Keras Core now has support for sparse tensors (with the TF backend only for now — we'll add JAX next). This makes it possible to use Keras Core to build recommender systems.
@fchollet
-
Symbolic Plugin Revealed as Code Interpreter Plugin
By
–
Following up on this: surprise surprise, it was a symbolic plugin (a code interpreter plugin).
-
Train KerasNLP Models Faster with Data Parallel Training
By
–
Learn how to train KerasNLP models faster with data parallel training: https://
keras.io/examples/nlp/d
ata_parallel_training_with_keras_nlp/
… -
AI Generated Images Enough to Pitch New City Development
By
–
Those images look extremely AI generated. Maybe this is all you need to pitch "let's build a new city" at this moment in time
-
Automated PyTorch Module Integration with Keras Layers
By
–
By popular demand, we've automated the "turn the Module into a Keras Layer" step. So now you can just… use torch Modules in a Keras model/layer. No extra step needed. Just like you can use Keras models/layers as Modules.
-
BART Abstractive Text Summarization with KerasNLP and Keras Core
By
–
Abstractive text summarization with BART — using KerasNLP and Keras Core, with JAX
-
Keras Core Models Compatible with PyTorch Module Objects
By
–
Inversely, Keras Core models behave like torch `Module` objects and can be used anywhere you'd use a `Module`.
-
Keras Core now supports arbitrary PyTorch Modules with TorchModuleWrapper
By
–
Now, it's possible to use arbitrary `Modules` inside a Keras Core model/layer (with the torch backend). The `TorchModuleWrapper` class turns them into Keras layers, keeping tracking of trainable weights / etc.
-
Efficient object detection with YOLOv8 and KerasCV
By
–
Efficient object detection with YOLOv8 and KerasCV:
-
LLMs More Intelligent Than Self-Driving Cars Despite Different Embodiment
By
–
SDCs are much more like embodied autonomous agents, but I would say that LLMs are actually more intelligent.