Skip to content

Conversation

@chai-xiaonan
Copy link
Contributor

The checkpoint conversion tool must be updated accordingly following the Megatron update.
Error occurred when converting the trained torch-format ckpt to HF safetensor format:
1)The parameter args.expert_tensor_parallel_sizecannot be retrieved.
2)The model_provideris missing the model_builderparameter.
Fixed the two issues mentioned above.

@CLAassistant
Copy link

CLA assistant check
Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
You have signed the CLA already but the status is still pending? Let us recheck it.

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