ms-marco-MiniLM-L-12-v2 / quantize_config.json
Xenova's picture
Xenova HF staff
Upload folder using huggingface_hub
a229a7c
raw
history blame contribute delete
722 Bytes
{
"per_channel": true,
"reduce_range": true,
"per_model_config": {
"model": {
"op_types": [
"Div",
"Slice",
"Tanh",
"Cast",
"Reshape",
"Erf",
"Shape",
"Gemm",
"Concat",
"Add",
"Gather",
"Pow",
"Unsqueeze",
"MatMul",
"Softmax",
"Constant",
"Sqrt",
"Mul",
"ReduceMean",
"Sub",
"Transpose"
],
"weight_type": "QInt8"
}
}
}