Skip to content

Merge main to ray [1/n] - #1345

Closed
luccabb wants to merge 99 commits into
dong/launch_via_ray_mainfrom
merge_main_to_ray
Closed

Merge main to ray [1/n]#1345
luccabb wants to merge 99 commits into
dong/launch_via_ray_mainfrom
merge_main_to_ray

Conversation

@luccabb

@luccabb luccabb commented Sep 30, 2025

Copy link
Copy Markdown
Contributor

What does this PR do? Please describe:
A summary of the change or the issue that is fixed.

Fixes #{issue number}

Does your PR introduce any breaking changes? If yes, please list them:
List of all backwards-incompatible changes.

Check list:

  • Was the content of this PR discussed and approved via a GitHub issue? (no need for typos or documentation improvements)
  • Did you read the contributor guideline?
  • Did you make sure that your PR does only one thing instead of bundling different changes together?
  • Did you make sure to update the documentation with your changes? (if necessary)
  • Did you write any new necessary tests?
  • Did you verify new and existing tests pass locally with your changes?
  • Did you update the CHANGELOG? (no need for typos, documentation, or minor internal changes)

cbalioglu added 30 commits April 16, 2025 12:28
@cbalioglu
Add TrainUnit.set_data_epoch_nr (#1110)
0855ce4
@cbalioglu
Call setup_fairseq2 automatically (#1111)
f397cd0
@cbalioglu
Introduce keep_checkpoint_every_n_steps (#1112)
c7b58b2
@cbalioglu
Introduce save_model_only (#1113)
8020519
@cbalioglu
Dump batch when CUDA OOM (#1114)
4e8da04
@cbalioglu
Split internal package imports (#1115)
152a0a6
@cbalioglu
Move device_stat to its own file (#1116)
e043665
@cbalioglu
Remove NormFormer (#1117)
74be268
@cbalioglu
Introduce LayerStack module (#1118)
d115c48
@cbalioglu
Revise fairseq2.nn modules (#1119)
9c8e546
@cbalioglu
Rename TransformerDecoderModel to TransformerLanguageModel (#1120)
20e8d37
@cbalioglu
Move Transformer modules to model directory (#1121)
09fda22
@cbalioglu
Move collective functions to data_reader (#1122)
a42ae15
@cbalioglu
Introduce TransformerLMDecoder (#1123)
765ac9d
@cbalioglu
Support tensor parallel for unsharded checkpoints (#1127)
063ce06
@cbalioglu
Improve sub-gang initialization (#1128)
6fe3dc0
@cbalioglu
Introduce mmap configuration (#1129)
f68e220
@cbalioglu
Handle split_group errors (#1130)
03ff405
@cbalioglu
Remove StateBag (#1131)
884def8
@cbalioglu
Move FileSystem to root (#1132)
5dceaa3
@cbalioglu
Rename model.config to model.config_overrides (#1133)
a81668d
@cbalioglu
Refactor FSDP and activation checkpointing (#1135)
26772da
@cbalioglu
Handle DTensor correctly in get_module_size_info (#1136)
d7b022c
@cbalioglu
Revise CheckpointState (#1137)
33b4f5d
@cbalioglu
Move torch_compile recipe settings to model section (#1138)
a3aefe2
@cbalioglu
Extract model hubs to their own file (#1139)
3296850
@cbalioglu
Revise recipe metric bags (#1141)
b5789f8
@cbalioglu
Support activation checkpointing of every nth layer (#1142)
190fac3
@cbalioglu
Improve non-blocking host-to-device transfer (#1145)
c7bb138
@cbalioglu
Introduce LM pretrain recipe (#1150)
db75e86
cbalioglu and others added 25 commits September 7, 2025 08:06
@cbalioglu
Remove safetensors import check (#1283)
3dc155a
@cbalioglu
Improve CLI error messaging for unknown assets (#1282)
a809853
@cbalioglu
Fix HG export bug for LLaMA 1 and 2 (#1287)
a9d5360
@cbalioglu
Introduce register_in_memory_assets (#1288)
dad5abc
@zyaoj
[documentation] update doc for v0.5 (#1285)
8a1d4bf 
* update documentation

* address comments

* address comments

* address more comments + doctest in ci

* move doctest
@cbalioglu
Revise safe type check in CheckpointManager (#1284)
fd2f0bd
@cbalioglu
Bump to v0.5.2 (#1289)
8ae890e
@cbalioglu
Bump to v0.6.0.dev0 (#1297)
950c719
@cbalioglu
Relax default 'resume_mode' for wandb (#1299)
9387eca
@cbalioglu
Introduce a new 'all_but_last' option for regime.save_model_only (#…
f1d1cc2 
…1298)
@cbalioglu
Improve metric value handling in recorders (#1314)
97b2f14
@zyaoj
[doc] enhance v0.5.* documentation (sprint 1 - round 1) (#1306)
8d3263d 
* add doc about hub (model/tokenizer/dataset)

* add doc for llama

* address comments

* address comment

* fix index error

* add simple skeleton for recipe register

* address comments
@zyaoj
bump setuptools + update setup (#1319)
4090720
@cbalioglu
Support multiple parameter groups in LR scheduler factories (#1320)
4f21a56
@zyaoj
Remove legacy CLI entrypoint (#1324)
9794c71
@cbalioglu
Ensure device index set for eager NCCL initialization (#1329)
0632c13
@cbalioglu
Extend ObjectValidator to support composite objects (#1328)
aa7fac5
@cirquit
Highlight asset store functionality in docs (#1330)
3d446f7
@cbalioglu
Extend the gang documentation (#1334)
a5de3d7
@cbalioglu
Introduce support for optimizer parameter groups (#1332)
064b164
@cbalioglu
Introduce exit-on-error command line flag (#1337)
800af1b
@jacklanchantin
Add SFT Recipe (#1271)
479c502
@uralik
Add more verbosity in structured errors (#1339)
68f7d56
@cbalioglu
Add family to ModelArchitectureNotKnownError (#1340)
fe61c6e
@luccabb
merge main
8afb430
@meta-cla meta-cla Bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Sep 30, 2025
@luccabb luccabb changed the title Merge main to ray Merge main to ray [1/n] Sep 30, 2025
@cbalioglu

cbalioglu commented Nov 3, 2025

Copy link
Copy Markdown
Contributor

Closing as this is severely stale. We might consider opening another PR in the future.

@cbalioglu cbalioglu closed this Nov 3, 2025
@cbalioglu
cbalioglu deleted the merge_main_to_ray branch November 3, 2025 17:16
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@zyaoj zyaoj Awaiting requested review from zyaoj zyaoj will be requested when the pull request is marked ready for review zyaoj is a code owner
@artemru artemru Awaiting requested review from artemru artemru will be requested when the pull request is marked ready for review artemru is a code owner
@cbalioglu cbalioglu Awaiting requested review from cbalioglu cbalioglu will be requested when the pull request is marked ready for review cbalioglu is a code owner

Assignees

No one assigned

Labels

CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

10 participants

@luccabb @cbalioglu @mattsetz @uralik @cirquit @artemru @heffernankevin @avidale @zyaoj @jacklanchantin