Alternatives and similar repositories for out-of-the-box-fp8-training

Users that are interested in out-of-the-box-fp8-training are comparing it to the libraries listed below

• srush / triton-autodiff
  Experiment of using Tangent to autodiff triton
  ☆80Updated last year
• cloneofsimo / min-fsdp
  ☆91Updated last year
• drisspg / transformer_nuggets
  A place to store reusable transformer components of my own creation or found on the interwebs
  ☆60Updated last week
• lucidrains / autoregressive-linear-attention-cuda
  CUDA implementation of autoregressive linear attention, with all the latest research findings
  ☆45Updated 2 years ago
• graphcore-research / unit-scaling
  A library for unit scaling in PyTorch
  ☆130Updated 3 months ago
• alexzhang13 / Triton-Puzzles-Solutions
  Personal solutions to the Triton Puzzles
  ☆20Updated last year
• mgmalek / efficient_cross_entropy
  ☆121Updated last year
• meta-pytorch / float8_experimental
  This repository contains the experimental PyTorch native float8 training UX
  ☆223Updated last year
• ezyang / torchdbg
  PyTorch centric eager mode debugger
  ☆48Updated 10 months ago
• GindaChen / FlexFlashAttention3
  FlexAttention w/ FlashAttention3 Support
  ☆27Updated last year
• UmerHA / triton_util
  Make triton easier
  ☆48Updated last year
• open-lm-engine / flash-model-architectures
  A bunch of kernels that might make stuff slower 😉
  ☆62Updated this week
• softmax1 / Flash-Attention-Softmax-N
  CUDA and Triton implementations of Flash Attention with SoftmaxN.
  ☆73Updated last year
• lernapparat / torchhacks
  Hacks for PyTorch
  ☆19Updated 2 years ago
• HazyResearch / train-tk
  train with kittens!
  ☆63Updated 11 months ago
• lianakoleva / no-libtorch-compile
  ☆21Updated 7 months ago
• amirzandieh / HyperAttention
  Triton Implementation of HyperAttention Algorithm
  ☆48Updated last year
• shikaiqiu / compute-better-spent
  ☆58Updated last year
• Zyphra / tree_attention
  Tree Attention: Topology-aware Decoding for Long-Context Attention on GPU clusters
  ☆130Updated 10 months ago
• cloneofsimo / min-max-gpt
  Minimal (400 LOC) implementation Maximum (multi-node, FSDP) GPT training
  ☆132Updated last year
• Jokeren / triton-samples
  ☆28Updated 9 months ago
• nshepperd / flash_attn_jax
  JAX bindings for Flash Attention v2
  ☆97Updated this week
• lucaslingle / mu_transformer
  Transformer with Mu-Parameterization, implemented in Jax/Flax. Supports FSDP on TPU pods.
  ☆32Updated 4 months ago
• stanford-futuredata / stk
  ☆112Updated last year
• dvruette / barrel-rec-pytorch
  ☆53Updated last year
• eth-easl / fmengine
  Utilities for Training Very Large Models
  ☆58Updated last year
• epfml / dynamic-sparse-flash-attention
  ☆149Updated 2 years ago
• insuhan / hyper-attn
  ☆83Updated last year
• ClashLuke / tpucare
  Automatically take good care of your preemptible TPUs
  ☆37Updated 2 years ago
• google-deepmind / asyncdiloco
  ☆46Updated last year