Add professional Ternary Bonsai 1.7B MLC model card

README.md

@@ -0,0 +1,115 @@
+---
+license: apache-2.0
+base_model: prism-ml/Ternary-Bonsai-1.7B-unpacked
+library_name: mlc-llm
+pipeline_tag: text-generation
+tags:
+- mlc-llm
+- web-llm
+- webgpu
+- qwen3
+- bonsai
+- ternary
+- prismml
+- 2-bit
+- quantized
+- experimental
+---
+
+# Ternary-Bonsai-1.7B `bonsai_tq_f32` for MLC/WebLLM
+
+This repository contains an experimental MLC/WebLLM conversion of
+[`prism-ml/Ternary-Bonsai-1.7B-unpacked`](https://huggingface.co/prism-ml/Ternary-Bonsai-1.7B-unpacked).
+It is a browser-runtime artifact, not a new model, fine-tune, GGUF, MLX, or ONNX
+mirror.
+
+The source checkpoint is Prism ML's unpacked FP16 Ternary Bonsai model. This
+conversion uses a local MLC `bonsai_tq_f32` profile: symmetric 2-bit group
+quantization with `uint32` storage, group size 128, and FP32 scales. The encoded
+values represent the ternary lane `-scale`, `0`, and `+scale`.
+
+## Artifact Summary
+
+| Field | Value |
+| --- | --- |
+| Source checkpoint | `prism-ml/Ternary-Bonsai-1.7B-unpacked` |
+| Architecture | Qwen3-shaped decoder |
+| MLC model type | `qwen3` |
+| Quantization | `bonsai_tq_f32` |
+| Quantized storage | 2-bit symmetric group quantization in `uint32` |
+| Conversation template | `qwen3_nothink` |
+| Context window in config | `32768` |
+| Prefill chunk in config | `2048` |
+| Total parameters | 1,720,028,160 |
+| Quantized parameter size | 0.451 GB |
+| Bits per parameter | 2.252 |
+| Parameter shards | 13 |
+| Artifact size | about 460 MB |
+| WebGPU library | `libs/ternary-bonsai-1.7b-bonsai_tq_f32-webgpu.wasm` |
+
+## Runtime Requirement
+
+This artifact requires an MLC/WebLLM runtime with the local `bonsai_tq_f32`
+quantization profile registered. It is not expected to load in an unmodified
+upstream WebLLM build until this profile is upstreamed or otherwise carried in
+the runtime.
+
+This first ternary path uses MLC's group-quantized graph path. It is a compact
+WebGPU artifact and a correctness/release milestone, but it is not yet a custom
+fused ternary matmul kernel. Benchmark it before making speed claims.
+
+## WebLLM Configuration
+
+```js
+const appConfig = {
+  model_list: [
+    {
+      model: "https://huggingface.co/welcoma/Ternary-Bonsai-1.7B-bonsai_tq_f32-MLC/resolve/main/",
+      model_id: "Ternary-Bonsai-1.7B-tq-MLC",
+      model_lib:
+        "https://huggingface.co/welcoma/Ternary-Bonsai-1.7B-bonsai_tq_f32-MLC/resolve/main/libs/ternary-bonsai-1.7b-bonsai_tq_f32-webgpu.wasm",
+      overrides: {
+        context_window_size: 4096,
+        prefill_chunk_size: 512,
+      },
+    },
+  ],
+};
+```
+
+The smaller override values above are intended for local browser smoke tests.
+Increase them only after measuring browser memory and cache behavior on the
+target device.
+
+## Validation
+
+The artifact was converted and WebGPU-compiled on the GCP MLC/WebLLM builder VM,
+not on a local laptop.
+
+- Source: `prism-ml/Ternary-Bonsai-1.7B-unpacked`
+- Quantization: `bonsai_tq_f32`
+- Quantization profile: `int2` values, `uint32` packed storage, FP32 scales
+- Conversion peak RAM: 3.204 GB on CPU
+- WebGPU compile completed successfully
+- Compile estimate without KV cache: 1886.70 MB
+- Compile estimate with 4K KV cache: 2782.70 MB
+
+## Limitations
+
+- This is an experimental runtime artifact, not a general `transformers` model
+  checkpoint.
+- This repo does not claim the same runtime performance as Prism ML's native MLX
+  2-bit release.
+- Quality evaluation is limited to conversion and WebGPU compile checks; no
+  benchmark score is claimed by this repository.
+- Browser success depends on WebGPU support, available GPU memory, cache quota,
+  and a compatible patched WebLLM runtime.
+
+## Provenance
+
+Original model by Prism ML:
+
+- [prism-ml/Ternary-Bonsai-1.7B-unpacked](https://huggingface.co/prism-ml/Ternary-Bonsai-1.7B-unpacked)
+- [prismml.com](https://prismml.com/)
+
+MLC/WebLLM conversion by `welcoma`.