You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Are there any training recipes or recommended configs for fine-tuning the parakeet-ctc-0.6b model? Specifically, I'm curious about how long it took to train on 1 hr of input audio.
I've been fine-tuning on an AWS p3.16xlarge instance, which has 8 V100 GPUs with 16 GB memory each. I run into persistent memory errors (torch.cuda.OutOfMemoryError) when training on more than 4 GPUs or with a batch size larger than 2. The quickest parameter set I've found takes 5 minutes to train on 10hr of data.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Are there any training recipes or recommended configs for fine-tuning the
parakeet-ctc-0.6bmodel? Specifically, I'm curious about how long it took to train on 1 hr of input audio.I've been fine-tuning on an AWS p3.16xlarge instance, which has 8 V100 GPUs with 16 GB memory each. I run into persistent memory errors (
torch.cuda.OutOfMemoryError) when training on more than 4 GPUs or with a batch size larger than 2. The quickest parameter set I've found takes 5 minutes to train on 10hr of data.Beta Was this translation helpful? Give feedback.
All reactions