mirror of https://github.com/hpcaitech/ColossalAI.git synced 2026-07-15 06:54:52 +00:00

Files

YeAnbang 3bd6fa3c67 [hot-fix] Fix memory leakage bug, support TP+PP (#6258 )

* update help information

* update style

* fix

* minor fix

* support PP training

* add pp support

* remove unused code

* address conversation

* fix memory leakage support tp+pp

* move empty cache

* move empty cache

---------

Co-authored-by: Tong Li <tong.li35271158@gmail.com>

2025-08-05 13:59:02 +08:00

reward

fix reward score

2025-08-05 13:59:02 +08:00

__init__.py

[chat] add distributed impl (#6210 )

2025-08-05 13:59:01 +08:00

comm.py

[chat] add distributed impl (#6210 )

2025-08-05 13:59:01 +08:00

consumer.py

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

2025-08-05 13:59:02 +08:00

grpo_consumer.py

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

2025-08-05 13:59:02 +08:00

inference_backend.py

simplify vllm preprocessing input ids

2025-08-05 13:59:02 +08:00

launch.py

[Distributed RLHF] Integration of PP (#6257 )

2025-08-05 13:59:02 +08:00

loss.py

fix vllm

2025-08-05 13:59:02 +08:00

producer.py

[feat] add microbatch forwarding (#6251 )

2025-08-05 13:59:02 +08:00

README.md

[chat] add distributed impl (#6210 )

2025-08-05 13:59:01 +08:00

utils.py

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

2025-08-05 13:59:02 +08:00

README.md

Requirements

pip install cupy-cuda12x
python -m cupyx.tools.install_library --cuda 12.x --library nccl