| license: mit | |
| library_name: onnxruntime | |
| tags: | |
| - onnx | |
| - qwen3 | |
| - browser-inference | |
| - distributed-inference | |
| # WebAI@Home Qwen3-0.6B ONNX shards | |
| This repository contains the three ONNX shards used by the [WebAI@Home Worker](https://github.com/webai-at-home/worker) to run the Qwen3-0.6B language model across browser workers. | |
| ## Files | |
| - `shards/shard-1.onnx` contains the token embedding and the first group of decoder layers. | |
| - `shards/shard-2.onnx` contains the middle group of decoder layers. | |
| - `shards/shard-3.onnx` contains the final group of decoder layers and the language-model output head. | |
| The three files are static model assets. A Worker browser downloads only the shard assigned to its stage and stores the downloaded bytes in its browser cache for later use. | |
| ## Source and version | |
| The shards were split from the Qwen3-0.6B ONNX model published as [`onnx-community/Qwen3-0.6B-ONNX`](https://huggingface.co/onnx-community/Qwen3-0.6B-ONNX). The WebAI@Home Worker uses this repository by immutable revision, not by the moving `main` branch. | |
| This repository is part of the [WebAI@Home](https://github.com/webai-at-home/webai-at-home) project. It is intended to be downloaded by browser code through the Hugging Face Hub and does not contain executable application code. | |