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
Nightly integration tests #1664
Conversation
|
The documentation is not available anymore as the PR was closed or merged. |
| @@ -212,211 +213,215 @@ def test_stable_diffusion_img2img_num_images_per_prompt(self): | |||
|
|
|||
| @slow | |||
| @require_torch_gpu | |||
| class StableDiffusionImg2ImgPipelineIntegrationTests(unittest.TestCase): | |||
| class StableDiffusionImg2ImgPipelineSlowTests(unittest.TestCase): | |||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Img2Img is the most flaky, max_diff > 1e-4 even with 3 steps. If I won't find the cause today, will raise the tolerances
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sounds good 1e-3 is totally fine. FYI in transformers we only have 5e-3 for some one-forward pass tests
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
In an effort to speed up our
slow/pushtests, this PR:@nightlydecorator for tests that can be enabled withRUN_NIGHTLY=1num_inference_steps=3for all@slowtests@slowtests to@nightlynum_inference_steps~50for all@nightlytestsNightly tests will be run daily at midnight by our CI.