Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[bug] Failed to load pretrained model with huggingface transformers #1317

Open
kehuanfeng opened this issue Nov 6, 2024 · 1 comment · May be fixed by #1335
Open

[bug] Failed to load pretrained model with huggingface transformers #1317

kehuanfeng opened this issue Nov 6, 2024 · 1 comment · May be fixed by #1335
Labels
bug Something isn't working

Comments

@kehuanfeng
Copy link

kehuanfeng commented Nov 6, 2024 •
edited
Loading

torch 2.4.0a0+3bcc3cddb5.nv24.7
transformer-engine 1.11.0+c27ee60
transformers 4.45.0

[rank5]: Traceback (most recent call last):
[rank5]:   File "/data/kehuan/LLaMA-Factory/src/train.py", line 28, in <module>
[rank5]:     main()
[rank5]:   File "/data/kehuan/LLaMA-Factory/src/train.py", line 19, in main
[rank5]:     run_exp()
[rank5]:   File "/data/kehuan/LLaMA-Factory/src/llamafactory/train/tuner.py", line 50, in run_exp
[rank5]:     run_sft(model_args, data_args, training_args, finetuning_args, generating_args, callbacks)
[rank5]:   File "/data/kehuan/LLaMA-Factory/src/llamafactory/train/sft/workflow.py", line 48, in run_sft
[rank5]:     model = load_model(tokenizer, model_args, finetuning_args, training_args.do_train)
[rank5]:   File "/data/kehuan/LLaMA-Factory/src/llamafactory/model/loader.py", line 162, in load_model
[rank5]:     model = load_class.from_pretrained(**init_kwargs)
[rank5]:   File "/usr/local/lib/python3.10/dist-packages/transformers/models/auto/auto_factory.py", line 559, in from_pretrained
[rank5]:     return model_class.from_pretrained(
[rank5]:   File "/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py", line 4008, in from_pretrained
[rank5]:     ) = cls._load_pretrained_model(
[rank5]:   File "/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py", line 4272, in _load_pretrained_model
[rank5]:     if param.device == torch.device("meta"):
[rank5]: AttributeError: '_io.BytesIO' object has no attribute 'device'

I know it's due to the missing of _extra_state related to fp8, but have no idea how to fix this kind of issue?
@timmoon10 timmoon10 linked a pull request Nov 15, 2024 that will close this issue
Open
13 tasks
@timmoon10 timmoon10 added the bug Something isn't working label Nov 15, 2024
@timmoon10
Copy link
Collaborator

Can you try running with #1335?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
bug Something isn't working
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

[PyTorch] Store module extra state in tensor timmoon10/TransformerEngine
2 participants
@timmoon10 @kehuanfeng