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

Enabled Models for TF-4.50.0 #340

Merged
merged 9 commits into from
Apr 8, 2025

Conversation

qcdipankar
Copy link
Contributor

@qcdipankar qcdipankar commented Apr 3, 2025

Enabled Models for TF-4.50.0

Models Enabled are

  1. GPT2
  2. GPTJ
  3. Granite
  4. Phi
  5. Phi3
  6. Whisper

This is the same PR that was raised as #334 which was closed due to rebasing issue

Code Cleaned and requested changes done and marked for review

QEff Dynamic Cache Added

@qcdipankar qcdipankar self-assigned this Apr 3, 2025
@qcdipankar qcdipankar marked this pull request as draft April 3, 2025 08:21
@quic-amitraj quic-amitraj changed the title Enabled Models for TF-4.50.50 Enabled Models for TF-4.50.0 Apr 3, 2025
@quic-amitraj
Copy link
Contributor

Please rebase

@qcdipankar qcdipankar force-pushed the transformer_migration branch 5 times, most recently from aac0964 to 9f5fcf2 Compare April 4, 2025 08:27
@qcdipankar qcdipankar marked this pull request as ready for review April 4, 2025 08:29
@qcdipankar qcdipankar added the enhancement New feature or request label Apr 4, 2025
@qcdipankar qcdipankar requested a review from vbaddi April 4, 2025 08:50
Copy link
Contributor

@quic-amitraj quic-amitraj left a comment

Choose a reason for hiding this comment

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

I still see the training code in the whisper, could you please once verify all the models?

@qcdipankar qcdipankar force-pushed the transformer_migration branch 2 times, most recently from 550b4ec to 04694ff Compare April 4, 2025 09:39
@quic quic deleted a comment from quic-amitraj Apr 4, 2025
@qcdipankar qcdipankar marked this pull request as draft April 4, 2025 11:46
@qcdipankar qcdipankar force-pushed the transformer_migration branch 2 times, most recently from 92b1c04 to a3ae6eb Compare April 4, 2025 12:14
@qcdipankar qcdipankar marked this pull request as ready for review April 4, 2025 14:54
@qcdipankar qcdipankar requested a review from quic-amitraj April 4, 2025 15:57
@qcdipankar
Copy link
Contributor Author

qcdipankar commented Apr 5, 2025

I still see the training code in the whisper, could you please once verify all the models?

All models full and one layer are verified I have kept logs for all and since whisper main code did not change at all for 4.50.0 so older code was retained with some changes but now I have removed the training parameters also.

@quic quic deleted a comment from quic-amitraj Apr 7, 2025
@quic quic deleted a comment from quic-amitraj Apr 7, 2025
@qcdipankar qcdipankar dismissed quic-amitraj’s stale review April 7, 2025 07:07

This has been attended and changes made.

@qcdipankar qcdipankar force-pushed the transformer_migration branch 2 times, most recently from d31c9ea to b79a9ea Compare April 7, 2025 15:53
Signed-off-by: Dipankar Sarkar <[email protected]>
Signed-off-by: Dipankar Sarkar <[email protected]>
Signed-off-by: Dipankar Sarkar <[email protected]>
In order to keep the same method for DynamicCache loading for all models we change the method for GPTJ also.

Signed-off-by: Dipankar Sarkar <[email protected]>
Signed-off-by: Dipankar Sarkar <[email protected]>
@qcdipankar qcdipankar force-pushed the transformer_migration branch from b79a9ea to dabbcf7 Compare April 7, 2025 15:53
Copy link
Contributor

@quic-amitraj quic-amitraj left a comment

Choose a reason for hiding this comment

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

LGTM

@quic-amitraj quic-amitraj merged commit 08bcf42 into quic:tf_upgrade_4.50 Apr 8, 2025
3 checks passed
quic-amitraj pushed a commit that referenced this pull request Apr 8, 2025
Enabled Models for TF-4.50.0

Models Enabled are

1. GPT2
2. GPTJ
3. Granite
4. Phi
5. Phi3
6. Whisper

This is the same PR that was raised as #334 which was closed due to
rebasing issue

Code Cleaned and requested changes done and marked for review

QEff Dynamic Cache Added

---------

Signed-off-by: Dipankar Sarkar <[email protected]>
Signed-off-by: Amit Raj <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
1.20.0 enhancement New feature or request
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants