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.

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

[Bug_Fix]Modify the config of args.deepspeed_config to args.deepspeed_config_dict #328

Merged
merged 1 commit into from
Jan 10, 2024
Merged

Conversation

ys950902
Copy link

This pr is to fix the potential bug on init, according to the method _create_ds_config_dict() https://github.com/microsoft/Megatron-DeepSpeed/blob/main/megatron/training.py#L148, the flag args.deepspeed_config is set to None and the deepspeed_config is set to the args.deepspeed_config_dict, so in init the config_dict_or_path=None.
When we change to MiCS_Init it will cause error because the deepspeed_config is None, should set to args.deepspeed_config_dict.

@tjruwase tjruwase merged commit a4f8079 into microsoft:main Jan 10, 2024
1 check passed
@ys950902 ys950902 deleted the fix_config branch January 11, 2024 02:18
zdaiot pushed a commit to zdaiot/Megatron-DeepSpeed that referenced this pull request Jan 20, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants