Fault tolerance for PyTorch (HSDP, LocalSGD, DiLoCo, Streaming DiLoCo)
☆540Sep 21, 2026Updated this week
Alternatives and similar repositories for torchft
Users that are interested in torchft are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- PyTorch Single Controller☆1,076Updated this week
- torchcomms: a modern PyTorch communications API☆397Updated this week
- NVIDIA Resiliency Extension is a python package for framework developers and users to implement fault-tolerant features. It improves the …☆336Updated this week
- A storage solution for PyTorch tensors with distributed tensor support.☆89Updated this week
- A PyTorch native platform for training generative AI models☆5,757Updated this week
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- An experimental implementation of compiler-driven automatic sharding of models across a given device mesh.☆97Updated this week
- A Python-embedded DSL that makes it easy to write fast, scalable ML kernels with minimal boilerplate.☆950Updated this week
- 🚀 Efficiently (pre)training foundation models with native PyTorch features, including FSDP for training and SDPA implementation of Flash…☆288Nov 24, 2025Updated 9 months ago
- Byted PyTorch Distributed for Hyperscale Training of LLMs and RLs☆1,040Mar 3, 2026Updated 6 months ago
- A Quirky Assortment of CuTe Kernels☆1,151Updated this week
- PyTorch native quantization for training and inference☆2,983Updated this week
- PCCL (Prime Collective Communications Library) implements fault tolerant collective communications over IP☆159Sep 12, 2025Updated last year
- Tritonbench is a collection of PyTorch custom operators with example inputs to measure their performance.☆369Updated this week
- TorchX is a universal job launcher for PyTorch applications. TorchX is designed to have fast iteration time for training/research and sup…☆427Updated this week
- Proton VPN Special Offer - Get 70% off • AdSpecial partner offer. Trusted by over 100 million users worldwide. Tested, Approved and Recommended by Experts.
- Helpful tools and examples for working with flex-attention☆1,252Updated this week
- Distributed Compiler and Optimized Parallel Kernels☆1,549Updated this week
- A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit and 4-bit floating point (FP8 and FP4) precision on H…☆3,549Updated this week
- A library to analyze PyTorch traces.☆556Sep 11, 2026Updated last week
- CUDA checkpoint and restore utility☆491Jul 6, 2026Updated 2 months ago
- Accelerating MoE with IO and Tile-aware Optimizations☆768Aug 29, 2026Updated 3 weeks ago
- A fast communication-overlapping library for tensor/expert parallelism on GPUs.☆1,363Aug 28, 2025Updated last year
- Zero Bubble Pipeline Parallelism☆466May 7, 2025Updated last year
- ☆1,593Mar 25, 2026Updated 5 months ago
- End-to-end encrypted email - Proton Mail • AdSpecial offer: 40% Off Yearly / 80% Off First Month. All Proton services are open source and independently audited for security.
- ☆171Dec 27, 2024Updated last year
- Fast low-bit matmul kernels in Triton☆486Aug 9, 2026Updated last month
- PyTorch-native post-training at scale☆703Sep 9, 2026Updated last week
- A CPU+GPU Profiling library that provides access to timeline traces and hardware performance counters.☆995Updated this week
- A performant, memory-efficient checkpointing library for PyTorch applications, designed with large, complex distributed workloads in mind…☆166Aug 19, 2026Updated last month
- TORCH_TRACE parser for PT2☆94May 11, 2026Updated 4 months ago
- Minimalistic large language model 3D-parallelism training☆2,825Updated this week
- CCLoco: Scaling Up Top-K Error Feedback with Local Optimizers☆27Aug 22, 2025Updated last year
- This repository contains the experimental PyTorch native float8 training UX☆226Aug 1, 2024Updated 2 years ago
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- Tile primitives for speedy kernels☆3,724Sep 12, 2026Updated last week
- This module defines a type system for distributed training code, based off of JAX's sharding in types, but adapted for the PyTorch ecosys…☆45Updated this week
- Minimalistic 4D-parallelism distributed training framework for education purpose☆2,307Aug 26, 2025Updated last year
- GPUd automates monitoring, diagnostics, and issue identification for GPUs☆496Updated this week
- Checkpoint-engine is a simple middleware to update model weights in LLM inference engines☆1,005Sep 15, 2026Updated last week
- Scalable and Performant Data Loading☆397Updated this week
- Efficient Triton Kernels for LLM Training☆6,627Updated this week