Unable to do QAT via nvidia h100 gpu

quic / aimet

AIMET is a library that provides advanced quantization and compression techniques for trained neural network models.

https://quic.github.io/aimet-pages/index.html

Other

2k stars 361 forks source link

Unable to do QAT via nvidia h100 gpu #2654

Open hcqylymzc opened 5 months ago

hcqylymzc commented 5 months ago

Hi, Since h100 requires pytorch2 and cuda118, I cannot complete aimet's QAT through h100. Is there any plan or time for releasing the Pytorch2_cuda118 version of aimet? Thank you very much.

quic-hitameht commented 5 months ago

Hi @hcqylymzc

Thank you for reaching out. Currently, we are in the process of upgrading PyTorch version to 2.x but don't have an ETA yet. We will update you shortly.