Commits

Breadcrumbs

History for

on5d709af

User selector

All users

Datepicker

All time

Commit History

Commits on Feb 3, 2023

• [PTD] Land 'to_std' utility parser fix pytorch#93209 (pytorch#94023)

  Show description for e98a942
  

  fduwjj

  authored and

  pytorchmergebot

  committed
  e98a942

  Copy full SHA for e98a942

• [dtensor] add split_with_sizes op (pytorch#93957)

  Show description for d05ec0e
  

  wanchaol

  authored and

  pytorchmergebot

  committed
  d05ec0e

  Copy full SHA for d05ec0e

• [FSDP][optim_state_dict] Ignore the state check on rank that does not own the corresponding parameter (pytorch#93318)

  Show description for 0f5b6ca
  

  fegin

  authored and

  pytorchmergebot

  committed
  0f5b6ca

  Copy full SHA for 0f5b6ca

Commits on Feb 2, 2023

• [FSDP][optim_state_dict] Make FSDP.optim_state_dict compatbile with DMP (pytorch#93285)

  Show description for e32d99a
  

  fegin

  authored and

  pytorchmergebot

  committed
  e32d99a

  Copy full SHA for e32d99a

• Basic Validation for FSDP state_dict transformations of modules with persistent buffers (pytorch#93396)

  Show description for 769eca6
  

  speediedan

  authored and

  pytorchmergebot

  committed
  769eca6

  Copy full SHA for 769eca6

• [FSDP][3/N] Refactor summon_full_params unit tests (pytorch#92298)

  Show description for 481a334
  

  awgu

  authored and

  pytorchmergebot

  committed
  481a334

  Copy full SHA for 481a334

• [FSDP][2/N] _summon_full_params -> _unshard_params (pytorch#92297)

  Show description for 1099073
  

  awgu

  authored and

  pytorchmergebot

  committed
  1099073

  Copy full SHA for 1099073

• [FSDP][optim_state_dict][10/N] Make optim_state_dict and optim_state_dict_to_load public (pytorch#92118)

  Show description for 4b0f1cc
  

  fegin

  authored and

  pytorchmergebot

  committed
  4b0f1cc

  Copy full SHA for 4b0f1cc

• [DTensor] implement dist_split as a sharding prop rule (pytorch#93306)

  Show description for 6f3018d
  

  XilunWu

  authored and

  pytorchmergebot

  committed
  6f3018d

  Copy full SHA for 6f3018d

• [DTensor] fix DTensorSpec dim_map description (pytorch#93160)

  Show description for b82f93d
  

  XilunWu

  authored and

  pytorchmergebot

  committed
  b82f93d

  Copy full SHA for b82f93d

• [PT-D][NamedOptimizer][6/N] Upstream init_state from keyed to NamedOptimizer (pytorch#93887)

  Show description for e7ace1f
  

  fduwjj

  authored and

  pytorchmergebot

  committed
  e7ace1f

  Copy full SHA for e7ace1f

Commits on Feb 1, 2023

• [PT-D][TP] Fix the module registration in TP API (pytorch#93412)

  Show description for 3fb6e11
  

  fduwjj

  authored and

  pytorchmergebot

  committed
  3fb6e11

  Copy full SHA for 3fb6e11

• [mta][foreach] Implement fused adamw (pytorch#88015)

  Show description for a23ed38
  

  crcrpar

  authored and

  pytorchmergebot

  committed
  a23ed38

  Copy full SHA for a23ed38

• [dtensor][6/N] change to a better/safer op registration (pytorch#90735)

  Show description for 60e503d
  

  wanchaol

  authored and

  pytorchmergebot

  committed
  60e503d

  Copy full SHA for 60e503d

• [dtensor][5/N] add cached propagator for TP (pytorch#90734)

  Show description for 9a56997
  

  wanchaol

  authored and

  pytorchmergebot

  committed
  9a56997

  Copy full SHA for 9a56997

• [dtensor][4/N] refactor dispatching logic and add propagator (pytorch#90733)

  Show description for b072245
  

  wanchaol

  authored and

  pytorchmergebot

  committed
  b072245

  Copy full SHA for b072245

Commits on Jan 31, 2023

• [FSDP][optim_state_dict] Fix _is_named_optimizer when the state is empty (pytorch#93303)

  Show description for 888771d
  

  fegin

  authored and

  pytorchmergebot

  committed
  888771d

  Copy full SHA for 888771d

Commits on Jan 30, 2023

• [pytorch] fix backend_type for backend/PG plugin (pytorch#93129)

  Show description for 1fa68d4
  

  kingchc

  authored and

  pytorchmergebot

  committed
  1fa68d4

  Copy full SHA for 1fa68d4

• [Reland][FSDP] Do not clean FQNs for use_orig_params=True (pytorch#92662)

  Show description for 3e4d0e8
  

  awgu

  authored and

  pytorchmergebot

  committed
  3e4d0e8

  Copy full SHA for 3e4d0e8

Commits on Jan 27, 2023

• [FSDP][optim_state_dict] Ensure correct devices for tensors when doing all_gather (pytorch#92992)

  Show description for a423897
  

  fegin

  authored and

  pytorchmergebot

  committed
  a423897

  Copy full SHA for a423897

• [FSDP][optim_state_dict] Use all_gather to deal with uneven size tensors (pytorch#92991)

  Show description for 8b1b47c
  

  fegin

  authored and

  pytorchmergebot

  committed
  8b1b47c

  Copy full SHA for 8b1b47c

• [DTensor] implement dist_cat as a sharding prop rule (pytorch#92677)

  Show description for 8b3e01c
  

  XilunWu

  authored and

  pytorchmergebot

  committed
  8b3e01c

  Copy full SHA for 8b3e01c

Commits on Jan 26, 2023

• [PT-D][TP] Fix TP API for FQN path based parallelization (pytorch#93029)

  Show description for 913866e
  

  fduwjj

  authored and

  pytorchmergebot

  committed
  913866e

  Copy full SHA for 913866e

• [PT-D] Enable Meta Tensor Support for DTensor (pytorch#92652)

  Show description for 77f3366
  

  fduwjj

  authored and

  pytorchmergebot

  committed
  77f3366

  Copy full SHA for 77f3366

• [nn] zero_grad() set_to_none default True (pytorch#92731)

  Show description for b90496e
  

  janeyx99

  authored and

  pytorchmergebot

  committed
  b90496e

  Copy full SHA for b90496e

Commits on Jan 25, 2023

• [PT-D][Checkpoint] Rename DCP storage layer init() (pytorch#92869)

  Show description for dd05f02
  

  wz337

  authored and

  pytorchmergebot

  committed
  dd05f02

  Copy full SHA for dd05f02

• [FSDP] Fix no_sync(), use_orig_params=True, mixed precision, sharded (pytorch#92874)

  Show description for c0ed0f2
  

  awgu

  authored and

  pytorchmergebot

  committed
  c0ed0f2

  Copy full SHA for c0ed0f2

• add number of cuda retries into tracker (pytorch#92557)

  Show description for 077e135
  

  zhaojuanmao

  authored and

  pytorchmergebot

  committed
  077e135

  Copy full SHA for 077e135

Commits on Jan 24, 2023

• Fix MPI backend PG initialization (pytorch#92847)

  Show description for 2503a4a
  

  H-Huang

  authored and

  pytorchmergebot

  committed
  2503a4a

  Copy full SHA for 2503a4a

• Add documentation for DCP. (pytorch#92813)

  Show description for 9e56378
  

  Rodrigo Kumpera

  authored and

  pytorchmergebot

  committed
  9e56378

  Copy full SHA for 9e56378

• [PT-D][Checkpoint]rename init() (pytorch#92829)

  Show description for dd4b46e
  

  wz337

  authored and

  pytorchmergebot

  committed
  dd4b46e

  Copy full SHA for dd4b46e

Commits on Jan 23, 2023

• [PT-D][Checkpoint]Resolve issue pytorch#89501: Rename _nested_tensor.py to (pytorch#92705)

  Show description for f7e1f3e
  

  wz337

  authored and

  pytorchmergebot

  committed
  f7e1f3e

  Copy full SHA for f7e1f3e

• [FSDP][optim_state_dict] Fix the conditions to check non-parameter associated states (pytorch#92744)

  Show description for 8f294f7
  

  fegin

  authored and

  pytorchmergebot

  committed
  8f294f7

  Copy full SHA for 8f294f7

• [FSDP][Easy] Remove outdated comment (pytorch#92739)

  Show description for 63d6ee7
  

  awgu

  authored and

  pytorchmergebot

  committed
  63d6ee7

  Copy full SHA for 63d6ee7

• [PT-D] Enable init ops for DTensor (pytorch#92651)

  Show description for b985c2e
  

  fduwjj

  authored and

  pytorchmergebot

  committed
  b985c2e

  Copy full SHA for b985c2e

Pagination

Previous

Next