-
Notifications
You must be signed in to change notification settings - Fork 233
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
CI - Add kaggle creds to pull model #1459
CI - Add kaggle creds to pull model #1459
Conversation
We can skip these by default, for users who have not yet set them up. We will need to set them up for CI, see keras-team#1459
We can skip these by default, for users who have not yet set them up. We will need to set them up for CI, see #1459
2b4bf1d
to
d49329a
Compare
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.
lgtm!
def test_smallest_preset(self): | ||
# TODO: Fails with OOM on current GPU CI |
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.
Another option would be to take in a dtype
to run_preset_test
and pass is it as kwarg to from_preset
. Then we could try bfloat16 precision, which might avoid the oom?
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.
I tried keras.config.set_floatx("float16")
, but that didn't help. So going to land this PR now. We can address it in another PR
We can skip these by default, for users who have not yet set them up. We will need to set them up for CI, see keras-team#1459
* CI - Add kaggle creds to pull model * add kaggle env variables * Kaggle env: * Kaggle env: * Kaggle env: * Kaggle env: * Update Build script for Kokoro * Add Kaggle env var * set gemma preset to extra_large * Change Gemma small preset to bfloat16 * Change Gemma small preset to xlarge
We can skip these by default, for users who have not yet set them up. We will need to set them up for CI, see keras-team/keras-hub#1459
We can skip these by default, for users who have not yet set them up. We will need to set them up for CI, see keras-team/keras-hub#1459
keras_nlp/models/gemma/gemma_backbone_test.py::GemmaBackboneTest::test_smallest_preset
- Failing on OOM.Marked the test as
extra_large
so GPU CI passes.