Blake
Blake
The model weights didn't use nn.linear so its not an out of the box solution. There are models out there that have been converted that work right away. I was...
Might this be a race condition between the two GPUs?
It works with the change you made? Perhaps an argument to pass would be allowing one to set the device index
Glad you found my code helpful and added it here. The whisper model takes either a wav file or an array(not sure of the format). However, I could not get...
Good idea to add. Could add a ```listen_to_file``` method
@arschoon yes I take PRs
I am having this issue as well. I tried 4.26 and 4.25.1. I am gonna try @sam-ulrich1 solution.
The fix did not help. Neither using DeepSpeed nor using vanilla Transformers. Using bfloat16 gives me expected results(but I need float16 for DeepSpeed)
I take back what I said. I am not having this issue at all. With or withou t @sam-ulrich1 fix, it is working fine. The issue is with DeepSpeed.
It is very hard to get this working on a Mac. I had to build some parent software from the source. I forget exactly what I did to get it...