Back to search
Top Contributors
118
commits
62
commits
42
commits
40
commits
25
commits
24
commits
23
commits
23
commits
18
commits
18
commits
Efficient Triton Kernels for LLM Training
Stars
6.3k
Forks
509
Watchers
6.3k
Open Issues
140
Overall repository health assessment
No package.json found
This might not be a Node.js project
118
commits
62
commits
42
commits
40
commits
25
commits
24
commits
23
commits
23
commits
18
commits
18
commits
[Feature] Use Liger's Relu_Squared kernel for Nemotron models (#1176)
1c013e2View on GitHubImprove liger-autopatch skill to enable modifying existing monkey-patches (#1177)
f16c9f7View on GitHub[NPU]: optimize cross entropy kernel gradient computation (#1178)
35cf6a0View on GitHub[Benchmark]: Add --sweep-mode and --bt to benchmark CLI. (#1163)
d991472View on GitHubAdd Kimi AttentionResiduals (AttnRes) kernelFeature/add attn res kernel (#1161)
3f85b72View on GitHub[bugfix] Remove the pos_ids parameter from apply_rotary_pos_emb (#1162)
cc7d605View on GitHubClaude skill for automatically monkey patching HF transformers models (liger-autopatch) (#1167)
f1b7e47View on GitHub[Feature] Add Triton kernel for ReLU Squared activation (#1171)
39a1f45View on GitHubClaude skill for Triton kernel development (liger-kernel-dev) (#1170)
0169c4dView on GitHubfix: guard save_for_backward on grad_bias not bias in fused linear CE forward (#1157)
d8d6630View on GitHub