Skip to content

Pull requests: Lightning-AI/litgpt

New pull request New
34 Open 1,318 Closed
34 Open 1,318 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

fix(tokenizer): extend decoding fix to cover Mistral no-spaces bug
#2285 opened Jul 15, 2026 by eeshsaxena Loading…
1
1
feat: Add OLMoE-1B-7B-0924 support
#2282 opened Jul 11, 2026 by prabhashj07 Loading…
1
[pre-commit.ci] pre-commit suggestions
#2281 opened Jul 6, 2026 by pre-commit-ci Bot Loading…
feat: optional torchembed fused Triton RoPE kernel (~3-4x training speedup)
#2280 opened Jul 6, 2026 by py-ai-dev Loading…
Handle tokenizer configs without tokenizer_class
#2278 opened Jun 25, 2026 by CodingFeng101 Loading…
build(deps): bump the gha-updates group across 1 directory with 4 updates CI / actions Continuous integration
#2274 opened Jun 25, 2026 by dependabot Bot Loading…
ci: add Lightning registry fallback for gated tokenizer tests
#2271 opened Jun 24, 2026 by bhimrazy Collaborator Loading…
1
Fix malformed <|im_start> token in Phi4Reasoning prompt style
#2267 opened Jun 16, 2026 by JSap0914 Loading…
1
fix(finetune): all_reduce val loss across devices for initial and final validation
#2266 opened Jun 13, 2026 by discobot Loading…
1
fix: make convert_from_litgpt output loadable with AutoModel.from_pretrained
#2265 opened Jun 13, 2026 by discobot Loading…
1
fix(finetune): apply train.max_norm gradient clipping in finetune scripts
#2264 opened Jun 13, 2026 by discobot Loading…
1
Add Lightning Model Registry fallback for gated HF tokenizer tests
#2262 opened Jun 11, 2026 by bhimrazy Collaborator • Draft
4 of 8 tasks
Attempt to fix #2220: Enable Flash Attention in KV-cache path
#2248 opened May 9, 2026 by woaiwang Loading…
1
feat(lora): add LoRA+ support via lora_plus_lr_ratio
#2247 opened May 2, 2026 by ramkrishs Loading…
1
build(deps-dev): bump litdata from 0.2.59 to 0.2.61 dependencies
#2246 opened May 1, 2026 by dependabot Bot Loading…
build(deps): update jsonargparse requirement from <=4.41,>=4.37 to >=4.37,<=4.48.0 dependencies
#2245 opened May 1, 2026 by dependabot Bot Loading…
build(deps-dev): update transformers requirement from <4.57,>=4.51.3 to >=4.51.3,<5.8 dependencies
#2244 opened May 1, 2026 by dependabot Bot Loading…
chore(model,prompts): add type annotations to module-level functions and __init__ methods
#2242 opened Apr 18, 2026 by Koushik-Salammagari Loading…
3 tasks done
chore(api): add type annotations to LLM class and module-level functions
#2241 opened Apr 18, 2026 by Koushik-Salammagari Loading…
3 tasks done
chore(data): add return type annotations to FLAN dataloader methods
#2240 opened Apr 18, 2026 by Koushik-Salammagari Loading…
3 tasks done
chore(utils): add type annotations to public functions in utils.py
#2237 opened Apr 14, 2026 by nuthalapativarun Loading…
3
docs(pretrain): add TinyStories pretraining section
#2236 opened Apr 14, 2026 by nuthalapativarun Loading…
1
2
fix(model): convert bool mask_cache to float additive mask for softcapping
#2235 opened Apr 14, 2026 by nuthalapativarun Loading…
1
3
Fix division by zero in LR scheduler when max_steps equals warmup_steps
#2212 opened Mar 9, 2026 by Br1an67 Contributor Loading…
1
Fix Mistral tokenizer missing spaces in decode_stream (Issue #1822)
#2211 opened Mar 5, 2026 by mrshibly Loading…
1
1
ProTip! Adding no:label will show everything without a label.