 'user' -%}
+        {%- set ns_turn.last_user_idx = i -%}
+    {%- endif -%}
+{%- endfor -%}
+
+{#- Loop through messages -#}
+{%- for message in loop_messages -%}
+    {%- if message['role'] != 'tool' -%}
+    {%- set ns.prev_message_type = None -%}
+    {%- set role = 'model' if message['role'] == 'assistant' else message['role'] -%}
+    {#- Detect continuation: suppress duplicate <|turn>model when previous non-tool message was also assistant -#}
+    {%- set prev_nt = namespace(role=None, found=false) -%}
+    {%- if loop.index0 > 0 -%}
+        {%- for j in range(loop.index0 - 1, -1, -1) -%}
+            {%- if not prev_nt.found -%}
+                {%- if loop_messages[j]['role'] != 'tool' -%}
+                    {%- set prev_nt.role = loop_messages[j]['role'] -%}
+                    {%- set prev_nt.found = true -%}
+                {%- endif -%}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- endif -%}
+    {%- set continue_same_model_turn = (role == 'model' and prev_nt.role == 'assistant') -%}
+    {%- if not continue_same_model_turn -%}
+        {{- '<|turn>' + role + '\n' }}
+    {%- endif -%}
+
+    {#- Render reasoning/reasoning_content as thinking channel -#}
+    {%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}
+    {%- if thinking_text and loop.index0 > ns_turn.last_user_idx and message.get('tool_calls') -%}
+        {{- '<|channel>thought\n' + thinking_text + '\n<channel|>' -}}
+    {%- endif -%}
+
+            {%- if message['tool_calls'] -%}
+                {%- for tool_call in message['tool_calls'] -%}
+                    {%- set function = tool_call['function'] -%}
+                    {{- '<|tool_call>call:' + function['name'] + '{' -}}
+                    {%- if function['arguments'] is mapping -%}
+                        {%- set ns_args = namespace(found_first=false) -%}
+                        {%- for key, value in function['arguments'] | dictsort -%}
+                            {%- if ns_args.found_first %},{% endif -%}
+                            {%- set ns_args.found_first = true -%}
+                            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
+                        {%- endfor -%}
+                    {%- elif function['arguments'] is string -%}
+                        {{- function['arguments'] -}}
+                    {%- endif -%}
+                    {{- '}<tool_call|>' -}}
+                {%- endfor -%}
+                {%- set ns.prev_message_type = 'tool_call' -%}
+            {%- endif -%}
+
+            {%- set ns_tr_out = namespace(flag=false) -%}
+            {%- if message.get('tool_responses') -%}
+                {#- Legacy: tool_responses embedded on the assistant message (Google/Gemma native) -#}
+                {%- for tool_response in message['tool_responses'] -%}
+                    {{- format_tool_response_block(tool_response['name'] | default('unknown'), tool_response['response']) -}}
+                    {%- set ns_tr_out.flag = true -%}
+                    {%- set ns.prev_message_type = 'tool_response' -%}
+                {%- endfor -%}
+            {%- elif message.get('tool_calls') -%}
+                {#- OpenAI Chat Completions: forward-scan consecutive role:tool messages -#}
+                {%- set ns_tool_scan = namespace(stopped=false) -%}
+                {%- for k in range(loop.index0 + 1, loop_messages | length) -%}
+                    {%- if ns_tool_scan.stopped -%}
+                    {%- elif loop_messages[k]['role'] != 'tool' -%}
+                        {%- set ns_tool_scan.stopped = true -%}
+                    {%- else -%}
+                        {%- set follow = loop_messages[k] -%}
+                        {#- Resolve tool_call_id to function name -#}
+                        {%- set ns_tname = namespace(name=follow.get('name') | default('unknown')) -%}
+                        {%- for tc in message['tool_calls'] -%}
+                            {%- if tc.get('id') == follow.get('tool_call_id') -%}
+                                {%- set ns_tname.name = tc['function']['name'] -%}
+                            {%- endif -%}
+                        {%- endfor -%}
+                        {#- Handle content as string or content-parts array -#}
+                        {%- set tool_body = follow.get('content') -%}
+                        {%- if tool_body is string -%}
+                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}
+                        {%- elif tool_body is sequence and tool_body is not string -%}
+                            {%- set ns_txt = namespace(s='') -%}
+                            {%- for part in tool_body -%}
+                                {%- if part.get('type') == 'text' -%}
+                                    {%- set ns_txt.s = ns_txt.s + (part.get('text') | default('')) -%}
+                                {%- endif -%}
+                            {%- endfor -%}
+                            {{- format_tool_response_block(ns_tname.name, ns_txt.s) -}}
+                        {%- else -%}
+                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}
+                        {%- endif -%}
+                        {%- set ns_tr_out.flag = true -%}
+                        {%- set ns.prev_message_type = 'tool_response' -%}
+                    {%- endif -%}
+                {%- endfor -%}
+            {%- endif -%}
+
+            {%- set captured_content -%}
+            {%- if message['content'] is string -%}
+                {%- if role == 'model' -%}
+                    {{- strip_thinking(message['content']) -}}
+                {%- else -%}
+                    {{- message['content'] | trim -}}
+                {%- endif -%}
+            {%- elif message['content'] is sequence -%}
+                {%- for item in message['content'] -%}
+                    {%- if item['type'] == 'text' -%}
+                        {%- if role == 'model' -%}
+                            {{- strip_thinking(item['text']) -}}
+                        {%- else -%}
+                            {{- item['text'] | trim -}}
+                        {%- endif -%}
+                    {%- elif item['type'] == 'image' -%}
+                        {{- '<|image|>' -}}
+                        {%- set ns.prev_message_type = 'image' -%}
+                    {%- elif item['type'] == 'audio' -%}
+                        {{- '<|audio|>' -}}
+                        {%- set ns.prev_message_type = 'audio' -%}
+                    {%- elif item['type'] == 'video' -%}
+                        {{- '<|video|>' -}}
+                        {%- set ns.prev_message_type = 'video' -%}
+                    {%- endif -%}
+                {%- endfor -%}
+            {%- endif -%}
+            {%- endset -%}
+
+            {{- captured_content -}}
+            {%- set has_content = captured_content | trim | length > 0 -%}
+
+        {%- if ns.prev_message_type == 'tool_call' and not ns_tr_out.flag -%}
+            {{- '<|tool_response>' -}}
+        {%- elif not (ns_tr_out.flag and not has_content) -%}
+            {{- '<turn|>\n' -}}
+        {%- endif -%}
+    {%- endif -%}
+{%- endfor -%}
+
+{%- if add_generation_prompt -%}
+    {%- if ns.prev_message_type != 'tool_response' and ns.prev_message_type != 'tool_call' -%}
+        {{- '<|turn>model\n' -}}
+    {%- endif -%}
+{%- endif -%}

bf16/onnx-standard/decoder/model-00001-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e554c5c22e59f60d8d5eee7505bac60e74da7a8d4be57666b3e9a4c2e2fba046
+size 4999430520

bf16/onnx-standard/decoder/model-00002-of-00002.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e3a718a5ed0fbfe379dbd95d01ccb26a59522db7d6490645a83e6c20a63dd6a
+size 4459103336

bf16/onnx-standard/decoder/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0051af80094cb4f7bf92e81192f48513f34331eae8f64bd35b0cc85adf08e0bd
+size 914388

bf16/onnx-standard/decoder/model.safetensors.index.json

@@ -0,0 +1,550 @@
+{
+  "metadata": {
+    "total_size": 9458465280
+  },
+  "weight_map": {
+    "decoder.model.per_layer_projection_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.0.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.1.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.2.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.3.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.4.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.5.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.6.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.7.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.8.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.9.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.10.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.11.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.12.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.13.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.14.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.15.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.16.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.17.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.18.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.19.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.20.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.21.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.22.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.23.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.24.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.25.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.26.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.27.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.28.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.29.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.30.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.31.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.32.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.33.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.embed_tokens_per_layer.34.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_local.cos_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_local.sin_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_global.cos_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.rotary_emb_global.sin_cache": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.2.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.3.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.4.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.5.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.6.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.7.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.8.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.9.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.10.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.11.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.12.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.13.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.k_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.14.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.15.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.16.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.17.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.18.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.19.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.20.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.21.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.22.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.23.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.24.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.25.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.26.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.27.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.28.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.29.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.30.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.31.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.32.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.33.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.input_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.self_attn.q_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.post_attention_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.pre_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.post_feedforward_layernorm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.34.post_per_layer_input_norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.norm.weight": "model-00001-of-00002.safetensors",
+    "decoder.model.per_layer_model_projection.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.q_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.k_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.v_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.self_attn.o_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.mlp.gate_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.mlp.up_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.mlp.down_proj.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.per_layer_input_gate.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.0.per_layer_projection.weight_t": "model-00001-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.1.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.2.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.3.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.4.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.5.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.6.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.7.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.8.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.9.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.10.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.11.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.12.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.13.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.k_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.v_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.14.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.15.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.16.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.17.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.18.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.19.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.20.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.21.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.22.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.23.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.24.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.25.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.26.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.27.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.28.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.29.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.30.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.31.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.32.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.33.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.self_attn.q_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.self_attn.o_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.mlp.gate_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.mlp.up_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.mlp.down_proj.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.per_layer_input_gate.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.model.layers.34.per_layer_projection.weight_t": "model-00002-of-00002.safetensors",
+    "decoder.lm_head.weight_t": "model-00002-of-00002.safetensors"
+  }
+}

bf16/onnx-standard/embedding/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47d105fc7ae637554a6552b90e4335caa175427b8770d1891e3602cb869776e0
+size 8656

bf16/onnx-standard/embedding/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:741c3297844e13c22a8eae3db118a94ae7d296e490d25e12e1a90c3d143f8681
+size 805312824

bf16/onnx-standard/genai_config.json

@@ -0,0 +1,100 @@
+{
+    "model": {
+        "type": "gemma4",
+        "vocab_size": 262144,
+        "context_length": 131072,
+        "decoder": {
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": []
+            },
+            "filename": "decoder/model.onnx",
+            "head_size": 256,
+            "hidden_size": 1536,
+            "inputs": {
+                "inputs_embeds": "inputs_embeds",
+                "attention_mask": "attention_mask",
+                "position_ids": "position_ids",
+                "input_ids": "input_ids",
+                "past_key_names": "past_key_values.%d.key",
+                "past_value_names": "past_key_values.%d.value"
+            },
+            "outputs": {
+                "logits": "logits",
+                "present_key_names": "present.%d.key",
+                "present_value_names": "present.%d.value"
+            },
+            "num_attention_heads": 8,
+            "num_hidden_layers": 35,
+            "num_key_value_heads": 1
+        },
+        "bos_token_id": 2,
+        "eos_token_id": [
+            1,
+            106
+        ],
+        "pad_token_id": 0,
+        "vision": {
+            "filename": "vision_encoder/model.onnx",
+            "config_filename": "image_processor.json",
+            "inputs": {
+                "pixel_values": "pixel_values",
+                "pixel_position_ids": "pixel_position_ids"
+            },
+            "outputs": {
+                "image_features": "image_features"
+            },
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": []
+            },
+            "spatial_merge_size": 2
+        },
+        "embedding": {
+            "filename": "embedding/model.onnx",
+            "inputs": {
+                "input_ids": "input_ids",
+                "image_features": "image_features",
+                "audio_features": "audio_features"
+            },
+            "outputs": {
+                "inputs_embeds": "inputs_embeds"
+            },
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": []
+            }
+        },
+        "speech": {
+            "filename": "audio_encoder/model.onnx",
+            "config_filename": "audio_feature_extraction.json",
+            "inputs": {
+                "audio_embeds": "input_features",
+                "attention_mask": "input_features_mask"
+            },
+            "outputs": {
+                "audio_features": "audio_features"
+            },
+            "session_options": {
+                "log_id": "onnxruntime-genai",
+                "provider_options": []
+            }
+        },
+        "image_token_id": 258880,
+        "audio_token_id": 258881,
+        "boa_token_id": 256000
+    },
+    "search": {
+        "do_sample": true,
+        "early_stopping": true,
+        "max_length": 131072,
+        "min_length": 0,
+        "num_beams": 1,
+        "num_return_sequences": 1,
+        "past_present_share_buffer": false,
+        "repetition_penalty": 1.0,
+        "temperature": 1.0,
+        "top_k": 1,
+        "top_p": 1.0
+    }
+}

bf16/onnx-standard/image_processor.json

@@ -0,0 +1,27 @@
+{
+    "processor": {
+        "name": "gemma_4_image_processing",
+        "transforms": [
+            {
+                "operation": {
+                    "name": "decode_image",
+                    "type": "DecodeImage",
+                    "attrs": {
+                        "color_space": "RGB"
+                    }
+                }
+            },
+            {
+                "operation": {
+                    "name": "gemma4_image_transform",
+                    "type": "Gemma4ImageTransform",
+                    "attrs": {
+                        "patch_size": 16,
+                        "max_soft_tokens": 280,
+                        "pooling_kernel_size": 3
+                    }
+                }
+            }
+        ]
+    }
+}

bf16/onnx-standard/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc8d3a0ce36466ccc1278bf987df5f71db1719b9ca6b4118264f45cb627bfe0f
+size 32169626

bf16/onnx-standard/tokenizer_config.json

@@ -0,0 +1,75 @@
+{
+  "audio_token": "<|audio|>",
+  "backend": "tokenizers",
+  "boa_token": "<|audio>",
+  "boi_token": "<|image>",
+  "bos_token": "<bos>",
+  "eoa_token": "<audio|>",
+  "eoc_token": "<channel|>",
+  "eoi_token": "<image|>",
+  "eos_token": "<eos>",
+  "eot_token": "<turn|>",
+  "escape_token": "<|\"|>",
+  "etc_token": "<tool_call|>",
+  "etd_token": "<tool|>",
+  "etr_token": "<tool_response|>",
+  "extra_special_tokens": [
+    "<|video|>"
+  ],
+  "image_token": "<|image|>",
+  "mask_token": "<mask>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "padding_side": "left",
+  "processor_class": "Gemma4Processor",
+  "response_schema": {
+    "type": "object",
+    "properties": {
+      "role": {
+        "const": "assistant"
+      },
+      "thinking": {
+        "type": "string"
+      },
+      "content": {
+        "type": "string"
+      },
+      "tool_calls": {
+        "x-regex-iterator": "<\\|tool_call>(.*?)<tool_call\\|>",
+        "type": "array",
+        "items": {
+          "type": "object",
+          "properties": {
+            "type": {
+              "const": "function"
+            },
+            "function": {
+              "type": "object",
+              "x-regex": "call\\:(?P<name>\\w+)(?P<arguments>\\{.*\\})",
+              "properties": {
+                "name": {
+                  "type": "string"
+                },
+                "arguments": {
+                  "type": "object",
+                  "x-parser": "gemma4-tool-call",
+                  "additionalProperties": {}
+                }
+              }
+            }
+          }
+        }
+      }
+    },
+    "x-regex": "(\\<\\|channel\\>thought\\n(?P<thinking>.*?)\\<channel\\|\\>)?(?P<tool_calls>\\<\\|tool_call\\>.*\\<tool_call\\|\\>)?(?P<content>(?:(?!\\<turn\\|\\>)(?!\\<\\|tool_response\\>).)+)?(?:\\<turn\\|\\>|\\<\\|tool_response\\>)?"
+  },
+  "soc_token": "<|channel>",
+  "sot_token": "<|turn>",
+  "stc_token": "<|tool_call>",
+  "std_token": "<|tool>",
+  "str_token": "<|tool_response>",
+  "think_token": "<|think|>",
+  "tokenizer_class": "GemmaTokenizer",
+  "unk_token": "<unk>",
+  "chat_template": "{%- macro format_parameters(properties, required, filter_keys=false) -%}\n    {%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}\n    {%- set ns = namespace(found_first=false) -%}\n    {%- for key, value in properties | dictsort -%}\n        {%- set add_comma = false -%}\n        {%- if not filter_keys or key not in standard_keys -%}\n            {%- if ns.found_first %},{% endif -%}\n            {%- set ns.found_first = true -%}\n            {{ key }}:{\n            {%- if value['description'] -%}\n                description:<|\"|>{{ value['description'] }}<|\"|>\n                {%- set add_comma = true -%}\n            {%- endif -%}\n            {%- if value['type'] | upper == 'STRING' -%}\n                {%- if value['enum'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    enum:{{ format_argument(value['enum']) }}\n                {%- endif -%}\n            {%- elif value['type'] | upper == 'ARRAY' -%}\n                {%- if value['items'] is mapping and value['items'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    items:{\n                    {%- set ns_items = namespace(found_first=false) -%}\n                    {%- for item_key, item_value in value['items'] | dictsort -%}\n                        {%- if item_value is not none -%}\n                            {%- if ns_items.found_first %},{% endif -%}\n                            {%- set ns_items.found_first = true -%}\n                            {%- if item_key == 'properties' -%}\n                                properties:{\n                                {%- if item_value is mapping -%}\n                                    {{- format_parameters(item_value, value['items']['required'] | default([])) -}}\n                                {%- endif -%}\n                                }\n                            {%- elif item_key == 'required' -%}\n                                required:[\n                                {%- for req_item in item_value -%}\n                                    <|\"|>{{- req_item -}}<|\"|>\n                                    {%- if not loop.last %},{% endif -%}\n                                {%- endfor -%}\n                                ]\n                            {%- elif item_key == 'type' -%}\n                                {%- if item_value is string -%}\n                                    type:{{ format_argument(item_value | upper) }}\n                                {%- else -%}\n                                    type:{{ format_argument(item_value | map('upper') | list) }}\n                                {%- endif -%}\n                            {%- else -%}\n                                {{ item_key }}:{{ format_argument(item_value) }}\n                            {%- endif -%}\n                        {%- endif -%}\n                    {%- endfor -%}\n                    }\n                {%- endif -%}\n            {%- endif -%}\n            {%- if value['nullable'] %}\n                {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                nullable:true\n            {%- endif -%}\n            {%- if value['type'] | upper == 'OBJECT' -%}\n                {%- if value['properties'] is defined and value['properties'] is mapping -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    properties:{\n                    {{- format_parameters(value['properties'], value['required'] | default([])) -}}\n                    }\n                {%- elif value is mapping -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    properties:{\n                    {{- format_parameters(value, value['required'] | default([]), filter_keys=true) -}}\n                    }\n                {%- endif -%}\n                {%- if value['required'] -%}\n                    {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n                    required:[\n                    {%- for item in value['required'] | default([]) -%}\n                        <|\"|>{{- item -}}<|\"|>\n                        {%- if not loop.last %},{% endif -%}\n                    {%- endfor -%}\n                    ]\n                {%- endif -%}\n            {%- endif -%}\n            {%- if add_comma %},{%- else -%} {%- set add_comma = true -%} {% endif -%}\n            type:<|\"|>{{ value['type'] | upper }}<|\"|>}\n        {%- endif -%}\n    {%- endfor -%}\n{%- endmacro -%}\n{%- macro format_function_declaration(tool_data) -%}\n    declaration:{{- tool_data['function']['name'] -}}{description:<|\"|>{{- tool_data['function']['description'] -}}<|\"|>\n    {%- set params = tool_data['function']['parameters'] -%}\n    {%- if params -%}\n        ,parameters:{\n        {%- if params['properties'] -%}\n            properties:{ {{- format_parameters(params['properties'], params['required']) -}} },\n        {%- endif -%}\n        {%- if params['required'] -%}\n            required:[\n            {%- for item in params['required'] -%}\n                <|\"|>{{- item -}}<|\"|>\n                {{- ',' if not loop.last -}}\n            {%- endfor -%}\n            ],\n        {%- endif -%}\n        {%- if params['type'] -%}\n            type:<|\"|>{{- params['type'] | upper -}}<|\"|>}\n        {%- endif -%}\n    {%- endif -%}\n    {%- if 'response' in tool_data['function'] -%}\n        {%- set response_declaration = tool_data['function']['response'] -%}\n        ,response:{\n        {%- if response_declaration['description'] -%}\n            description:<|\"|>{{- response_declaration['description'] -}}<|\"|>,\n        {%- endif -%}\n        {%- if response_declaration['type'] | upper == 'OBJECT' -%}\n            type:<|\"|>{{- response_declaration['type'] | upper -}}<|\"|>}\n        {%- endif -%}\n    {%- endif -%}\n    }\n{%- endmacro -%}\n{%- macro format_argument(argument, escape_keys=True) -%}\n    {%- if argument is string -%}\n        {{- '<|\"|>' + argument + '<|\"|>' -}}\n    {%- elif argument is boolean -%}\n        {{- 'true' if argument else 'false' -}}\n    {%- elif argument is mapping -%}\n        {{- '{' -}}\n        {%- set ns = namespace(found_first=false) -%}\n        {%- for key, value in argument | dictsort -%}\n            {%- if ns.found_first %},{% endif -%}\n            {%- set ns.found_first = true -%}\n            {%- if escape_keys -%}\n                {{- '<|\"|>' + key + '<|\"|>' -}}\n            {%- else -%}\n                {{- key -}}\n            {%- endif -%}\n            :{{- format_argument(value, escape_keys=escape_keys) -}}\n        {%- endfor -%}\n        {{- '}' -}}\n    {%- elif argument is sequence -%}\n        {{- '[' -}}\n        {%- for item in argument -%}\n            {{- format_argument(item, escape_keys=escape_keys) -}}\n            {%- if not loop.last %},{% endif -%}\n        {%- endfor -%}\n        {{- ']' -}}\n    {%- else -%}\n        {{- argument -}}\n    {%- endif -%}\n{%- endmacro -%}\n{%- macro strip_thinking(text) -%}\n    {%- set ns = namespace(result='') -%}\n    {%- for part in text.split('<channel|>') -%}\n        {%- if '<|channel>' in part -%}\n            {%- set ns.result = ns.result + part.split('<|channel>')[0] -%}\n        {%- else -%}\n            {%- set ns.result = ns.result + part -%}\n        {%- endif -%}\n    {%- endfor -%}\n    {{- ns.result | trim -}}\n{%- endmacro -%}\n\n{%- macro format_tool_response_block(tool_name, response) -%}\n    {{- '<|tool_response>' -}}\n    {%- if response is mapping -%}\n        {{- 'response:' + tool_name + '{' -}}\n        {%- for key, value in response | dictsort -%}\n            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}\n            {%- if not loop.last %},{% endif -%}\n        {%- endfor -%}\n        {{- '}' -}}\n    {%- else -%}\n        {{- 'response:' + tool_name + '{value:' + format_argument(response, escape_keys=False) + '}' -}}\n    {%- endif -%}\n    {{- '<tool_response|>' -}}\n{%- endmacro -%}\n\n{%- set ns = namespace(prev_message_type=None) -%}\n{%- set loop_messages = messages -%}\n{{- bos_token -}}\n{#- Handle System/Tool Definitions Block -#}\n{%- if (enable_thinking is defined and enable_thinking) or tools or messages[0]['role'] in ['system', 'developer'] -%}\n    {{- '<|turn>system\\n' -}}\n    {#- Inject Thinking token at the very top of the FIRST system turn -#}\n    {%- if enable_thinking is defined and enable_thinking -%}\n        {{- '<|think|>\\n' -}}\n        {%- set ns.prev_message_type = 'think' -%}\n    {%- endif -%}\n    {%- if messages[0]['role'] in ['system', 'developer'] -%}\n        {%- if messages[0]['content'] is string -%}\n            {{- messages[0]['content'] | trim -}}\n        {%- elif messages[0]['content'] is sequence -%}\n            {%- for item in messages[0]['content'] -%}\n                {{- item['text'] | trim + ' '-}}\n            {%- endfor -%}\n        {%- endif -%}\n        {%- set loop_messages = messages[1:] -%}\n    {%- endif -%}\n    {%- if tools -%}\n        {%- for tool in tools %}\n            {{- '<|tool>' -}}\n            {{- format_function_declaration(tool) | trim -}}\n            {{- '<tool|>' -}}\n        {%- endfor %}\n        {%- set ns.prev_message_type = 'tool' -%}\n    {%- endif -%}\n    {{- '<turn|>\\n' -}}\n{%- endif %}\n\n{#- Pre-scan: find last user message index for reasoning guard -#}\n{%- set ns_turn = namespace(last_user_idx=-1) -%}\n{%- for i in range(loop_messages | length) -%}\n    {%- if loop_messages[i]['role'] == 'user' -%}\n        {%- set ns_turn.last_user_idx = i -%}\n    {%- endif -%}\n{%- endfor -%}\n\n{#- Loop through messages -#}\n{%- for message in loop_messages -%}\n    {%- if message['role'] != 'tool' -%}\n    {%- set ns.prev_message_type = None -%}\n    {%- set role = 'model' if message['role'] == 'assistant' else message['role'] -%}\n    {#- Detect continuation: suppress duplicate <|turn>model when previous non-tool message was also assistant -#}\n    {%- set prev_nt = namespace(role=None, found=false) -%}\n    {%- if loop.index0 > 0 -%}\n        {%- for j in range(loop.index0 - 1, -1, -1) -%}\n            {%- if not prev_nt.found -%}\n                {%- if loop_messages[j]['role'] != 'tool' -%}\n                    {%- set prev_nt.role = loop_messages[j]['role'] -%}\n                    {%- set prev_nt.found = true -%}\n                {%- endif -%}\n            {%- endif -%}\n        {%- endfor -%}\n    {%- endif -%}\n    {%- set continue_same_model_turn = (role == 'model' and prev_nt.role == 'assistant') -%}\n    {%- if not continue_same_model_turn -%}\n        {{- '<|turn>' + role + '\\n' }}\n    {%- endif -%}\n\n    {#- Render reasoning/reasoning_content as thinking channel -#}\n    {%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}\n    {%- if thinking_text and loop.index0 > ns_turn.last_user_idx and message.get('tool_calls') -%}\n        {{- '<|channel>thought\\n' + thinking_text + '\\n<channel|>' -}}\n    {%- endif -%}\n\n            {%- if message['tool_calls'] -%}\n                {%- for tool_call in message['tool_calls'] -%}\n                    {%- set function = tool_call['function'] -%}\n                    {{- '<|tool_call>call:' + function['name'] + '{' -}}\n                    {%- if function['arguments'] is mapping -%}\n                        {%- set ns_args = namespace(found_first=false) -%}\n                        {%- for key, value in function['arguments'] | dictsort -%}\n                            {%- if ns_args.found_first %},{% endif -%}\n                            {%- set ns_args.found_first = true -%}\n                            {{- key -}}:{{- format_argument(value, escape_keys=False) -}}\n                        {%- endfor -%}\n                    {%- elif function['arguments'] is string -%}\n                        {{- function['arguments'] -}}\n                    {%- endif -%}\n                    {{- '}<tool_call|>' -}}\n                {%- endfor -%}\n                {%- set ns.prev_message_type = 'tool_call' -%}\n            {%- endif -%}\n\n            {%- set ns_tr_out = namespace(flag=false) -%}\n            {%- if message.get('tool_responses') -%}\n                {#- Legacy: tool_responses embedded on the assistant message (Google/Gemma native) -#}\n                {%- for tool_response in message['tool_responses'] -%}\n                    {{- format_tool_response_block(tool_response['name'] | default('unknown'), tool_response['response']) -}}\n                    {%- set ns_tr_out.flag = true -%}\n                    {%- set ns.prev_message_type = 'tool_response' -%}\n                {%- endfor -%}\n            {%- elif message.get('tool_calls') -%}\n                {#- OpenAI Chat Completions: forward-scan consecutive role:tool messages -#}\n                {%- set ns_tool_scan = namespace(stopped=false) -%}\n                {%- for k in range(loop.index0 + 1, loop_messages | length) -%}\n                    {%- if ns_tool_scan.stopped -%}\n                    {%- elif loop_messages[k]['role'] != 'tool' -%}\n                        {%- set ns_tool_scan.stopped = true -%}\n                    {%- else -%}\n                        {%- set follow = loop_messages[k] -%}\n                        {#- Resolve tool_call_id to function name -#}\n                        {%- set ns_tname = namespace(name=follow.get('name') | default('unknown')) -%}\n                        {%- for tc in message['tool_calls'] -%}\n                            {%- if tc.get('id') == follow.get('tool_call_id') -%}\n                                {%- set ns_tname.name = tc['function']['name'] -%}\n                            {%- endif -%}\n                        {%- endfor -%}\n                        {#- Handle content as string or content-parts array -#}\n                        {%- set tool_body = follow.get('content') -%}\n                        {%- if tool_body is string -%}\n                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}\n                        {%- elif tool_body is sequence and tool_body is not string -%}\n                            {%- set ns_txt = namespace(s='') -%}\n                            {%- for part in tool_body -%}\n                                {%- if part.get('type') == 'text' -%}\n                                    {%- set ns_txt.s = ns_txt.s + (part.get('text') | default('')) -%}\n                                {%- endif -%}\n                            {%- endfor -%}\n                            {{- format_tool_response_block(ns_tname.name, ns_txt.s) -}}\n                        {%- else -%}\n                            {{- format_tool_response_block(ns_tname.name, tool_body) -}}\n                        {%- endif -%}\n                        {%- set ns_tr_out.flag = true -%}\n                        {%- set ns.prev_message_type = 'tool_response' -%}\n                    {%- endif -%}\n                {%- endfor -%}\n            {%- endif -%}\n\n            {%- set captured_content -%}\n            {%- if message['content'] is string -%}\n                {%- if role == 'model' -%}\n                    {{- strip_thinking(message['content']) -}}\n                {%- else -%}\n                    {{- message['content'] | trim -}}\n                {%- endif -%}\n            {%- elif message['content'] is sequence -%}\n                {%- for item in message['content'] -%}\n                    {%- if item['type'] == 'text' -%}\n                        {%- if role == 'model' -%}\n                            {{- strip_thinking(item['text']) -}}\n                        {%- else -%}\n                            {{- item['text'] | trim -}}\n                        {%- endif -%}\n                    {%- elif item['type'] == 'image' -%}\n                        {{- '<|image|>' -}}\n                        {%- set ns.prev_message_type = 'image' -%}\n                    {%- elif item['type'] == 'audio' -%}\n                        {{- '<|audio|>' -}}\n                        {%- set ns.prev_message_type = 'audio' -%}\n                    {%- elif item['type'] == 'video' -%}\n                        {{- '<|video|>' -}}\n                        {%- set ns.prev_message_type = 'video' -%}\n                    {%- endif -%}\n                {%- endfor -%}\n            {%- endif -%}\n            {%- endset -%}\n\n            {{- captured_content -}}\n            {%- set has_content = captured_content | trim | length > 0 -%}\n\n        {%- if ns.prev_message_type == 'tool_call' and not ns_tr_out.flag -%}\n            {{- '<|tool_response>' -}}\n        {%- elif not (ns_tr_out.flag and not has_content) -%}\n            {{- '<turn|>\\n' -}}\n        {%- endif -%}\n    {%- endif -%}\n{%- endfor -%}\n\n{%- if add_generation_prompt -%}\n    {%- if ns.prev_message_type != 'tool_response' and ns.prev_message_type != 'tool_call' -%}\n        {{- '<|turn>model\\n' -}}\n    {%- endif -%}\n{%- endif -%}"
+}

bf16/onnx-standard/vision_encoder/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2b1224078aae32a065b043107176d8a4de0c339b841e5944b850254697b395a
+size 1238953

bf16/onnx-standard/vision_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33fd860d1ce013444d2f16667e4625045971ad0ca905d1cb3b62feae2e937ba9
+size 358083464

f16/cuda/audio_encoder/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be721e5ebbdf37908cabf4b0e354761214fd44ab8fccbd07d45d98bbfacf930a
+size 1239093

f16/cuda/audio_encoder/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71ea13edb1d00a3c23fde4816433649bcbfdd245ddffb6769b491984f914a27f
+size 614724736