Skip to content

Pull requests: sgl-project/SpecForge

New pull request New
16 Open 162 Closed
16 Open 162 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

delete last message when regenerating train data
#248 opened Oct 8, 2025 by Shang-Pin Loading…
6 tasks done
3
fix: remove attention mask shift in flex attention implementation
#245 opened Oct 4, 2025 by jialefu Loading…
6 tasks
2
fix: remove attention mask shift & add pe shift
#244 opened Oct 3, 2025 by Liyuhui-12 Loading…
6 tasks
3
Fix resume offline train logic. Add loading optimizer state
#243 opened Sep 29, 2025 by hanq-moreh Loading…
6 tasks
3
Apply FSDP2 to offline training
#242 opened Sep 26, 2025 by j1young Loading…
6 tasks
4
fix: mid aux hidden layer id calculation in online mode
#240 opened Sep 24, 2025 by Liu-Xue-Song Loading…
6 tasks
2
support deepseek-v2-lite online train and support yarn rope
#224 opened Sep 8, 2025 by jiapingW Loading…
6 tasks
6
Added mistral model support
#208 opened Sep 1, 2025 by ValeGian Loading…
3 of 6 tasks
21
[Feature] VLM model support tp
#206 opened Sep 1, 2025 by KerwinKai • Draft
6 tasks
1
Support Train Eagle-3 By DeepSpeed
#197 opened Sep 1, 2025 by xq25478 Loading…
12
Adapt Eagle3 for Deepseek architecture
#186 opened Aug 28, 2025 by xuhaojie-2025 Loading…
6 tasks
1
Fix for when draft model hidden dimension is different from target model hidden dimension
#183 opened Aug 26, 2025 by yilian49 Loading…
6
supported think mode
#182 opened Aug 26, 2025 by jiapingW Loading…
6 tasks
25
Add Draft LoRA scripts high priority
#138 opened Aug 13, 2025 by shuaills • Draft
6 tasks
10
Feat: Support TP for long-context draft model training high priority
#117 opened Aug 6, 2025 by yd-oom Loading…
6 tasks
6
Added Eagle training support for Kimi-K2
#108 opened Aug 3, 2025 by xuhaojie-2025 Loading…
6
ProTip! no:milestone will show everything without a milestone.