nanoGPT icon indicating copy to clipboard operation
nanoGPT copied to clipboard

checkpoints don't seem to be working

Open eniompw opened this issue 2 years ago • 2 comments

I'm testing out train.py on google colab but no checkpoints are created, even after iter 1000 .

I'm using this command:

!cd /content/nanoGPT/ && python train.py --dataset=shakespeare --compile=False --n_layer=4 --n_head=4 --n_embd=64 --eval_iters=20 --block_size=64 --batch_size=8 --init_from=gpt2 --always_save_checkpoint=True --dtype=float32

eniompw avatar Jan 17 '23 13:01 eniompw

The eval_interval is 2000 by default, maybe then?

karpathy avatar Jan 17 '23 15:01 karpathy

just tested it, seems to be 2000 thanks :) for ref its line 36: eval_interval = 2000

If anyone is curious, I've done a small writeup of using Google Colab to finetune nanoGPT on shakespeare. GitHub Repo Link

eniompw avatar Jan 17 '23 16:01 eniompw