Skip to content

Pull requests: linkedin/Liger-Kernel

New pull request New
40 Open 702 Closed
40 Open 702 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 memory requirements for benchmark_jsd.py and benchmark_distill_jsd_loss.py
#1050 opened Jan 29, 2026 by Egor-Krivov Loading…
3 tasks done
1
NPU adds support for grpo loss
#1049 opened Jan 29, 2026 by TianHao324 Loading…
2 of 3 tasks
docs: fix typos and formatting in README
#1047 opened Jan 29, 2026 by hashwnath Loading…
Set transformers upper bound
#1046 opened Jan 27, 2026 by Tcc0403 Loading…
3 tasks
1
Moe kernel for transformers v5
#1045 opened Jan 27, 2026 by Ramshankar07 Loading…
3 tasks
12
[NPU] FIX fused_linear_jsd ub overflow and OOM on NPU
#1043 opened Jan 27, 2026 by MAYUNHUI666 Loading…
3 tasks done
[WIP]support fused neighborhood attention for npu
#1034 opened Jan 20, 2026 by Hailey-Zh Loading…
3 tasks
2
Transformers 5.0 Support for Qwen3 MOE models
#1019 opened Jan 15, 2026 by edbeeching • Draft
3 tasks
Fix runtime issues for FSDP/DeepSpeed training
#1017 opened Jan 15, 2026 by SajanGhimire1 Loading…
3 tasks
5
Support transfomers v5
#994 opened Jan 2, 2026 by Tcc0403 • Draft
3 tasks
4
[GRPO] add grpo loss types
#993 opened Jan 1, 2026 by kashif Loading…
2 of 3 tasks
1
Add mlp support for qwen3vl series and little refactor
#957 opened Nov 28, 2025 by Tcc0403 Loading…
1 of 3 tasks
1
@momochen
2
Add GLM4_MOE model support
#952 opened Nov 25, 2025 by vvvdwbvvv Loading…
2 of 3 tasks
@momochen
17
Fix visualization when data is from more than 1 gpu
#938 opened Nov 11, 2025 by Tcc0403 Loading…
3 tasks
feat(FLCE): add helion version of fused linear cross entropy
#928 opened Nov 3, 2025 by Tcc0403 • Draft
3 tasks
9
Fix #920: Check cross entropy monkey patch
#927 opened Nov 3, 2025 by 0xtoward Loading…
3 tasks
1
1
Add automatically-request-copilot-review.yaml workflow
#877 opened Sep 8, 2025 by ChrisCarini Loading…
@shimizust
6
Add support for GPT-OSS models
#850 opened Aug 9, 2025 by Comet0322 • Draft
1 of 3 tasks
@momochen
11
fixed_fused_linear_cross_entropy should pass through kwargs
#832 opened Jul 22, 2025 by npuichigo Loading…
3 tasks done
4
Optimize Runtime Perf
#806 opened Jul 12, 2025 by AndreSlavescu Loading…
3 tasks done
7
fix vocab_size path for gemma3
#774 opened Jun 25, 2025 by winglian Loading…
2
Chunked DPO
#721 opened May 21, 2025 by cyr0930 Loading…
3 tasks
1
Fix errorbar in benchmarks_visualizer.py
#711 opened May 13, 2025 by Stonepia Loading…
1 of 3 tasks
1
Solar pro implementation
#699 opened May 4, 2025 by KareemMusleh Loading…
2 of 4 tasks
3
Modified block and warp sizes for improved performance on XPU for both layernnorm and rmsnorm
#661 opened Apr 10, 2025 by Tarakarevu1 Loading…
3 tasks done
6
ProTip! Adding no:label will show everything without a label.