bge-m3 / quantize_config.json
anpigon's picture
Upload 11 files
daa992f verified
{
"per_channel": true,
"reduce_range": true,
"per_model_config": {
"model": {
"op_types": [
"Reshape",
"Equal",
"Where",
"Expand",
"Shape",
"Pow",
"Mul",
"Gather",
"Softmax",
"MatMul",
"Sub",
"Transpose",
"Add",
"Div",
"CumSum",
"Unsqueeze",
"ReduceMean",
"Erf",
"Sqrt",
"Not",
"Concat",
"Cast",
"Constant",
"Slice",
"ConstantOfShape"
],
"weight_type": "QInt8"
}
}
}