nlp-qual-q3i / quantize_config.json
kiranbandi's picture
Upload 8 files
98828b2 verified
{
"per_channel": true,
"reduce_range": true,
"per_model_config": {
"model": {
"op_types": [
"Gemm",
"Add",
"Pow",
"ReduceMean",
"Div",
"Reshape",
"Softmax",
"Concat",
"Constant",
"Sqrt",
"Mul",
"Gather",
"Unsqueeze",
"Erf",
"Slice",
"Transpose",
"MatMul",
"Sub",
"Cast",
"Shape",
"Tanh"
],
"weight_type": "QInt8"
}
}
}