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

Import punet tests from sharktank to iree_tests. #277

Merged
merged 8 commits into from
Jul 8, 2024

Conversation

ScottTodd
Copy link
Member

Related to nod-ai/SHARK-Platform#76

Only testing compilation and basic execution for now. Result comparisons will follow once they are enabled upstream. TBD if the quantized outputs will be different enough across backends to need extra infrastructure configuration there.

@ScottTodd ScottTodd requested a review from saienduri July 3, 2024 01:03
Copy link
Contributor

@stellaraccident stellaraccident left a comment

Choose a reason for hiding this comment

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

Thanks. Not terribly surprising that this is not compiling well. We just got it compiling (barely) on rocm on Sunday on a branch. I've been using CPU as a proxy. It'll get there.

@ScottTodd ScottTodd marked this pull request as ready for review July 3, 2024 19:35
@ScottTodd
Copy link
Member Author

This will have conflicts with #280, but otherwise it should be ready to land.

@ScottTodd ScottTodd merged commit 2c0a312 into nod-ai:main Jul 8, 2024
3 checks passed
@ScottTodd ScottTodd deleted the punet-ireetests branch July 8, 2024 18:56
renxida pushed a commit that referenced this pull request Jul 18, 2024
Related to nod-ai/SHARK-Platform#76

* Model source:
https://github.com/nod-ai/sharktank/tree/main/sharktank/sharktank/models/punet
* Python test:
https://github.com/nod-ai/sharktank/blob/main/sharktank/integration/models/punet/integration_test.py
* Python run script:
https://github.com/nod-ai/sharktank/blob/main/sharktank/sharktank/models/punet/tools/run_punet.py
* Weights are labeled in the sharktank sources. The primarily come from
https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0 and
https://huggingface.co/amd-shark/sdxl-quant-models

Only testing compilation and basic execution for now. Result comparisons
will follow once they are enabled upstream. TBD if the quantized outputs
will be different enough across backends to need extra infrastructure
configuration there.
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