 Huang Zhihong
|
a620831762
Fix the bug when continue the peft. (#717)
|
6 months ago |
 Kai Wu
|
9c7a5b421f
fix AutoModel and bump transformers version to 4.45 (#686)
|
6 months ago |
|
Kai Wu
|
57afa0b51e
use AutoModel
|
7 months ago |
|
Kai Wu
|
2730bcaab7
fix readme and fsdp logic
|
7 months ago |
|
Kai Wu
|
50dff0b78e
gradient_checkpointing_enable()
|
7 months ago |
|
Kai Wu
|
c18a0d277f
changed dataset to ocrvqa
|
7 months ago |
|
Kai Wu
|
bd22f407d5
changed to aid2 dataset
|
7 months ago |
|
Kai Wu
|
1a76080807
lora+fsdp working
|
7 months ago |
|
Kai Wu
|
8a11b48022
lora+fsdp not working
|
7 months ago |
|
Kai Wu
|
79dbe05a94
batch fine-tuning lmm working
|
7 months ago |
|
Kai Wu
|
ce299b3439
add get_custom_data_collator feature
|
7 months ago |
|
Kai Wu
|
12da109823
Merge branch 'main' into lmm_finetune
|
7 months ago |
|
Kai Wu
|
bb990be967
not working, need create dataloader function
|
7 months ago |
 Matthias Reso
|
778e31e35c
Fix checkpoint saving (#650)
|
7 months ago |
|
Kai Wu
|
ee204ccb98
working now
|
7 months ago |
|
Kai Wu
|
b566582a86
finetune not working with fsdp
|
7 months ago |
 Hamid Shojanazeri
|
808a3f7a0c
Adding support for FSDP+Qlora. (#572)
|
9 months ago |
 haozhx23
|
e6b0f97199
Fix hsdp_device_mesh=None when enable HSDP and HYBRID_SHARD (#402)
|
10 months ago |
|
Kai Wu
|
41a46d811d
fix alpaca dataset by using 5% of the data as eval and make sure len((eval_loader)>0
|
11 months ago |
|
Kai Wu
|
f1d90d0ff0
fix wandb config update
|
11 months ago |
|
Kai Wu
|
480c4f2b5e
resume the finetuning given the path of the previous peft checkpoint folder
|
11 months ago |
 Merovingian
|
b4e1a420c8
Freeze layer bug fix
|
11 months ago |
 Pavel Belevich
|
fb2e802cef
Fix param_init_fn: move if-statement out of lambda
|
11 months ago |
 Beto
|
65559c1640
Merge branch 'main' of github.com:albertodepaola/llama-recipes-private into main
|
1 year ago |
|
Matthias Reso
|
43cb6a2db4
Remove check for nighlies for low_cpu_fsdp and bump torch version to 2.2 instead
|
1 year ago |
|
Matthias Reso
|
113ea18bf1
Replace LlamaTokenizer with AutoTokenizer
|
1 year ago |
 Rahul A R
|
f8183b96fe
use new tokenizer_name argument and resize embeddings if required
|
1 year ago |
|
Hamid Shojanazeri
|
11f51db28c
adding the kbit prep in the code
|
1 year ago |
|
Hamid Shojanazeri
|
f058ff6ccd
update due to peft new release
|
1 year ago |
|
Hamid Shojanazeri
|
ffdc93f00a
Merge branch 'main' into wandb_logging
|
1 year ago |
