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

Update VAE MI3000-QPX spec file (no perf improvement) #125

Merged
merged 2 commits into from
Feb 24, 2025

Conversation

rkayaith
Copy link

This new spec file is tuned on the SDXL MLPERF IREE branch and has some slight differences, however I didn't see any actual perf improvement compared to the previous tuning when benchmarking.

This new spec file is tuned on the SDXL MLPERF IREE branch and has some
slight differences, however I didn't see any actual perf improvement
compared to the previous tuning when benchmarking.
@rkayaith rkayaith requested review from Max191 and monorimet February 24, 2025 18:07
@monorimet
Copy link
Collaborator

(just a thought: can these be consolidated into the fp8-model specs directory? We shouldn't need different specs for each model and batch size.)

@rkayaith
Copy link
Author

I added a symlink, since the scripts to compile/benchmark vae still reference the specs under int8-model/ currently

@rkayaith rkayaith merged commit 4aa0577 into shared/sdxl_on_main Feb 24, 2025
@rkayaith rkayaith deleted the rahul/sdxl_on_main-vae-tuning branch February 24, 2025 20:44
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