Meta PyTorch ORG

Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

PyTorch native post-training library

Model interpretability and understanding for PyTorch

Bayesian optimization in PyTorch

Pytorch domain library for recommendation systems

Training PyTorch models with differential privacy

A lightweight library for PyTorch training tools and utilities

A batched offline inference oriented version of segment-anything

A PyTorch repo for data loading and utilities to be shared by the PyTorch domain libraries.

Helpful tools and examples for working with flex-attention

PyTorch media decoding and encoding

PyTorch Single Controller

LeanRL is a fork of CleanRL, where selected PyTorch scripts optimized for performance using compile and cudagraphs.

PyTorch-native post-training at scale

Fault tolerance for PyTorch (HSDP, LocalSGD, DiLoCo, Streaming DiLoCo)

Autonomous GPU Kernel Generation & Optimization via Deep Agents

TorchX is a universal job launcher for PyTorch applications. TorchX is designed to have fast iteration time for training/research and support for E2E production ML pipelines when you're ready.

torchcomms: a modern PyTorch communications API

Tritonbench is a collection of PyTorch custom operators with example inputs to measure their performance.

Applied AI experiments and examples for PyTorch

A library that contains a rich collection of performant PyTorch model metrics, a simple interface to create new metrics, a toolkit to facilitate metric computation in distributed training and tools for PyTorch model evaluations.

This is a repository for all workshop related materials.

Configuration classes enabling type-safe PyTorch configuration for Hydra apps

This repository contains the experimental PyTorch native float8 training UX

TritonParse: A Compiler Tracer, Visualizer, and Reproducer for Triton Kernels

torch::deploy (multipy for non-torch uses) is a system that lets you get around the GIL problem by running multiple Python interpreters in a single C++ process.

A performant, memory-efficient checkpointing library for PyTorch applications, designed with large, complex distributed workloads in mind.

Ship correct and fast LLM kernels to PyTorch

TorchFix - a linter for PyTorch-using code with autofix support

Cryptographically secure pseudorandom number generators for PyTorch

MSLK (Meta Superintelligence Labs Kernels) is a collection of PyTorch GPU operator libraries that are designed and optimized for GenAI training and inference, such as FP8 row-wise quantization and collective communications.

Triton-based Symmetric Memory operators and examples

TORCH_TRACE parser for PT2

An experimental implementation of compiler-driven automatic sharding of models across a given device mesh.

Example apps and demos using PyTorch's ExecuTorch framework

A storage solution for PyTorch tensors with distributed tensor support.

C++ implementations for various tokenizers (sentencepiece, tiktoken etc).

An experimental CPU backend for Triton (https//github.com/openai/triton)

A block oriented training approach for inference time optimization.

Conveniently export torch.compile compiled products into self-contained Python files

This module defines a type system for distributed training code, based off of JAX's sharding in types, but adapted for the PyTorch ecosystem.

Framework for Algorithmic Correctness Testing of Operators

torch_remat fine-grained activation checkpointing API

Chakra et_replay is a tool designed for replaying Chakra Execution Traces (ET) from machine learning models using PyTorch.

For building the world's largest dataset of GPU kernels.

An extension of `torchvision` for decoding AVIF and HEIC images.

Monarch CRD and operator for native kubernetes support

This repository contains the source code for a static website that provides documentation for each "Graph Break" identified by a Graph Break ID (GBID).

A curated collection of OpenEnv environments, versioned against OpenEnv core

Staging ground for pytorch/pytorch maintainers to collaborate on release notes for PyTorch releases. Houses the artifacts that chronicle each release.

This is a mirror of https//git.mlplatform.org/ml/ethos-u/ethos-u-core-driver.git/

The accel_kernels repository is under pytorch-labs, which houses experimental GPU and accelerator kernels implemented in Triton or other DSLs, for SoTA algorithms.

User facing skill definitions for PyTorch users

A data parallelism that shards a model block-wise to support optimizers like Shampoo and Muon

The website for Meta-PyTorch.org

No description.

This is a mirror of https//git.mlplatform.org/tosa/serialization_lib.git/