Keras 3.0 is here.
Keras 3.0 - Install & Setup
More details: Getting started with Keras
What about Keras 2?
- Now available as a standalone package, tf_keras
- pip install tf_keras
- import tf_keras as keras
- Starting with TF 2.15, if Keras 3 is installed, tf.keras will be Keras 3
- Note that TF 2.15 installs Keras 2.15 by default
- But it’s also compatible with Keras 3 if it’s installed
- TF 2.16+ will install Keras 3
- Want to keep tf.keras pointing to Keras 2 after TF 2.16? You can!
- Install tf_keras
- Set TF_USE_LEGACY_KERAS=1
Kaggle + Keras
- Work is underway to integrate KerasNLP & KerasCV with Kaggle Models.
- Here’s a sneak peak!
- More details soon…
Come join us and make an impact!
- Call for contributions
- Example: develop a CTC API.
- Contribute to KerasCV
- Contribute to KerasNLP