Qwen2.5-Coder-7B-Instruct-ONNX-WebGPU

Qwen/Qwen2.5-Coder-7B-Instruct converted to ONNX for running in a web browser with ONNX Runtime Web's WebGPU backend, as used by Quick Inference.

Changes from the original

• Converted to ONNX with ONNX Runtime GenAI's model builder (-p int4 -e webgpu).
• The weights of the linear layers are quantised to 4 bits (k-quant, block size 32); the rest are float16.
• Activations and the KV cache are float16.

These are modified files: their outputs are close to, but not identical to, the original model's.

Licence

Licensed under the same terms as the original model: see LICENSE.

Files

• LICENSE
• chat_template.jinja
• genai_config.json
• model.onnx
• model.onnx.data
• tokenizer.json
• tokenizer_config.json