# December 2023


Build with multimodal Gemini. Discover the latest in ML research, applications, and more.

|---|
| |---| | |---| | [![](https://gstatic.com/growthlab/api/5bVWsgT8ttKvlVaE5gRkjgAiBv9mwnyIQkgpcp0q.png)](https://www.tensorflow.org/subscribe/2023/12#) | | | |---| | |---| | |---| | Catch up on the latest in machine learning and AI | | | |---| | |---| | |---| | |---| | Explore sessions from the WiML Symposium 2023 to learn all the latest developer updates, including the future of open ML frameworks, generative AI tools, and more. | | | |---| | [Watch on YouTube](https://goo.gle/WiML23-all?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | | | | | |---| |   | | | |---| | |---| | |---| | |---| | |---| | |---| | |---| | [![](https://services.google.com/fh/files/emails/tf_building.png)](https://ai.google.dev/?utm_source=newsletter&utm_medium=email&utm_campaign=tf_dec) | | Build with the new Gemini multimodal model | | Grab your API key and try out Gemini's text and image reasoning capabilities with Google AI Studio, available in [38 languages across 180+ countries](https://ai.google.dev/available_regions/?utm_source=newsletter&utm_medium=email&utm_campaign=tf_dec&utm_content) . | | |---| | |---| | |---| | |---|---| | |---| | [Try Gemini](https://ai.google.dev/?utm_source=newsletter&utm_medium=email&utm_campaign=tf_dec) | | |---| | [![](https://www.gstatic.com/growthlab/api/Obmfy5bg9fJotE3zdZDXu4HiZx1oT3KnprcfW3ED.png)](https://ai.google.dev/?utm_source=newsletter&utm_medium=email&utm_campaign=tf_dec) | | | | | | | | | | | |---| | |---| | |---| | |---| | [![](https://services.google.com/fh/files/emails/tf_precision.jpg)](https://blog.tensorflow.org/2023/11/half-precision-inference-doubles-on-device-inference-performance.html?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | Half-precision inference with XNNPack doubles on-device inference performance | | Deploy AI-powered features to older and lower tier devices using floating-point inference in TensorFlow Lite's XNNPack backend for ARM CPUs. Explore the benchmarks on computer vision tasks. | | |---| | |---| | |---| | |---|---| | |---| | [Read the blog](https://blog.tensorflow.org/2023/11/half-precision-inference-doubles-on-device-inference-performance.html?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | |---| | [![](https://www.gstatic.com/growthlab/api/Obmfy5bg9fJotE3zdZDXu4HiZx1oT3KnprcfW3ED.png)](https://blog.tensorflow.org/2023/11/half-precision-inference-doubles-on-device-inference-performance.html?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | | | | | | | | | | |---| | |---| | |---| | |---| | [![](https://services.google.com/fh/files/emails/tf_training.jpg)](https://www.youtube.com/watch?v=jQDO20AzTM4?&utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | Training a TensorFlow Decision Forests model using gradient boosted trees | | Gradient boosting is one of the most powerful decision forest algorithms, often outperforming random forests. Start with a simple model and correct errors by adding new trees. | | |---| | |---|---|---| | |---| | |---|---| | |---| | [Watch the video](https://www.youtube.com/watch?v=jQDO20AzTM4?&utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | |---| | [![](https://www.gstatic.com/growthlab/api/Obmfy5bg9fJotE3zdZDXu4HiZx1oT3KnprcfW3ED.png)](https://www.youtube.com/watch?v=jQDO20AzTM4?&utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | | | ![](https://services.google.com/fh/files/emails/spacer_11.gif) | |---| | |---|---| | |---| | [Try the code in Colab](https://colab.sandbox.google.com/github/tensorflow/decision-forests/blob/main/documentation/tutorials/beginner_colab.ipynb?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | |---| | [![](https://www.gstatic.com/growthlab/api/Obmfy5bg9fJotE3zdZDXu4HiZx1oT3KnprcfW3ED.png)](https://colab.sandbox.google.com/github/tensorflow/decision-forests/blob/main/documentation/tutorials/beginner_colab.ipynb?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | | | | | | | | | | |---| | |---| | |---| | |---| | [![](https://services.google.com/fh/files/emails/tf_improving.jpg)](https://blog.research.google/2023/11/improving-simulations-of-clouds-and.html?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | Improving simulations of clouds with graph computation and XLA | | Learn how TPUs are enabling new computational experiments and substantially enlarging the sample of large eddy simulations (LES) to train parameterizations of clouds for global climate models. | | |---| | |---| | |---| | |---|---| | |---| | [Read the blog](https://blog.research.google/2023/11/improving-simulations-of-clouds-and.html?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | |---| | [![](https://www.gstatic.com/growthlab/api/Obmfy5bg9fJotE3zdZDXu4HiZx1oT3KnprcfW3ED.png)](https://blog.research.google/2023/11/improving-simulations-of-clouds-and.html?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | | | | | | | | | | |---| | |---| | |---| | |---| | [![](https://services.google.com/fh/files/emails/tf_enable.jpg)](https://blog.google/products/shopping/ar-beauty-shopping-ads/?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | Enabling AR tools on mobile for Google Shopping | | Check out AR Beauty Try On, a new feature in Google Search powered by MediaPipe, TensorFlow Lite, and WebGL infrastructure. Mobile users can now try on beauty products without the native Search app. | | |---| | |---| | |---| | |---|---| | |---| | [Read announcement](https://blog.google/products/shopping/ar-beauty-shopping-ads/?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | |---| | [![](https://www.gstatic.com/growthlab/api/Obmfy5bg9fJotE3zdZDXu4HiZx1oT3KnprcfW3ED.png)](https://blog.google/products/shopping/ar-beauty-shopping-ads/?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | | | | | | | | | | | | | |---| |   | | | |---| | ![](https://www.gstatic.com/gumdrop/files/spacer-11.gif) | | Stay Connected | | |---|---|---|---|---|---|---|---|---| | [![TensorFlow blog](https://www.gstatic.com/growthlab/api/e7WEFvQxwbiFjhuJkQsF5Fq09BSs8t0t7zjVc2ZZ.png "TensorFlow blog")](https://blog.tensorflow.org?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) |   | [![Github](https://www.gstatic.com/growthlab/api/3oLbQMNQ8xdPTrZaFXgbpypikCrz1cifbZnJ7hYv.png "Github")](https://github.com/tensorflow?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) |   | [![X](https://gstatic.com/growthlab/api/qTLa5CDPfCaU74g6OzQPE546CJPaqLheI2Vw8qev.png "X")](https://twitter.com/tensorflow?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) |   | [![Youtube](https://www.gstatic.com/growthlab/api/Hmw6zspJ6zRcW9dalh8m4Yztyv80TMDVsycNLpCi.png "Youtube")](https://www.youtube.com/tensorflow?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) |   | [![TensorFlow forum](https://www.gstatic.com/growthlab/api/o2mjGQiE1CeSDhViQ9I47utHTpyx929ckrYzUOoR.png "TensorFlow forum")](https://discuss.tensorflow.org?utm_source=email&utm_medium=newsletter&utm_campaign=tensorflow-dec23) | | | | | |

![](https://notifications.google.com/g/proof/img/ANiao5r3lcgkYR99slRmpQGVAd1vFx8xr5nah7IWMGO1XaMmpEgVuoRrtGstIl_LTpL7Dk_XWo42kaPgtiwUU6JY91t0fM4hE5w4OcypgK2RNyJrs-A.gif)