News for hackers. Stuff that matters.

A comment on Show HN: Web demo of 13B Alpaca-LLaMA trained on improved Stanford dataset. Parent comment

Re:Show HN: Web demo of 13B Alpaca-LLaMA trained on improved... by sockeye on Thursday March 23, 2023 @02:54PM (#35279580)
Thanks!

- Training code is https://github.com/tatsu-lab/stanford_alpaca

- Params were mostly default (as in stanford_alpaca README), except for: per_device_train_batch_size=1, per_device_eval_batch_size=1

- Fine-tuning dataset was based on https://github.com/tloen/alpaca-lora/raw/81eb72f707b0505a03b... with minor improvements; I'm going to publish my version soon

- The training itself took about 3 hours on 8x Nvidia A100 80GB

[ Parent | on HN ]
Re:Show HN: Web demo of 13B Alpaca-LLaMA trained on improved... by musabg on Thursday March 23, 2023 @03:19PM (#35279941)
Where should We follow you so we immediately know when you published that?
[ Parent | on HN ]
Re:Show HN: Web demo of 13B Alpaca-LLaMA trained on improved... by james7777 on Thursday March 23, 2023 @09:21PM (#35283876)
if the 7b model took 3 hours with the same hardware and codes and parameters, how can the 13b model be done in 3 hours?
[ Parent | on HN ]
Re:Show HN: Web demo of 13B Alpaca-LLaMA trained on improved... by cglan on Friday March 24, 2023 @12:21PM (#35291347)
will you be publishing the trained 13b model?
[ Parent | on HN ]