search

huggingface / optimum-quanto

A pytorch quantization backend for optimum
Apache License 2.0
645 stars 36 forks source link

Should we stop using quanto without the optimum? #215

Open kechan opened 1 week ago

kechan commented 1 week ago

Is this going to replace this?

from quanto import Calibration, freeze, qfloat8, qint4, qint8, quantize

and from this point, we should

from optimum.quanto import Calibration, freeze, qfloat8, qint4, qint8, quantize

Is the "old" quanto not going to be supported?

dacorvo commented 1 week ago

@kechan you're right, the "old" quanto is not going to be supported anymore. All future releases will be optimum-quanto.