fantastic

#1
by Yoshiii - opened

image.png

Looking forward the final checkpoint.

Looking forward the final checkpoint.

Looks like training is has been paused/stopped. Just a guess but maybe checkpoint-800 is as final as it'll get.

Looking forward the final checkpoint.

Looks like training is has been paused/stopped. Just a guess but maybe checkpoint-800 is as final as it'll get.

The latest checkpoint is 800/1200, I would continue training but it's simply too slow (60s/it) for some reason. If anybody has a solution to this problem let me know, I would be happy to let these 4xA100 continue working or even train larger models such as 30B with more gpus. More info: https://github.com/tatsu-lab/stanford_alpaca/issues/32

Got it fixed...

there it is, uploaded the full model

Sign up or log in to comment