We’re advancing on-device AI with ExecuTorch, now deployed across devices including Meta Quest 3, Ray-Ban Meta, Oakley Meta Vanguard and Meta Ray-Ban Display. By eliminating conversion steps and supporting pre-deployment validation in PyTorch, ExecuTorch accelerates the path

🔥The future of AI apps is on the device. With react-native-executorch (from @swmansion), you can run models locally in your Expo app: ♢ Private by design ♢ Zero API fees ♢ Works offline ♢ Instant response Read how it works in the blog below from @_darthez_ and

🚀 NVIDIA and Meta have joined forces to bring effortless Federated Learning to mobile devices with NVIDIA FLARE & ExecuTorch! This collaboration empowers developers to leverage cross-device learning while ensuring user privacy and data security. 🌐🔒 With a robust hierarchical

Feels good to be highlighted by the ExecuTorch team as a success story as they release v1.0 ExecuTorch community has been amazing to work with and vastly accelerated working with newer models while giving amazing performance! Look forward to exciting times for on-device AI and

New from @Arm + PyTorch 👇 Train smarter for the edge: • Hardware–software co-design • Extreme + QAT quantization • Mixture-of-Experts for lower latency Runs even faster on Arm with KleidiAI + ExecuTorch. 👉 Blog + 10 hrs of notebooks: https://t.co/cBktKnMv3n #PyTorch

https://t.co/XIkqmYLv32 @qualcomm engineers Derrick Johnson and Kartikey Rawat going through running LLMs on-device with Qualcomm Hardware with ExecuTorch

I’ve been following what the Unsloth team is doing, and their latest collab with Meta’s ExecuTorch is actually pretty impressive. They’ve basically bridged the gap between fine-tuning a model and getting it to run smoothly on a phone. It’s not just "it works" it’s actually

We just hit 1000 stars on react-native-executorch! When we started this project nearly a year ago, I didn't imagine we'd get here so fast. Grateful for every contributor, issue reporter, and user who's helped shape this. The best is yet to come.

Executorch: On-device AI across mobile, embedded and edge for PyTorch https://t.co/FbsQ6IGS75

Run LLMs on-device using @expo react native and #ExecuTorch 🔥 with this #oss library by @swmansion! Demo & code 👇

On-device AI is here. @RunAnywhereAI lets teams ship models to any edge device, at scale. We’ve just added @Microsoft ONNX support to the SDK (alongside llamacpp). ExecuTorch support next 👀 @Meta :

@rcanand @UnslothAI @PyTorch Oh technically yes - we made a notebook for it as well, so you just need to change the loading part with `qat_scheme = "phone-deployment" and save it via Executorch -

Part 2 of our tutorial where we build an AI-powered note-taking app using React Native ExecuTorch and React Native RAG is out. 🧠 In this part, we add image semantic search, all locally!

This Week In React 231 Special edition with msg from the React team 😀 https://t.co/ll5BbvYLZU ⚛️ - React Labs - React Compiler - React Router - TanStack Query - Next.js - c15t - RTK - Base UI 📱 - Legend List - FlashList - Versioning Policy - Brownfield - Metro - ExecuTorch

WWDC24: 在Apple GPU上训练您的机器学习和AI模型 | Apple 了解如何使用Metal在Apple Silicon上训练支持PyTorch、JAX和TensorFlow的模型。通过新的注意力操作和量化支持，提升设备上Transformer模型的性能。 00:00 - 介绍 01:36 - 在Apple silicon上训练框架 04:16 - PyTorch改进 11:26 - ExecuTorch

We recently helped with enabling OpenAI’s Whisper support in @huggingface Optimum ExecuTorch 🎉 Optimum ExecuTorch is a bridge between 🤗 Transformers and ExecuTorch — making it easy to run AI models on-device. Check out how you can run exported models in React Native ⚛️ 🧵

New low-bit embedding and linear operators for Arm CPUs are now in TorchAO, PyTorch’s low-precision library! Compatible with eager, torch.compile, AOTI, and ExecuTorch, they’re built for flexibility and ready for #torchchat. 👋Learn more: https://t.co/YkTKm2mrz3 #pytorch

2x faster LLM inference coming to ExecuTorch soon. Went from 25 tok/s to 62 tok/s on the same prompts. Patch for react-native-executorch drops in a few days 👀

Last week, I created a POC for work that involved an LLM with executorch + Embedding + Vector Database, all running locally on an old Android! Crazy times.