Training state-of-the-art large language models (LLMs) with billions of parameters requires distributed training across hundreds or thousands of GPUs. At this scale, hardware failures are not exceptional events—they are expected....
TL;DR PyTorch tests are often generated at import time, so CI failures may show device/dtype-specific names that differ from the source template. For local debugging, pytest -k and test/run_test.py are...
This past weekend in San Francisco, builders, researchers, mobile developers, and AI practitioners came together for the ExecuTorch Hackathon, a two-day, on-site event focused on a practical and increasingly important...
TL;DR Miles is RadixArk’s open source framework for large-scale LLM RL post-training. It composes SGLang for rollout, NVIDIA Megatron-LM for training, Ray orchestration, and PyTorch-native extensibility behind a small, pluggable...
TL;DR PyTorch now has a Cross-Repository CI Relay (CRCR) that automatically triggers and tracks CI in downstream repositories whenever a PR is opened or a commit is pushed against pytorch/pytorch....
TL;DR Eighty engineers, researchers, and community builders gathered for the inaugural PyTorch Meetup Singapore. Hosted at the Red Hat Asia Pacific office and organised by Sudhir Dharanendraiah, Ayush Satyam, Sumantro...
TL;DR: This case study demonstrates how LinkedIn re-architected its distributed linear programming solver, DuaLip, by developing a GPU-accelerated PyTorch version to handle extreme-scale optimization challenges like web applications. This transition...
TL;DR: The TokenSpeed inference engine achieved a record-breaking 580 tps running the Qwen3.5-397B-A17B model on GPUs. This extreme performance for agentic workloads is driven by systematic elimination of memory copies,...
A little over a year ago, the PyTorch Foundation launched the Ambassador Program, an initiative that recognizes and supports independent, trusted voices in the PyTorch community who are passionate about...
Thank you to everyone who participated in the PyTorch Docathon 2026! Once again, the community showed up with incredible energy and dedication to make PyTorch documentation better for developers everywhere....
TL;DR: Introducing the ExecuTorch MLX Delegate The new MLX delegate enables optimized, GPU-accelerated inference for PyTorch models on Apple Silicon Macs, using Apple’s MLX framework. The delegate seamlessly integrates with...
We are excited to announce the release of PyTorch® 2.12 (release notes)! The PyTorch 2.12 release features the following changes: Batched linalg.eigh on CUDA is up to 100x faster due...
TL;DR: ExecuTorch extends the PyTorch ecosystem to deliver local AI inference on constrained edge devices. To provide a practical entry point, Arm has created a set of Jupyter Labs that...
TL;DR: Traditional RecSys inference explicitly replicates shared user embeddings/sequences for every candidate. In-Kernel Broadcast Optimization (IKBO) eliminates this overhead via a kernel-model-system co-design that fuses broadcast logic directly into user-candidate...
