Skip to content

[Feature] add HFCheckpointHook to auto save hf model after the whole training phase #621

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
Merged
pppppM merged 6 commits into from
May 10, 2024
Merged

[Feature] add HFCheckpointHook to auto save hf model after the whole training phase #621

pppppM merged 6 commits into from
May 10, 2024

Conversation

@HIT-cwh
Copy link
Collaborator

@HIT-cwh HIT-cwh commented Apr 26, 2024 •
edited
Loading

Usage

+ from xtuner.engine.hooks import HFCheckpointHook

custom_hooks = [
+    dict(type=HFCheckpointHook),
]

The HF model will be saved at work_dir/timestamp/hf_model

Notes

  1. As the priority of HFCheckpointHook is 95 which is lower than CheckpointHook in MMEngine, the hf model is saved only after the normal checkpoint saving process is complete. So feel free to use it.
@pppppM pppppM merged commit fc41943 into InternLM:main May 10, 2024
@HIT-cwh HIT-cwh mentioned this pull request Jun 11, 2024
Closed
@HIT-cwh HIT-cwh deleted the pth_to_hf_hook branch July 26, 2024 11:48
llkn-2 pushed a commit to llkn-2/xtuner that referenced this pull request Jul 31, 2024
@HIT-cwh
[Feature] add HFCheckpointHook to auto save hf model after the whole …
37e730b 
…training phase (InternLM#621)

* add HFCheckpointHook to auto save hf model after the whole training phase

* refinie HFCheckpointHook

* fix lint

* delete useless codes

* fix bugs

* support non-dist training
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

2 participants

@HIT-cwh @pppppM