Skip to content
@meta-pytorch

Meta PyTorch

Popular repositories Loading

  1. gpt-fast gpt-fast Public

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

    Python 6.3k 582

  2. torchtune torchtune Public

    PyTorch native post-training library

    Python 5.8k 756

  3. captum captum Public

    Model interpretability and understanding for PyTorch

    Python 5.7k 571

  4. botorch botorch Public

    Bayesian optimization in PyTorch

    Jupyter Notebook 3.6k 502

  5. torchrec torchrec Public

    Pytorch domain library for recommendation systems

    Python 2.6k 704

  6. opacus opacus Public

    Training PyTorch models with differential privacy

    Python 2k 399

Repositories

Showing 10 of 63 repositories
  • torchcomms Public

    torchcomms: a modern PyTorch communications API

    meta-pytorch/torchcomms's past year of commit activity
    C++ 400 BSD-3-Clause 194 36 716 Updated Oct 7, 2026
  • remat Public

    torch_remat fine-grained activation checkpointing API

    meta-pytorch/remat's past year of commit activity
    Python 22 BSD-3-Clause 3 2 2 Updated Oct 7, 2026
  • torch_checkpointing Public

    A modular approach to solving distributed checkpointing in PyTorch.

    meta-pytorch/torch_checkpointing's past year of commit activity
    Python 2 BSD-3-Clause 4 1 2 Updated Oct 7, 2026
  • monarch Public

    PyTorch Single Controller

    meta-pytorch/monarch's past year of commit activity
    Rust 1,079 BSD-3-Clause 175 168 217 Updated Oct 7, 2026
  • flex_shard Public

    Flexible parameter sharding and distributed optimization utilities for PyTorch training

    meta-pytorch/flex_shard's past year of commit activity
    Python 10 3 0 2 Updated Oct 7, 2026
  • torchinfer Public

    PyTorch inference enablement and optimization solution for serving DLRMv3 https//mlcommons.org/2026/02/dlrmv3-inference-meta/ model families

    meta-pytorch/torchinfer's past year of commit activity
    Python 1 BSD-3-Clause 0 0 0 Updated Oct 7, 2026
  • autoparallel Public

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

    meta-pytorch/autoparallel's past year of commit activity
    Python 97 BSD-3-Clause 27 11 45 Updated Oct 7, 2026
  • tritonbench Public

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

    meta-pytorch/tritonbench's past year of commit activity
    Python 372 BSD-3-Clause 92 8 (3 issues need help) 29 Updated Oct 7, 2026
  • tokenizers Public

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

    meta-pytorch/tokenizers's past year of commit activity
    C++ 50 BSD-3-Clause 35 5 9 Updated Oct 7, 2026
  • spmd_types Public

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

    meta-pytorch/spmd_types's past year of commit activity
    Python 45 BSD-3-Clause 2 0 3 Updated Oct 7, 2026