gte-large / quantize_config.json
nanmoon's picture
gte-large
154a5ac
{
"per_channel": true,
"reduce_range": true,
"per_model_config": {
"model": {
"op_types": [
"Unsqueeze",
"Transpose",
"Softmax",
"Constant",
"Slice",
"Reshape",
"Add",
"ReduceMean",
"Pow",
"MatMul",
"Sub",
"Concat",
"Mul",
"Sqrt",
"Div",
"Cast",
"Shape",
"Erf",
"Gather"
],
"weight_type": "QInt8"
}
}
}