Inference

Open HaiFengZeng opened this issue 1 year ago • 2 comments

how to do inference?

Oct 29 '24 07:10 HaiFengZeng

Did you post the final models?

Nov 04 '24 02:11 josephflowers-ra

I did a straight run of this repo, unmodified (except for the obvious run parameters), on the cloud using one L40S. Results with the checkpoints/log/final model/inference code/online demo is posted at huggingface here: https://huggingface.co/lemonteaa/nanogpt-speedrun

(No claim that the codes are bug free though as I just ask AI to code it up quickly with some manual coding on the non core part)

Nov 09 '24 08:11 lemonteaa

I think gpt-fast will do a lot of help

Dec 12 '24 06:12 HaiFengZeng