gte-small / quantize_config.json
ggrn's picture
chore: initial commit
2c0d57d
raw
history blame
674 Bytes
{
"per_channel": true,
"reduce_range": true,
"per_model_config": {
"model": {
"op_types": [
"Erf",
"Shape",
"Sqrt",
"Constant",
"Concat",
"Mul",
"MatMul",
"Softmax",
"Slice",
"Add",
"ReduceMean",
"Cast",
"Unsqueeze",
"Gather",
"Sub",
"Div",
"Transpose",
"Reshape",
"Pow"
],
"weight_type": "QInt8"
}
}
}