fix: wrong dp-rank condition when enable pp#6374

Open

liuqh16 wants to merge 116 commits intohpcaitech:mainfrom

liuqh16:fix-dp-rank

Commits on Feb 21, 2025

Commits on Feb 23, 2025

Commits on Feb 25, 2025

update grpo
TongLi3701
committed

Commits on Feb 28, 2025

polish
TongLi3701
committed

Commits on Mar 6, 2025

modify data loader
Tong Li
committed
grpo consumer
Tong Li
committed
update loss
Tong Li
committed
update reward fn
Tong Li
committed
update example
Tong Li
committed
update loader
Tong Li
committed
add algo selection
Tong Li
committed
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot]
committed
add save
Tong Li
committed
update select algo
Tong Li
committed
Merge branch 'grpo-latest' of github.com:hpcaitech/ColossalAI into grpo-latest
Tong Li
committed
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot]
committed

Commits on Mar 10, 2025

Commits on Mar 11, 2025

Commits on Mar 13, 2025

fix tp bug
Tong Li
committed
fix consumer
Tong Li
committed
convert to 8 generation
Tong Li
committed
print results
Tong Li
committed
setup update
Tong Li
committed

Commits on Mar 14, 2025

fix transformers backend
YeAnbang
committed

Commits on Mar 18, 2025

[Feature] Support Distributed LogProb for GRPO Training (#6247)

duanjunwen
and
Tong Li
authored

Commits on Mar 19, 2025

fix vllm
YeAnbang
committed

Commits on Mar 28, 2025

[feat] add microbatch forwarding (#6251)

YeAnbang
and
pre-commit-ci[bot]
authored

Commits on Apr 9, 2025

[Distributed RLHF] Integration of PP (#6257)

YeAnbang
and
Tong Li
authored

Commits on Apr 10, 2025

[hot-fix] Fix memory leakage bug, support TP+PP (#6258)

YeAnbang
and
Tong Li
authored

Commits on Apr 22, 2025

add prompt template (#6273)

TongLi3701
and
Tong Li
authored

Commits on Apr 23, 2025

[feat] Add final save at the end (#6274)
TongLi3701
authored

Commits on Apr 25, 2025

[feat] Support DAPO (#6263)

authored

Commits on Apr 26, 2025

fix checkpoint naming; add num_epoch parameter (#6277)
YeAnbang
authored

Commits on Apr 27, 2025

fix save issue (#6279)

TongLi3701
and
Tong Li
authored

Commits on Apr 28, 2025

fix pp+tp, fix dataloader (#6280)
YeAnbang
authored

Commits on Apr 29, 2025

[feat] Support boxed math reward (#6284)
YeAnbang
authored

Commits on Apr 30, 2025

Commits on May 1, 2025

rewrite reward fn
YeAnbang
committed

Commits on May 3, 2025

Merge pull request #6292 from hpcaitech/grpo-latest-dev-reward-update
YeAnbang
authored

Commits on May 7, 2025

[fix] revert reward update and evaluation (#6295)
YeAnbang
authored

Commits on May 13, 2025

update pad seq (#6303)

TongLi3701
and
Tong Li
authored

Commits on May 14, 2025

Support evaluation during training
YeAnbang
committed
[feat] Support prompt level dynamic (#6300)

TongLi3701
and
Tong Li
authored
move logging to producer
YeAnbang
committed
merge
YeAnbang
committed

Commits on May 15, 2025

Commits on May 17, 2025

fix logging rollouts
YeAnbang
committed

Commits on May 19, 2025

merge reward and eval
YeAnbang
committed

Commits on May 20, 2025

Commits on May 21, 2025

fix missing tags parameter
YeAnbang
committed

Commits on May 22, 2025

fix default eval setting (#6321)

TongLi3701
and
Tong Li
authored

Commits on May 28, 2025

Commits on May 29, 2025

Commits on Jun 5, 2025

Commits on Jun 10, 2025

Commits on Jun 11, 2025

fix pp memory issue (#6344)

TongLi3701
and
Tong Li
authored

Commits on Jun 13, 2025

optimize pp log_softmax OOM
YeAnbang
committed

Commits on Jun 18, 2025

implement memory efficient logprob
YeAnbang
committed

Commits on Jun 19, 2025

Commits on Jun 30, 2025

[feat[ Support one-behind to reduce bubble time. Add profiling code (#6353)

YeAnbang
and
Tong Li
authored

Commits on Jul 14, 2025

Commits on Jul 17, 2025

add entropy (#6363)
YeAnbang
authored

Commits on Jul 22, 2025

hotfix entropy calculation (#6364)
YeAnbang
authored

Commits on Jul 30, 2025