onnx-gpt2-alpaca / quantize_config.json
Felladrin's picture
Upload folder using huggingface_hub
d15d881
raw
history blame contribute delete
813 Bytes
{
"per_channel": false,
"reduce_range": false,
"per_model_config": {
"model": {
"op_types": [
"Mul",
"Slice",
"Tanh",
"MatMul",
"Concat",
"Where",
"Gemm",
"Div",
"Transpose",
"Shape",
"Pow",
"Softmax",
"Unsqueeze",
"Sub",
"Add",
"ConstantOfShape",
"Reshape",
"Sqrt",
"ReduceMean",
"Squeeze",
"Constant",
"Cast",
"Gather",
"Split"
],
"weight_type": "QInt8"
}
}
}