Skip to content

Pull requests: alibaba/ChatLearn

New pull request New
13 Open 267 Closed
13 Open 267 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Force forward_step and train_step have same data order
#311 opened Jul 10, 2025 by yytang0204 Loading…
[WIP] MCore-vLLM Parameter Synchronization
#310 opened Jul 10, 2025 by lostkevin • Draft
[WIP] Fix dp consistency
#308 opened Jul 7, 2025 by yytang0204 Loading…
adopt per-token-loss for mcore GRPO
#307 opened Jul 7, 2025 by lostkevin Loading…
fix yaml empty conf
#292 opened Jun 12, 2025 by tianliplus Loading…
feat(wandb): enable logging
#225 opened Jan 31, 2025 by slimfrkha Loading…
12
regroup k/v weight.
#219 opened Jan 30, 2025 by adoda Loading…
support vllm-0.6.6
#214 opened Jan 22, 2025 by charles9304 Loading…
2
fix[vllm benchmark]: avoid num input tokens overflow
#213 opened Jan 22, 2025 by haolin-nju Loading…
fix(parameter_sync): default regrouping parameters to all-gather instead of all-to-all
#202 opened Jan 6, 2025 by haolin-nju Loading…
1
feature(parameter_sync): offer 2 GPU memory optimization levels in broadcasting parameters
#195 opened Dec 27, 2024 by haolin-nju Loading…
2
Enable context parallelism in SFT
#190 opened Dec 24, 2024 by slimfrkha Loading…
5
[WIP]feature(mixtral): support Mixtral-8x7B SFT, Reward, and Alignment
#95 opened Sep 24, 2024 by haolin-nju Loading…
8
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.