 Hamid Shojanazeri
|
b2a55022cb
clean up
|
2 lat temu |
|
Hamid Shojanazeri
|
0941a9ecf7
clean up
|
2 lat temu |
|
Hamid Shojanazeri
|
8601ae14ea
extend word list
|
2 lat temu |
|
Hamid Shojanazeri
|
44ef280d31
adding flash attention and xformer memory efficient through PT SDPA
|
2 lat temu |
 Geeta Chauhan
|
dd57dc5cf9
Update spellcheck.sh (#88)
|
2 lat temu |
 sekyonda
|
59b59afe88
Update spellcheck.sh
|
2 lat temu |
|
Geeta Chauhan
|
1387b76e11
fixing the full state path in checkpoint handler+loss report calculation (#51)
|
2 lat temu |
|
Hamid Shojanazeri
|
88d3e1febc
fix the save_train_param condition
|
2 lat temu |
|
Hamid Shojanazeri
|
b56028c98d
fixing the word list/spell check
|
2 lat temu |
|
Hamid Shojanazeri
|
62be60355a
resolving conflicts
|
2 lat temu |
|
Geeta Chauhan
|
174b856591
update README: python 3.9 rec + fix formatting (#63)
|
2 lat temu |
|
Geeta Chauhan
|
0cd5694a14
Fsdp inference checkpoints (#39)
|
2 lat temu |
|
Hamid Shojanazeri
|
c4e96af6ee
clean up
|
2 lat temu |
 Christian Miller
|
7c1884c690
recommend python 3.9
|
2 lat temu |
|
Hamid Shojanazeri
|
7d2e06821e
fixing the path to script
|
2 lat temu |
|
Hamid Shojanazeri
|
5f97db8f0c
fix spell check word list
|
2 lat temu |
|
Hamid Shojanazeri
|
017cadd04b
Merge branch 'checkpoint_handler_path_fix' of https://github.com/facebookresearch/llama-recipes into checkpoint_handler_path_fix
|
2 lat temu |
|
Hamid Shojanazeri
|
4f70348b94
remove the redundant lr step
|
2 lat temu |
|
Hamid Shojanazeri
|
9c95ed4bbe
clean up
|
2 lat temu |
|
Hamid Shojanazeri
|
311a5c1eec
add notes for train_param.yaml
|
2 lat temu |
|
Hamid Shojanazeri
|
5b916114eb
merge main branch
|
2 lat temu |
|
Hamid Shojanazeri
|
668c364f6b
add rank to save_train_params
|
2 lat temu |
|
Hamid Shojanazeri
|
231c9e7da9
adding train_param.yaml saving for fsdp checkpoint loading for inference
|
2 lat temu |
|
Hamid Shojanazeri
|
475e67b4ec
clean up
|
2 lat temu |
|
Hamid Shojanazeri
|
50e9d17045
add the default option for find the HF model_name/path from train_param.yaml
|
2 lat temu |
|
Hamid Shojanazeri
|
41dd7ff1cb
Merge branch 'main' into checkpoint_handler_path_fix
|
2 lat temu |
|
Hamid Shojanazeri
|
31d6ce8bf6
adding expnadable sgement and dist debug flag info
|
2 lat temu |
|
Hamid Shojanazeri
|
a955ed1999
added checks for dist barrier and commented cuda exapnadable segements and dist_dbug
|
2 lat temu |
|
Hamid Shojanazeri
|
a2403c7c1a
clean up
|
2 lat temu |
|
Hamid Shojanazeri
|
e9559d2669
fixing the train/eval_loss calcualtion
|
2 lat temu |
