LICENSE
NOTICE
README.md
pyproject.toml
assets/version.txt
torchtitan_npu/__init__.py
torchtitan_npu/entry.py
torchtitan_npu/train.py
torchtitan_npu.egg-info/PKG-INFO
torchtitan_npu.egg-info/SOURCES.txt
torchtitan_npu.egg-info/dependency_links.txt
torchtitan_npu.egg-info/requires.txt
torchtitan_npu.egg-info/top_level.txt
torchtitan_npu/config/custom_config.py
torchtitan_npu/converters/__init__.py
torchtitan_npu/converters/base_converter.py
torchtitan_npu/converters/convert_utils.py
torchtitan_npu/converters/npu_converter.py
torchtitan_npu/converters/quant_converter.py
torchtitan_npu/converters/registry.py
torchtitan_npu/converters/features/bypass_triton_codegen.py
torchtitan_npu/converters/kernels/deepseek_v4_sfa.py
torchtitan_npu/converters/kernels/dsa.py
torchtitan_npu/converters/kernels/fusion_attention.py
torchtitan_npu/converters/kernels/gmm.py
torchtitan_npu/converters/kernels/mhc_prepost.py
torchtitan_npu/converters/kernels/permutation.py
torchtitan_npu/converters/kernels/permute.py
torchtitan_npu/converters/kernels/quant_gmm.py
torchtitan_npu/converters/kernels/quant_linear.py
torchtitan_npu/converters/kernels/rms_norm.py
torchtitan_npu/converters/kernels/rope.py
torchtitan_npu/distributed/context_parallel/__init__.py
torchtitan_npu/distributed/context_parallel/custom_context_parallel.py
torchtitan_npu/distributed/context_parallel/deepseek_v4_cp.py
torchtitan_npu/distributed/context_parallel/dsa_cp.py
torchtitan_npu/distributed/context_parallel/ulysses_cp.py
torchtitan_npu/models/common/__init__.py
torchtitan_npu/models/common/dsa_indexer_loss.py
torchtitan_npu/models/deepseek_v3/__init__.py
torchtitan_npu/models/deepseek_v3/infra/parallelize.py
torchtitan_npu/models/deepseek_v3/model/args.py
torchtitan_npu/models/deepseek_v3/model/state_dict_adapter.py
torchtitan_npu/models/deepseek_v32/__init__.py
torchtitan_npu/models/deepseek_v32/infra/parallelize.py
torchtitan_npu/models/deepseek_v32/model/args.py
torchtitan_npu/models/deepseek_v32/model/model.py
torchtitan_npu/models/deepseek_v32/model/state_dict_adapter.py
torchtitan_npu/models/deepseek_v4/__init__.py
torchtitan_npu/models/deepseek_v4/text_datasets.py
torchtitan_npu/models/deepseek_v4/infra/parallelize.py
torchtitan_npu/models/deepseek_v4/model/args.py
torchtitan_npu/models/deepseek_v4/model/model.py
torchtitan_npu/models/deepseek_v4/model/moe.py
torchtitan_npu/models/deepseek_v4/model/state_dict_adapter.py
torchtitan_npu/models/llama4/model/state_dict_adapter.py
torchtitan_npu/ops/aclnn/__init__.py
torchtitan_npu/ops/aclnn/builder.py
torchtitan_npu/ops/triton/__init__.py
torchtitan_npu/ops/triton/add.py
torchtitan_npu/ops/triton/mhc_triton.py
torchtitan_npu/ops/triton/post_bmm1.py
torchtitan_npu/ops/triton/post_bmm2.py
torchtitan_npu/ops/triton/pre_bmm.py
torchtitan_npu/ops/triton/prepost_sinkhorn.py
torchtitan_npu/patches/distributed/cp_input_sharding.py
torchtitan_npu/patches/distributed/custom_context_parallel.py
torchtitan_npu/patches/distributed/mtp_context_parallel.py
torchtitan_npu/patches/distributed/utils.py
torchtitan_npu/patches/optimizer/muon_optimizer.py
torchtitan_npu/patches/optimizer/swap_muon_optimizer.py
torchtitan_npu/patches/optimizer/swap_optimizer.py
torchtitan_npu/patches/optimizer/virtual_allocator.py
torchtitan_npu/patches/optimizer/virtual_optimizer.py
torchtitan_npu/patches/quantization/quant_config.py
torchtitan_npu/patches/quantization/quantize.py
torchtitan_npu/patches/tools/metrics.py
torchtitan_npu/patches/torch/checkpoint.py
torchtitan_npu/patches/torch/clip_grad.py
torchtitan_npu/patches/torch/micro_pipeline_tp.py
torchtitan_npu/patches/torch/pipelining.py
torchtitan_npu/patches/torch/_inductor/graph.py
torchtitan_npu/patches/torch/testing/_internal/distributed/fake_pg.py
torchtitan_npu/patches/torch_npu/_meta_registrations.py
torchtitan_npu/patches/torch_npu/custom_shardings.py
torchtitan_npu/patches/torch_npu/_inductor/lowering.py
torchtitan_npu/patches/torchtitan/hf_datasets.py
torchtitan_npu/patches/torchtitan/loss.py
torchtitan_npu/patches/torchtitan/lr_scheduler.py
torchtitan_npu/tools/checkpoint_patch.py
torchtitan_npu/tools/flight_recorder.py
torchtitan_npu/tools/profiling.py
torchtitan_npu/tools/weight_utils.py