multilingual-embedding-base / onnx /quantization_check.json
Horizon-Labs's picture
Multilingual embedding base v1.0 (bge-m3 compatible)
1b929be verified
Raw History Blame Contribute Delete
158 Bytes
{
"gather_only": {
"min_cos": 0.9999687671661377,
"mb": 641.822905
},
"fp32": {
"cos_to_torch": 0.9999997615814209,
"mb": 1231.445467
},
"n": 4
}