Instructions to use Felladrin/onnx-Llama-68M-Chat-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers.js
How to use Felladrin/onnx-Llama-68M-Chat-v1 with Transformers.js:
// npm i @huggingface/transformers import { pipeline } from '@huggingface/transformers'; // Allocate pipeline const pipe = await pipeline('text-generation', 'Felladrin/onnx-Llama-68M-Chat-v1');
Download quantize_config.json from Felladrin/onnx-Llama-68M-Chat-v1: direct link, hf CLI and curl.
- Browser
- Download file 890 Bytes
-
https://huggingface.co/Felladrin/onnx-Llama-68M-Chat-v1/resolve/main/quantize_config.json
- Command line
-
hf download hf://Felladrin/onnx-Llama-68M-Chat-v1/quantize_config.json
-
curl -L -o quantize_config.json https://huggingface.co/Felladrin/onnx-Llama-68M-Chat-v1/resolve/main/quantize_config.json
890 Bytes
| { | |
| "per_channel": false, | |
| "reduce_range": false, | |
| "per_model_config": { | |
| "model": { | |
| "op_types": [ | |
| "Add", | |
| "Reshape", | |
| "Div", | |
| "Softmax", | |
| "Mul", | |
| "Sqrt", | |
| "Less", | |
| "Constant", | |
| "Where", | |
| "Pow", | |
| "ReduceMean", | |
| "Unsqueeze", | |
| "Expand", | |
| "MatMul", | |
| "Sub", | |
| "Range", | |
| "Shape", | |
| "Squeeze", | |
| "ConstantOfShape", | |
| "Transpose", | |
| "Neg", | |
| "Cast", | |
| "Sigmoid", | |
| "Concat", | |
| "Gather", | |
| "Equal", | |
| "Slice" | |
| ], | |
| "weight_type": "QInt8" | |
| } | |
| } | |
| } |