eycheung
eycheung
cc: @symphonylyh can you please help me take a look at this MR for the first part of T5 quantization?
thank you so much @Eddie-Wang1120! this indeed fixed the build issue I was having and now the output looks correct. This has unblocked me for the next steps now.
Great, thank you @symphonylyh ! And I appreciate the update and work on follow-up 1. And thank you @Eddie-Wang1120 for the help on this. To summarize the issue, before the...
> I'm trying to run this PR, but I'm getting an assertion error when running the `build.py` script even without weight-only quantization (with FP32 too). Couldn't figure out the reason....
Great, thank you @monatis ! Sorry i could not help with this specific issue though.