Morph MORPH
SIGN IN SIGN UP
verl-project / verl UNCLAIMED

verl: Volcano Engine Reinforcement Learning for LLMs

20361 0 0 Python

ACTIVITY

Commits Reviews
week month year
COMMITS
20
in the last week
CONTRIBUTORS
19
active
STARS
20361
total
FORKS
0
total
TOP CONTRIBUTORS
H
Huazhong
2 commits
M
Mingxuan Yu
1 commit
J
Joel
1 commit
F
Faraz
1 commit
K
khazzz1c
1 commit
H
Han Zhenyu 韩振宇
1 commit
R
Rinne
1 commit
N
Nurxat
1 commit
E
ellen ma
1 commit
Y
Yan Bai
1 commit
RECENT COMMITS
M
[megatron, fsdp] feat: DP workload balance for SFT (#5679)
Mingxuan Yu
e7af162
J
[megatron] feat: support cp for bshd format (#5826)
Joel
fe1abd8
F
[reward] fix: disable signal.alarm() in math_verify to fix silent scoring failure in Ray workers (#5635)
Faraz
9586c97
K
[megatron] fix: pass use_distributed_optimizer to ddp_config in vanilla mbridge path (#5775)
khazzz1c
3d274f6
H
[single_controller] fix: Set `device_name` for `split_resource_pool` to prevent failure on NPU environments (#5824)
Han Zhenyu 韩振宇
bba740e
R
[rollout] fix: processor does not have image_processor. (#5823)
Rinne
ad8c236
N
[env] fix: Modify the package installation sequence in the Ascend installation guide (#5819)
Nurxat
99d6199
E
[sglang, rollout] fix: wire up LoRA adapter path for engine_workers + sglang sleep (#5769)
ellen ma
a43b838
Y
[megatron] feat: support dynamic CP (#5057)
Yan Bai
7e9a07c
Y
[tool] chore: remove hard-code tool agent loop in fully async (#5816)
Yuyang Ding
60900f0
X
[ci] feat: support Ascend A2/A3 docker image build pipeline for sglang (#5804)
xiazhahe
36a1320
H
[trainer] feat: enable expandable segment support for npu (#5795)
Huazhong
7074405
A
[2/n][rollout] feat: flowgrpo - add diffusion agent loop support (#5716)
Andy Zhou
feff37a
Z
[fsdp, model] feat: add qwen3.5 fsdp grpo training support. (#5682)
Zhang1Sheng
4045d67
L
[ci] fix: resolve oom when allocating weight transfer buffer in fully async test cases (#5791)
lynn
fc3a5ac
H
[trainer] feat: support use_remove_padding=False for mindspeed backend (#5768)
Huazhong
57021e3
B
[ci] chore: add npu nightly ci for dapo-moonlight-16b-megatron and modify log path (#5734)
Beirong Zhou
ac6e46d
S
[trainer] fix: MLFlow publishing metrics failure should be non-blocking. (#5771)
Sheila
f88ed61
Z
[rollout, tool] feat: support multi-step in skip_rollout v2 (#5556)
zyang6
01a2bd3
Z
[hardware, rollout] feat: enable MXFP8 rollout on Ascend 950 devices (DV100 & DV120) (#5756)
Zhijie Xia
62212c0