# April 2024


Plus explore the latest in Keras, improvements to TensorFlow Lite, on-device LLMs, and more.

|---|
| |---| | |---| | [![](https://gstatic.com/growthlab/api/5bVWsgT8ttKvlVaE5gRkjgAiBv9mwnyIQkgpcp0q.png)](https://www.tensorflow.org/subscribe/2024/04#) | | | |---| | |---| | |---| | Learn the latest at Google I/O | | | |---| | |---| | |---| | |---| | Google I/O returns on May 14, 2024! Tune in for keynotes, sessions and workshops covering the latest in machine learning and AI. | | | |---| | [Register now](https://io.google/2024/?utm_source=newsletter&utm_medium=email&utm_campaign=april_tf&utm_content) | | | | | | |---| |   | | | |---| | |---| | |---| | |---| | |---| | |---| | |---| | [![](https://services.google.com/fh/files/emails/people_of_ai_tf.png)](https://www.youtube.com/watch?v=oe6fuxhGVRE) | | People of AI interview with Francois Chollet, creator of Keras | | Listen to a discussion on deep learning, the evolution of Keras from creation to multi-backend Keras 3, and the intelligence of LLMs. | | |---| | |---| | |---| | |---|---| | |---| | [Watch episode](https://www.youtube.com/watch?v=oe6fuxhGVRE) | | |---| | [![](https://www.gstatic.com/growthlab/api/Obmfy5bg9fJotE3zdZDXu4HiZx1oT3KnprcfW3ED.png)](https://www.youtube.com/watch?v=oe6fuxhGVRE) | | | | | | | | | | | |---| | |---| | |---| | |---| | [![](https://services.google.com/fh/files/emails/xxn_pack_tf.png)](https://blog.tensorflow.org/2024/04/faster-dynamically-quantized-inference-with-xnnpack.html) | | Faster dynamically quantized inference with XNNPack | | Learn about the benefits of dynamic range quantization and how you can combine it with half-precision to improve on-device CPU inference performance. | | |---| | |---| | |---| | |---|---| | |---| | [Read the blog](https://blog.tensorflow.org/2024/04/faster-dynamically-quantized-inference-with-xnnpack.html) | | |---| | [![](https://www.gstatic.com/growthlab/api/Obmfy5bg9fJotE3zdZDXu4HiZx1oT3KnprcfW3ED.png)](https://blog.tensorflow.org/2024/04/faster-dynamically-quantized-inference-with-xnnpack.html) | | | | | | | | | | | |---| | |---| | |---| | |---| | [![](https://services.google.com/fh/files/emails/gemma_on_device_tf.png)](https://www.youtube.com/watch?v=LMXX6Q0DHnA) | | Gemma on-device with MediaPipe and TensorFlow Lite | | Watch a demo from [Gemma Developer Day](https://youtube.com/playlist?list=PLOU2XLYxmsIL8TxLOcsVI_hCq0c6cO-WQ&si=3w4JxMnyB6Ufl9-h) showing how to deploy Gemma models on-device using the MediaPipe LLM Inference API. | | |---| | |---| | |---| | |---|---| | |---| | [Watch video](https://www.youtube.com/watch?v=LMXX6Q0DHnA) | | |---| | [![](https://www.gstatic.com/growthlab/api/Obmfy5bg9fJotE3zdZDXu4HiZx1oT3KnprcfW3ED.png)](https://www.youtube.com/watch?v=LMXX6Q0DHnA) | | | | | | | | | | | |---| | |---| | |---| | |---| | [![](https://services.google.com/fh/files/emails/training_gradient_tf.png)](https://ydf.readthedocs.io/en/latest/tutorial/custom_loss/) | | Training gradient boosted trees with custom loss | | Learn how to define a custom loss function for TF-DF models. For robust features, a simpler API, and faster training times, you can [migrate](https://ydf.readthedocs.io/en/latest/tutorial/custom_loss/) to the YDF library. | | |---| | |---| | |---| | |---|---| | |---| | [View tutorial](https://ydf.readthedocs.io/en/latest/tutorial/custom_loss/) | | |---| | [![](https://www.gstatic.com/growthlab/api/Obmfy5bg9fJotE3zdZDXu4HiZx1oT3KnprcfW3ED.png)](https://ydf.readthedocs.io/en/latest/tutorial/custom_loss/) | | | | | | | | | | | | | | |---| |   | | | |---| | ![](https://www.gstatic.com/gumdrop/files/spacer-11.gif) | | Stay Connected | | |---|---|---|---|---|---|---|---|---| | [![TensorFlow blog](https://www.gstatic.com/growthlab/api/e7WEFvQxwbiFjhuJkQsF5Fq09BSs8t0t7zjVc2ZZ.png "TensorFlow blog")](http://blog.tensorflow.org/?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow_april24) |   | [![GitHub](https://www.gstatic.com/growthlab/api/3oLbQMNQ8xdPTrZaFXgbpypikCrz1cifbZnJ7hYv.png "GitHub")](https://github.com/tensorflow?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow_april24) |   | [![X](https://gstatic.com/growthlab/api/qTLa5CDPfCaU74g6OzQPE546CJPaqLheI2Vw8qev.png "X")](https://twitter.com/tensorflow?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow_april24) |   | [![YouTube](https://www.gstatic.com/growthlab/api/Hmw6zspJ6zRcW9dalh8m4Yztyv80TMDVsycNLpCi.png "YouTube")](https://www.youtube.com/tensorflow?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow_april24) |   | [![TensorFlow forum](https://www.gstatic.com/growthlab/api/o2mjGQiE1CeSDhViQ9I47utHTpyx929ckrYzUOoR.png "TensorFlow forum")](http://discuss.tensorflow.org/?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow_april24) | | | | | |

![](https://notifications.google.com/g/proof/img/ANiao5pRlv6M_3ZYckcoRuJ1b7IpZH9F3x5BPg4qVW0DNl3R7_FSu7Jcd9PfnVFLkf9_aiFyZNjmIUaPRnRJB5WrOIjFwQt10xVvDTMMv5jKYNbn49F9LV5k-rxTwoNslzI3_j_Zj8Ee4GYObIiZIaBxZT7ftQ.gif)