Skip to content

Keras <> NNX integration #21252

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

Open
wants to merge 102 commits into
base: master
Choose a base branch
from

Conversation

divyashreepathihalli
Copy link
Collaborator

@divyashreepathihalli divyashreepathihalli commented May 5, 2025

The PR integrates NNX into JAX backend!

The following snippet shows how you would enable the nnx backend

import os
os.environ["KERAS_BACKEND"]="jax"
os.environ["KERAS_NNX_ENABLED"]="true"
import keras

Demo colab here : https://colab.sandbox.google.com/drive/1mK-4qbce2HGRIkcb4v5n4niWGDezL_6n#scrollTo=m-ZH9Mpnphfz
Added a github workflow action for nnx backend. Note this will fail - because this needs a new release of flax to work.

@divyashreepathihalli divyashreepathihalli marked this pull request as draft May 5, 2025 23:05
@codecov-commenter
Copy link

codecov-commenter commented May 5, 2025

Codecov Report

Attention: Patch coverage is 19.52663% with 136 lines in your changes missing coverage. Please review.

Project coverage is 61.78%. Comparing base (d55a767) to head (2214458).
Report is 4 commits behind head on master.

Files with missing lines Patch % Lines
keras/src/backend/jax/core.py 6.49% 71 Missing and 1 partial ⚠️
keras/src/backend/config.py 32.35% 18 Missing and 5 partials ⚠️
keras/src/ops/function.py 41.17% 8 Missing and 2 partials ⚠️
keras/src/backend/jax/trainer.py 0.00% 9 Missing ⚠️
keras/src/layers/layer.py 35.71% 7 Missing and 2 partials ⚠️
keras/src/backend/jax/layer.py 0.00% 5 Missing ⚠️
keras/src/backend/common/variables.py 0.00% 2 Missing and 1 partial ⚠️
keras/src/ops/operation.py 25.00% 2 Missing and 1 partial ⚠️
keras/api/_tf_keras/keras/config/__init__.py 0.00% 1 Missing ⚠️
keras/src/backend/__init__.py 0.00% 1 Missing ⚠️

❗ There is a different number of reports uploaded between BASE (d55a767) and HEAD (2214458). Click for more details.

HEAD has 6 uploads less than BASE
Flag BASE (d55a767) HEAD (2214458)
keras 5 2
keras-torch 1 0
keras-tensorflow 1 0
keras-jax 1 0
Additional details and impacted files
@@             Coverage Diff             @@
##           master   #21252       +/-   ##
===========================================
- Coverage   82.81%   61.78%   -21.03%     
===========================================
  Files         565      565               
  Lines       55520    55803      +283     
  Branches     8664     8716       +52     
===========================================
- Hits        45977    34476    -11501     
- Misses       7428    19190    +11762     
- Partials     2115     2137       +22     
Flag Coverage Δ
keras 61.77% <19.52%> (-20.85%) ⬇️
keras-jax ?
keras-numpy 58.48% <19.52%> (-0.11%) ⬇️
keras-openvino 33.91% <18.34%> (-0.08%) ⬇️
keras-tensorflow ?
keras-torch ?

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

import jax.numpy as jnp

x = ops.ones(3)

@jax.jit
@nnx.jit
Copy link
Collaborator

Choose a reason for hiding this comment

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

Would the integration prevent the use of jax.jit with Keras layers?

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

yes! it would only work with nnx.jit for now ( They might be working on adding support for jax.jit)

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

Added nnx as a opt in with this flag - os.environ["KERAS_NNX_ENABLED"]

@divyashreepathihalli
Copy link
Collaborator Author

divyashreepathihalli commented Jul 14, 2025

FLAX installation is pointing to my branch currently until the CL and PR changes makes it through the Flax Github

model(x)
model.predict(x)
Copy link
Member

Choose a reason for hiding this comment

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

why do we need this change? this seems to indicate a bug perhaps

Copy link
Collaborator

Choose a reason for hiding this comment

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

Looks like a model building bug? predict() should work out of the box because the model should auto-build using the data as reference

set_floatx(_floatx)
set_epsilon(_epsilon)
set_image_data_format(_image_data_format)
_BACKEND = _backend

# Save config file, if possible.
Copy link
Collaborator

Choose a reason for hiding this comment

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

Why did you need to move this block?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

8 participants