Skip to content

Support of Grok1 Model #373

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

Closed
wants to merge 14 commits into from
Closed

Support of Grok1 Model #373

wants to merge 14 commits into from

Conversation

quic-amitraj
Copy link
Contributor

Support of Grok1 Model

@quic-rishinr
Copy link
Contributor

quic-rishinr commented Jun 5, 2025

Add the mode to the validated list and please post the accuracy numbers as well

@quic-akuruvil
Copy link
Contributor

Add the mode to the validated list and please post the accuracy numbers as well

Also, grok1 2-layer model needs to be added to test files, and please paste with this PR the perplexity numbers in torch and AIC.

@quic-amitraj
Copy link
Contributor Author

Add the mode to the validated list and please post the accuracy numbers as well

Added.

@quic-amitraj
Copy link
Contributor Author

quic-amitraj commented Jun 6, 2025

Add the mode to the validated list and please post the accuracy numbers as well

Also, grok1 2-layer model needs to be added to test files, and please paste with this PR the perplexity numbers in torch and AIC.

All

image

Output matches end to end on Pytorch, ONNX and AI 100 for single layer.

Copy link
Contributor

@ochougul ochougul left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this tested via infer CLI command?

@quic-akuruvil
Copy link
Contributor

Add the mode to the validated list and please post the accuracy numbers as well

Also, grok1 2-layer model needs to be added to test files, and please paste with this PR the perplexity numbers in torch and AIC.

All

image

Output matches end to end on Pytorch, ONNX and AI 100 for single layer.

This is good. But Anuj has suggested to do a first order perplexity analysis from our side for text models, before we close a model, and pass on to accuracy team. Can check with Anuj and run the perplexity scripts if needed.

@quic-amitraj
Copy link
Contributor Author

Is this tested via infer CLI command?

It was not supported through infer as this model is not integrated with Transformers. With latest commit added support of this.

Copy link
Contributor

@quic-hemagnih quic-hemagnih left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good to me, we can plan to merge the changes.

@quic-hemagnih
Copy link
Contributor

Closing this Pull request as we have migrated these changes to another PR: #447

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

6 participants