Skip to content

Pull requests: Lightning-AI/litgpt

New pull request New
12 Open 1,166 Closed
12 Open 1,166 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

Qwen3 MoE Preliminary: add intermediate_size argument to MLP modules
#2046 opened May 17, 2025 by ysjprojects Loading…
LLaMAMoE fixes
#2014 opened Apr 13, 2025 by ysjprojects • Draft
7
(WIP) DeepseekV3 (and Multi-Head Latent Attention)
#2012 opened Apr 13, 2025 by ysjprojects • Draft
4
Add Multi-head Latent Attention (DeepSeekv2) enhancement New feature or request
#1945 opened Feb 25, 2025 by simoneangarano Loading…
2
Support for KV caching and batched inference enhancement New feature or request
#1934 opened Feb 6, 2025 by mseeger • Draft
31
qwen2.5 long context enhancement New feature or request waiting on author
#1933 opened Feb 4, 2025 by ysjprojects Loading…
2
Raise error if disk is full before downloading weights bug Something isn't working
#1903 opened Jan 8, 2025 by rasbt • Draft
1
7
OLMo 2 enhancement New feature or request waiting on author
#1897 opened Jan 4, 2025 by ysjprojects Loading…
17
OpenCoder series enhancement New feature or request waiting on author
#1880 opened Dec 21, 2024 by ysjprojects • Draft
4
WIP: TensorParallel with new strategy
#1421 opened May 14, 2024 by awaelchli • Draft
Add LongLora for both full and lora fine-tuning enhancement New feature or request
#1350 opened Apr 24, 2024 by belerico • Draft
19
[WIP] Simplified preparation of pretraining datasets enhancement New feature or request
#1057 opened Mar 7, 2024 by awaelchli • Draft
1
ProTip! Exclude everything labeled bug with -label:bug.