Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Nov 3, 2022

• Rename pod-config to tpu-config + docs (huggingface#818)

  muellerzr

  authored
  874c496

Commits on Nov 2, 2022

• Update CLI docs and use mps rather than mps_device (huggingface#814)

  muellerzr

  authored
  dc9966d

• Mlflow-tracker-v2 🔥 (huggingface#794)

  

  nbroad1881

  and

  sgugger

  authored
  e2cd36b

• Act on deprecations (huggingface#813)

  muellerzr

  authored
  6a0082d

• add recurse argument in remove_hook_from_module (huggingface#812)

  

  younesbelkada

  and

  sgugger

  authored
  102cf00

• adding support to pickle and unpickle AcceleratedOptimizer (huggingface#811)

  

  pacman100

  and

  BenjaminBossan

  authored
  359bd1b

• Refactor CLI to improve readability (huggingface#810)

  muellerzr

  authored
  0de1644

Commits on Nov 1, 2022

• Introduce a pod-config command (huggingface#802)

  

  3 peopleauthored
  b816e25

Commits on Oct 31, 2022

• Deal with optimizer.differentiable in PyTorch 1.13.0 (huggingface#803)

  

  comaniac

  and

  sgugger

  authored
  c4c444a

• fix: add pdsh as default launcher (huggingface#800)

  zanussbaum

  authored
  f3129d1

• Fix extraction of state dict in offload (huggingface#795)

  sgugger

  authored
  8c92805

• Fix device_map="auto" on CPU-only envs (huggingface#797)

  sgugger

  authored
  8c0505d

• Add even_batches keyword to Accelerator (huggingface#781)

  

  3 peopleauthored
  16d548c

Commits on Oct 27, 2022

• Consider top-level buffers when computing infer_auto_device_map (huggingface#792)

  

  younesbelkada

  and

  sgugger

  authored
  415b738

Commits on Oct 25, 2022

• separate dataloader generator from sampler generator (huggingface#789)

  

  3 peopleauthored
  a552540

Commits on Oct 21, 2022

• Add Dev Container configuration (huggingface#782)

  Chris-hughes10

  authored
  37b2aa0

Commits on Oct 20, 2022

• Work in kaggle! (huggingface#783)

  muellerzr

  authored
  4df576e

Commits on Oct 19, 2022

• fix transformers tests (huggingface#777)

  ArthurZucker

  authored
  87a7e07

• Add same_network + docs (huggingface#780)

  muellerzr

  authored
  5c8f181

• Make rich toggleable and seperate out a new environment utility file (huggingface#779)

  muellerzr

  authored
  6f7fa4f

• Allow BatchSamplerShard to not even out batches (huggingface#776)

  

  sgugger

  and

  muellerzr

  authored
  15a854e

• Add defaults for launchers (huggingface#778)

  muellerzr

  authored
  63d0653

Commits on Oct 18, 2022

• Fix flakey wandb test (huggingface#775)

  muellerzr

  authored
  21b7f15

• Fix all github actions issues + depreciations (huggingface#773)

  muellerzr

  authored
  49cd8d3

• Fix number of devices in get_balanced_memory (huggingface#774)

  sgugger

  authored
  1eafa55

• Regression cli tests (huggingface#772)

  muellerzr

  authored
  9114fb0

• Move io_same_device hook to before attach_align_device hook on cpu_offload and disk_offload. (huggingface#768)

  piEsposito

  authored
  5e8ab12

Commits on Oct 17, 2022

• updating docs to use fork of megatron-lm and minor example/docs fix (huggingface#766)

  pacman100

  authored
  a635111

• Only wrap modules in DDP if they require grad (huggingface#761)

  samuelstevens

  authored
  a7334df

Commits on Oct 15, 2022

• update docs (huggingface#759)

  

  pacman100

  and

  sgugger

  authored
  4a7268d

• refactor (huggingface#758)

  pacman100

  authored
  148f6dc

Commits on Oct 14, 2022

• Return unclipped gradient from grad_clip_norm_ (huggingface#756)

  samuelstevens

  authored
  693d468

• ensure megatron is 2.2.0+ (huggingface#755)

  jeffra

  authored
  dfba92a

Commits on Oct 12, 2022

• Change num_cpu_threads_per_process default (huggingface#753)

  

  muellerzr

  and

  sgugger

  authored
  4dc5049

• Megatron-LM integration (huggingface#667)

  

  pacman100

  and

  sgugger

  authored
  e3ebf17

Pagination

PreviousNext