.gitignore
.pre-commit-config.yaml
LICENSE
Makefile
NOTICE
README.md
mkdocs.yml
pyproject.toml
setup.py
.claude/skills/liger-autopatch/SKILL.md
.claude/skills/liger-autopatch/code-generator.md
.claude/skills/liger-autopatch/decision-matrix.md
.claude/skills/liger-autopatch/model-analyzer.md
.claude/skills/liger-autopatch/validator.md
.claude/skills/liger-autopatch/examples/gemma-profile.md
.claude/skills/liger-autopatch/examples/llama-profile.md
.claude/skills/liger-autopatch/templates/lce-forward-dense.md
.claude/skills/liger-autopatch/templates/lce-forward-moe.md
.claude/skills/liger-autopatch/templates/monkey-patch-fn.md
.claude/skills/liger-autopatch/templates/test-convergence.md
.claude/skills/liger-autopatch/templates/test-instance-patch.md
.claude/skills/liger-kernel-dev/SKILL.md
.claude/skills/liger-kernel-dev/analyzer.md
.claude/skills/liger-kernel-dev/generator.md
.claude/skills/liger-kernel-dev/kernel-profile-format.md
.claude/skills/liger-kernel-dev/validator.md
.claude/skills/liger-kernel-dev/examples/cross-entropy-profile.md
.claude/skills/liger-kernel-dev/examples/rms-norm-profile.md
.claude/skills/liger-kernel-dev/examples/swiglu-profile.md
.claude/skills/liger-kernel-dev/templates/benchmark.md
.claude/skills/liger-kernel-dev/templates/functional-api.md
.claude/skills/liger-kernel-dev/templates/module-wrapper.md
.claude/skills/liger-kernel-dev/templates/ops-kernel.md
.claude/skills/liger-kernel-dev/templates/unit-test.md
.claude/skills/liger-kernel-perf/SKILL.md
.claude/skills/liger-kernel-perf/finalizer.md
.claude/skills/liger-kernel-perf/optimization-strategies.md
.claude/skills/liger-kernel-perf/optimizer.md
.claude/skills/liger-kernel-perf/profiler.md
.claude/skills/liger-kernel-perf/templates/optimization-profile.md
.claude/skills/liger-kernel-perf/templates/variant-notes.md
.github/pull_request_template.md
.github/ISSUE_TEMPLATE/bug_report.yaml
.github/ISSUE_TEMPLATE/feature_request.yaml
.github/workflows/amd-ci.yml
.github/workflows/benchmark.yml
.github/workflows/checkstyle.yml
.github/workflows/docs.yml
.github/workflows/intel-ci.yml
.github/workflows/merge-queue-failure.yml
.github/workflows/nvi-ci.yml
.github/workflows/publish-nightly.yml
.github/workflows/publish-release.yml
benchmark/BENCHMARK_GUIDELINES.md
benchmark/README.md
benchmark/__init__.py
benchmark/benchmarks_visualizer.py
benchmark/data/all_benchmark_data.csv
benchmark/scripts/__init__.py
benchmark/scripts/benchmark_attn_res.py
benchmark/scripts/benchmark_cpo_loss.py
benchmark/scripts/benchmark_cross_entropy.py
benchmark/scripts/benchmark_distill_cosine_loss.py
benchmark/scripts/benchmark_distill_jsd_loss.py
benchmark/scripts/benchmark_dpo_loss.py
benchmark/scripts/benchmark_dyt.py
benchmark/scripts/benchmark_embedding.py
benchmark/scripts/benchmark_fused_add_rms_norm.py
benchmark/scripts/benchmark_fused_linear_cross_entropy.py
benchmark/scripts/benchmark_fused_linear_jsd.py
benchmark/scripts/benchmark_fused_moe.py
benchmark/scripts/benchmark_fused_neighborhood_attention.py
benchmark/scripts/benchmark_geglu.py
benchmark/scripts/benchmark_group_norm.py
benchmark/scripts/benchmark_grpo_loss.py
benchmark/scripts/benchmark_jsd.py
benchmark/scripts/benchmark_kl_div.py
benchmark/scripts/benchmark_kto_loss.py
benchmark/scripts/benchmark_layer_norm.py
benchmark/scripts/benchmark_llama4_rope.py
benchmark/scripts/benchmark_mhc.py
benchmark/scripts/benchmark_mhc_lm.py
benchmark/scripts/benchmark_model_configs.py
benchmark/scripts/benchmark_multi_token_attention.py
benchmark/scripts/benchmark_orpo_loss.py
benchmark/scripts/benchmark_poly_norm.py
benchmark/scripts/benchmark_qwen2vl_mrope.py
benchmark/scripts/benchmark_relu_squared.py
benchmark/scripts/benchmark_rms_norm.py
benchmark/scripts/benchmark_rope.py
benchmark/scripts/benchmark_simpo_loss.py
benchmark/scripts/benchmark_softmax.py
benchmark/scripts/benchmark_sparse_multi_token_attention.py
benchmark/scripts/benchmark_sparsemax.py
benchmark/scripts/benchmark_swiglu.py
benchmark/scripts/benchmark_tiled_mlp.py
benchmark/scripts/benchmark_tvd.py
benchmark/scripts/utils.py
dev/fmt-requirements.txt
dev/modal/benchmarks.py
dev/modal/tests.py
docs/Examples.md
docs/Getting-Started.md
docs/High-Level-APIs.md
docs/Low-Level-APIs.md
docs/acknowledgement.md
docs/contributing.md
docs/index.md
docs/license.md
docs/images/banner.GIF
docs/images/compose.gif
docs/images/e2e-memory.png
docs/images/e2e-tps.png
docs/images/logo-banner.png
docs/images/patch.gif
docs/images/post-training.png
examples/alignment/accelerate_config.yaml
examples/alignment/run_orpo.py
examples/huggingface/README.md
examples/huggingface/callback.py
examples/huggingface/launch_on_modal.py
examples/huggingface/requirements.txt
examples/huggingface/run_benchmarks.sh
examples/huggingface/run_gemma.sh
examples/huggingface/run_llama.sh
examples/huggingface/run_qwen.sh
examples/huggingface/run_qwen2_vl.sh
examples/huggingface/training.py
examples/huggingface/training_multimodal.py
examples/huggingface/config/fsdp_config.json
examples/huggingface/img/gemma_7b_mem.png
examples/huggingface/img/gemma_7b_tp.png
examples/huggingface/img/llama_mem_alloc.png
examples/huggingface/img/llama_tps.png
examples/huggingface/img/qwen_mem_alloc.png
examples/huggingface/img/qwen_tps.png
examples/lightning/README.md
examples/lightning/requirements.txt
examples/lightning/training.py
examples/medusa/README.md
examples/medusa/callback.py
examples/medusa/medusa_util.py
examples/medusa/requirements.txt
examples/medusa/train.py
examples/medusa/docs/images/Memory_Stage1_num_head_3.png
examples/medusa/docs/images/Memory_Stage1_num_head_5.png
examples/medusa/docs/images/Memory_Stage2_num_head_3.png
examples/medusa/docs/images/Memory_Stage2_num_head_5.png
examples/medusa/docs/images/Throughput_Stage1_num_head_3.png
examples/medusa/docs/images/Throughput_Stage1_num_head_5.png
examples/medusa/docs/images/Throughput_Stage2_num_head_3.png
examples/medusa/docs/images/Throughput_Stage2_num_head_5.png
examples/medusa/fsdp/acc-fsdp.conf
examples/medusa/scripts/llama3_8b_medusa.sh
licenses/LICENSE-Apache-2.0
licenses/LICENSE-MIT-AutoAWQ
licenses/LICENSE-MIT-Efficient-Cross-Entropy
licenses/LICENSE-MIT-llmc
licenses/LICENSE-MIT-triton
src/liger_kernel/__init__.py
src/liger_kernel/env_report.py
src/liger_kernel/utils.py
src/liger_kernel/chunked_loss/README.md
src/liger_kernel/chunked_loss/__init__.py
src/liger_kernel/chunked_loss/cosine_similarity_loss.py
src/liger_kernel/chunked_loss/cpo_loss.py
src/liger_kernel/chunked_loss/dpo_loss.py
src/liger_kernel/chunked_loss/functional.py
src/liger_kernel/chunked_loss/fused_linear_distillation.py
src/liger_kernel/chunked_loss/fused_linear_ppo.py
src/liger_kernel/chunked_loss/fused_linear_preference.py
src/liger_kernel/chunked_loss/fused_linear_unpaired_preference.py
src/liger_kernel/chunked_loss/grpo_loss.py
src/liger_kernel/chunked_loss/jsd_loss.py
src/liger_kernel/chunked_loss/kto_loss.py
src/liger_kernel/chunked_loss/orpo_loss.py
src/liger_kernel/chunked_loss/simpo_loss.py
src/liger_kernel/ops/__init__.py
src/liger_kernel/ops/attn_res.py
src/liger_kernel/ops/cross_entropy.py
src/liger_kernel/ops/dyt.py
src/liger_kernel/ops/fused_add_rms_norm.py
src/liger_kernel/ops/fused_linear_cross_entropy.py
src/liger_kernel/ops/fused_linear_jsd.py
src/liger_kernel/ops/fused_moe.py
src/liger_kernel/ops/fused_moe_kernels.py
src/liger_kernel/ops/fused_neighborhood_attention.py
src/liger_kernel/ops/geglu.py
src/liger_kernel/ops/group_norm.py
src/liger_kernel/ops/grpo_loss.py
src/liger_kernel/ops/jsd.py
src/liger_kernel/ops/kl_div.py
src/liger_kernel/ops/layer_norm.py
src/liger_kernel/ops/llama4_rope.py
src/liger_kernel/ops/mhc.py
src/liger_kernel/ops/multi_token_attention.py
src/liger_kernel/ops/poly_norm.py
src/liger_kernel/ops/qwen2vl_mrope.py
src/liger_kernel/ops/relu_squared.py
src/liger_kernel/ops/rms_norm.py
src/liger_kernel/ops/rope.py
src/liger_kernel/ops/softmax.py
src/liger_kernel/ops/sparsemax.py
src/liger_kernel/ops/swiglu.py
src/liger_kernel/ops/tiled_mlp.py
src/liger_kernel/ops/tvd.py
src/liger_kernel/ops/utils.py
src/liger_kernel/ops/backends/README.md
src/liger_kernel/ops/backends/__init__.py
src/liger_kernel/ops/backends/registry.py
src/liger_kernel/ops/backends/_ascend/__init__.py
src/liger_kernel/ops/backends/_ascend/ascend-ub-manager-design.md
src/liger_kernel/ops/backends/_ascend/ub_manager.py
src/liger_kernel/ops/backends/_ascend/ops/__init__.py
src/liger_kernel/ops/backends/_ascend/ops/attn_res.py
src/liger_kernel/ops/backends/_ascend/ops/cross_entropy.py
src/liger_kernel/ops/backends/_ascend/ops/dyt.py
src/liger_kernel/ops/backends/_ascend/ops/embedding.py
src/liger_kernel/ops/backends/_ascend/ops/fused_add_rms_norm.py
src/liger_kernel/ops/backends/_ascend/ops/fused_linear_cross_entropy.py
src/liger_kernel/ops/backends/_ascend/ops/fused_linear_jsd.py
src/liger_kernel/ops/backends/_ascend/ops/fused_moe.py
src/liger_kernel/ops/backends/_ascend/ops/fused_moe_kernels.py
src/liger_kernel/ops/backends/_ascend/ops/fused_neighborhood_attention.py
src/liger_kernel/ops/backends/_ascend/ops/geglu.py
src/liger_kernel/ops/backends/_ascend/ops/group_norm.py
src/liger_kernel/ops/backends/_ascend/ops/grpo_loss.py
src/liger_kernel/ops/backends/_ascend/ops/jsd.py
src/liger_kernel/ops/backends/_ascend/ops/kl_div.py
src/liger_kernel/ops/backends/_ascend/ops/layer_norm.py
src/liger_kernel/ops/backends/_ascend/ops/llama4_rope.py
src/liger_kernel/ops/backends/_ascend/ops/mhc.py
src/liger_kernel/ops/backends/_ascend/ops/poly_norm.py
src/liger_kernel/ops/backends/_ascend/ops/qwen2vl_mrope.py
src/liger_kernel/ops/backends/_ascend/ops/rms_norm.py
src/liger_kernel/ops/backends/_ascend/ops/rope.py
src/liger_kernel/ops/backends/_ascend/ops/softmax.py
src/liger_kernel/ops/backends/_ascend/ops/sparsemax.py
src/liger_kernel/ops/backends/_ascend/ops/swiglu.py
src/liger_kernel/ops/backends/_ascend/ops/tvd.py
src/liger_kernel/ops/experimental/embedding.py
src/liger_kernel/ops/experimental/mm_int8int2.py
src/liger_kernel/transformers/__init__.py
src/liger_kernel/transformers/auto_model.py
src/liger_kernel/transformers/cross_entropy.py
src/liger_kernel/transformers/dyt.py
src/liger_kernel/transformers/fsdp.py
src/liger_kernel/transformers/functional.py
src/liger_kernel/transformers/fused_add_rms_norm.py
src/liger_kernel/transformers/fused_linear_cross_entropy.py
src/liger_kernel/transformers/fused_linear_jsd.py
src/liger_kernel/transformers/fused_neighborhood_attention.py
src/liger_kernel/transformers/geglu.py
src/liger_kernel/transformers/group_norm.py
src/liger_kernel/transformers/grpo_loss.py
src/liger_kernel/transformers/jsd.py
src/liger_kernel/transformers/kl_div.py
src/liger_kernel/transformers/layer_norm.py
src/liger_kernel/transformers/llama4_rope.py
src/liger_kernel/transformers/mhc.py
src/liger_kernel/transformers/monkey_patch.py
src/liger_kernel/transformers/multi_token_attention.py
src/liger_kernel/transformers/poly_norm.py
src/liger_kernel/transformers/qwen2vl_mrope.py
src/liger_kernel/transformers/relu_squared.py
src/liger_kernel/transformers/rms_norm.py
src/liger_kernel/transformers/rope.py
src/liger_kernel/transformers/softmax.py
src/liger_kernel/transformers/sparsemax.py
src/liger_kernel/transformers/swiglu.py
src/liger_kernel/transformers/tiled_mlp.py
src/liger_kernel/transformers/trainer_integration.py
src/liger_kernel/transformers/tvd.py
src/liger_kernel/transformers/experimental/__init__.py
src/liger_kernel/transformers/experimental/embedding.py
src/liger_kernel/transformers/model/__init__.py
src/liger_kernel/transformers/model/exaone4.py
src/liger_kernel/transformers/model/falcon_h1.py
src/liger_kernel/transformers/model/gemma.py
src/liger_kernel/transformers/model/gemma2.py
src/liger_kernel/transformers/model/gemma3.py
src/liger_kernel/transformers/model/gemma4.py
src/liger_kernel/transformers/model/glm4.py
src/liger_kernel/transformers/model/glm4v.py
src/liger_kernel/transformers/model/glm4v_moe.py
src/liger_kernel/transformers/model/gpt_oss.py
src/liger_kernel/transformers/model/hunyuan_v1.py
src/liger_kernel/transformers/model/internvl.py
src/liger_kernel/transformers/model/llama.py
src/liger_kernel/transformers/model/llama4.py
src/liger_kernel/transformers/model/llava.py
src/liger_kernel/transformers/model/loss_utils.py
src/liger_kernel/transformers/model/ministral.py
src/liger_kernel/transformers/model/mistral.py
src/liger_kernel/transformers/model/mixtral.py
src/liger_kernel/transformers/model/mllama.py
src/liger_kernel/transformers/model/nemotron.py
src/liger_kernel/transformers/model/olmo2.py
src/liger_kernel/transformers/model/olmo3.py
src/liger_kernel/transformers/model/output_classes.py
src/liger_kernel/transformers/model/paligemma.py
src/liger_kernel/transformers/model/phi3.py
src/liger_kernel/transformers/model/pixtral.py
src/liger_kernel/transformers/model/qwen2.py
src/liger_kernel/transformers/model/qwen2_5_vl.py
src/liger_kernel/transformers/model/qwen2_vl.py
src/liger_kernel/transformers/model/qwen3.py
src/liger_kernel/transformers/model/qwen3_5.py
src/liger_kernel/transformers/model/qwen3_5_moe.py
src/liger_kernel/transformers/model/qwen3_moe.py
src/liger_kernel/transformers/model/qwen3_next.py
src/liger_kernel/transformers/model/qwen3_vl.py
src/liger_kernel/transformers/model/qwen3_vl_moe.py
src/liger_kernel/transformers/model/smollm3.py
src/liger_kernel/transformers/model/smolvlm.py
src/liger_kernel/transformers/trainer/__init__.py
src/liger_kernel/transformers/trainer/orpo_trainer.py
src/liger_kernel/triton/__init__.py
src/liger_kernel/triton/monkey_patch.py
src/liger_kernel_nightly.egg-info/PKG-INFO
src/liger_kernel_nightly.egg-info/SOURCES.txt
src/liger_kernel_nightly.egg-info/dependency_links.txt
src/liger_kernel_nightly.egg-info/requires.txt
src/liger_kernel_nightly.egg-info/top_level.txt
test/__init__.py
test/conftest.py
test/utils.py
test/chunked_loss/__init__.py
test/chunked_loss/test_cosine_loss.py
test/chunked_loss/test_cpo_loss.py
test/chunked_loss/test_dpo_loss.py
test/chunked_loss/test_grpo_loss.py
test/chunked_loss/test_jsd_loss.py
test/chunked_loss/test_kto_loss.py
test/chunked_loss/test_orpo_loss.py
test/chunked_loss/test_simpo_loss.py
test/convergence/__init__.py
test/convergence/bf16/__init__.py
test/convergence/bf16/test_mini_models.py
test/convergence/bf16/test_mini_models_multimodal.py
test/convergence/bf16/test_mini_models_with_logits.py
test/convergence/fp32/__init__.py
test/convergence/fp32/test_mini_models.py
test/convergence/fp32/test_mini_models_multimodal.py
test/convergence/fp32/test_mini_models_with_logits.py
test/resources/tiny_shakespeare.txt
test/resources/fake_configs/Google/Gemma3/gemma-3-4b-it/tokenizer_config.json
test/resources/fake_configs/Google/Paligemma/paligemma-3b-pt-224/tokenizer_config.json
test/resources/fake_configs/HuggingFaceTB/SmolVLM2-256M-Video-Instruct/tokenizer_config.json
test/resources/fake_configs/Llava/llava-1.5-7b-hf/preprocessor_config.json
test/resources/fake_configs/Llava/llava-1.5-7b-hf/processor_config.json
test/resources/fake_configs/Llava/llava-1.5-7b-hf/tokenizer_config.json
test/resources/fake_configs/OpenGVLab/InternVL3-1B-hf/tokenizer_config.json
test/resources/fake_configs/Qwen/Qwen2-VL-7B-Instruct/tokenizer_config.json
test/resources/fake_configs/Qwen/Qwen2.5-VL-7B-Instruct/tokenizer_config.json
test/resources/fake_configs/Qwen/Qwen3-VL-4B-Instruct/tokenizer_config.json
test/resources/fake_configs/meta-llama/Llama-3.2-11B-Vision-Instruct/tokenizer_config.json
test/resources/fake_configs/meta-llama/Llama-4-Scout-17B-16E-Instruct/tokenizer_config.json
test/resources/scripts/generate_tokenized_dataset.py
test/resources/tiny_shakespeare_tokenized/data-00000-of-00001.arrow
test/resources/tiny_shakespeare_tokenized/dataset_info.json
test/resources/tiny_shakespeare_tokenized/state.json
test/transformers/test_attn_res.py
test/transformers/test_auto_model.py
test/transformers/test_cross_entropy.py
test/transformers/test_dyt.py
test/transformers/test_embedding.py
test/transformers/test_flex_attention.py
test/transformers/test_fused_add_rms_norm.py
test/transformers/test_fused_linear_cross_entropy.py
test/transformers/test_fused_linear_jsd.py
test/transformers/test_fused_moe.py
test/transformers/test_fused_neighborhood_attention.py
test/transformers/test_geglu.py
test/transformers/test_group_norm.py
test/transformers/test_grpo_loss.py
test/transformers/test_jsd.py
test/transformers/test_kl_div.py
test/transformers/test_layer_norm.py
test/transformers/test_llama4_rope.py
test/transformers/test_mhc.py
test/transformers/test_mm_int8int2.py
test/transformers/test_monkey_patch.py
test/transformers/test_multi_token_attention.py
test/transformers/test_poly_norm.py
test/transformers/test_qwen2vl_mrope.py
test/transformers/test_relu_squared.py
test/transformers/test_rms_norm.py
test/transformers/test_rope.py
test/transformers/test_softmax.py
test/transformers/test_sparsemax.py
test/transformers/test_swiglu.py
test/transformers/test_tiled_mlp.py
test/transformers/test_trainer_integration.py
test/transformers/test_transformers.py
test/transformers/test_tvd.py
test/triton/test_triton_monkey_patch.py