LICENSE
MANIFEST.in
README.md
pyproject.toml
requirements-ascend.txt
requirements-dev.txt
requirements-nvidia.txt
requirements.txt
setup.py
csrc/asc_utils.cpp
csrc/cuda_utils.cpp
csrc/cuda_utils.hpp
csrc/pybind.cpp
csrc/engine/block_manager.hpp
csrc/engine/flops_estimator.hpp
csrc/engine/mtp_varlen.hpp
csrc/engine/outputs.hpp
csrc/engine/prepare_sample.cpp
csrc/engine/prepare_sample.hpp
csrc/engine/process_output.hpp
csrc/engine/pybind.cpp
csrc/engine/pybind.hpp
csrc/engine/pybind_output.cpp
csrc/engine/pybind_scheduler.cpp
csrc/engine/pybind_sequence.cpp
csrc/engine/scheduler.hpp
csrc/engine/sequence.hpp
csrc/engine/utils.hpp
csrc/engine/zos.cpp
csrc/engine/zos.hpp
csrc/engine/input_metadata/block_load_plan.cpp
csrc/engine/input_metadata/block_load_plan.hpp
csrc/engine/input_metadata/prepare_tensor_input.hpp
csrc/engine/input_metadata/prepare_tensor_input_base.cpp
csrc/engine/input_metadata/prepare_tensor_input_v4.cpp
csrc/graph_timer/graph_timer.hpp
csrc/graph_timer/graph_timer_asc.cpp
csrc/graph_timer/graph_timer_cuda.cpp
csrc/graph_timer/pybind.cpp
csrc/utils/lockable.hpp
csrc/utils/log.hpp
csrc/utils/object_pool.hpp
deep_cluster/__init__.py
deep_cluster/hardware/__init__.py
deep_cluster/hardware/ascend.py
deep_cluster/hardware/cpu.py
deep_cluster/hardware/device.py
deep_cluster/hardware/nvidia.py
deep_cluster/topology/__init__.py
deep_cluster/topology/bind.py
deepinfer/__init__.py
deepinfer/attrdict.py
deepinfer/config.py
deepinfer/environ.py
deepinfer/load_class.py
deepinfer/logger.py
deepinfer/model_config_loader.py
deepinfer/multimodal_tokens.py
deepinfer/platform.py
deepinfer/py.typed
deepinfer/sampling_params.py
deepinfer/sequence.py
deepinfer/shared_memory.py
deepinfer/testing.py
deepinfer/timer.py
deepinfer/utils.py
deepinfer.egg-info/PKG-INFO
deepinfer.egg-info/SOURCES.txt
deepinfer.egg-info/dependency_links.txt
deepinfer.egg-info/requires.txt
deepinfer.egg-info/top_level.txt
deepinfer/_C/__init__.pyi
deepinfer/_C/cuda_utils.pyi
deepinfer/_C/engine.pyi
deepinfer/_C/graph_timer.pyi
deepinfer/distributed/__init__.py
deepinfer/distributed/shm_broadcast.py
deepinfer/engine/__init__.py
deepinfer/engine/arg_utils.py
deepinfer/engine/engine_type.py
deepinfer/engine/llm_engine.py
deepinfer/engine/profile_server.py
deepinfer/engine/stats_logger.py
deepinfer/entrypoints/__init__.py
deepinfer/entrypoints/llm.py
deepinfer/hugging_face/__init__.py
deepinfer/hugging_face/config.py
deepinfer/hugging_face/loading.py
deepinfer/hugging_face/names.py
deepinfer/hugging_face/parameters.py
deepinfer/hugging_face/processor.py
deepinfer/hugging_face/tokenizer.py
deepinfer/hugging_face/weights.py
deepinfer/model_executor/__init__.py
deepinfer/model_executor/layer_utils.py
deepinfer/model_executor/model_loader.py
deepinfer/model_executor/model_metadata_utils.py
deepinfer/model_executor/model_output.py
deepinfer/model_executor/sampling_inputs.py
deepinfer/model_executor/utils.py
deepinfer/model_executor/eplb/__init__.py
deepinfer/model_executor/eplb/policy.py
deepinfer/model_executor/eplb/shuffle.py
deepinfer/model_executor/input_metadata/__init__.py
deepinfer/model_executor/input_metadata/dflash.py
deepinfer/model_executor/input_metadata/input_metadata.py
deepinfer/model_executor/input_metadata/prepare_tensor_input_zos.py
deepinfer/model_executor/input_metadata/yoco.py
deepinfer/model_executor/layers/__init__.py
deepinfer/model_executor/layers/attention_backend.py
deepinfer/model_executor/layers/layernorm.py
deepinfer/model_executor/layers/linear.py
deepinfer/model_executor/layers/mtp_info_pool.py
deepinfer/model_executor/layers/rotary_embedding.py
deepinfer/model_executor/layers/sampler.py
deepinfer/model_executor/layers/sampler_utils.py
deepinfer/model_executor/layers/vocab_parallel_embedding.py
deepinfer/model_executor/layers/tilelang_ops/__init__.py
deepinfer/model_executor/layers/tilelang_ops/cache_common.py
deepinfer/model_executor/layers/tilelang_ops/candidate_blocks.py
deepinfer/model_executor/layers/tilelang_ops/common.py
deepinfer/model_executor/layers/tilelang_ops/compress.py
deepinfer/model_executor/layers/tilelang_ops/compute_seqlens_k_zos.py
deepinfer/model_executor/layers/tilelang_ops/conftest.py
deepinfer/model_executor/layers/tilelang_ops/dflash.py
deepinfer/model_executor/layers/tilelang_ops/dflash_copy.py
deepinfer/model_executor/layers/tilelang_ops/flash_mla_utils.py
deepinfer/model_executor/layers/tilelang_ops/fused_sample.py
deepinfer/model_executor/layers/tilelang_ops/fused_transform.py
deepinfer/model_executor/layers/tilelang_ops/gather.py
deepinfer/model_executor/layers/tilelang_ops/gather_engram_tokens_zos.py
deepinfer/model_executor/layers/tilelang_ops/gather_input_ids_zos.py
deepinfer/model_executor/layers/tilelang_ops/get_kv_index.py
deepinfer/model_executor/layers/tilelang_ops/image_embed_ops.py
deepinfer/model_executor/layers/tilelang_ops/kv_index_cat.py
deepinfer/model_executor/layers/tilelang_ops/load_paged_kv_cache.py
deepinfer/model_executor/layers/tilelang_ops/norm.py
deepinfer/model_executor/layers/tilelang_ops/ops.py
deepinfer/model_executor/layers/tilelang_ops/outlier_detection.py
deepinfer/model_executor/layers/tilelang_ops/prepare_cp_mirror.py
deepinfer/model_executor/layers/tilelang_ops/prepare_v4_decode_metadata.py
deepinfer/model_executor/layers/tilelang_ops/random_sampling.py
deepinfer/model_executor/layers/tilelang_ops/sampler.py
deepinfer/model_executor/layers/tilelang_ops/schedule_mtp_varlen.py
deepinfer/model_executor/layers/tilelang_ops/state_cache.py
deepinfer/model_executor/layers/tilelang_ops/store_kv_cache.py
deepinfer/model_executor/layers/tilelang_ops/store_mtp_info.py
deepinfer/model_executor/layers/tilelang_ops/structure_output.py
deepinfer/model_executor/layers/tilelang_ops/sum_smaller_probs.py
deepinfer/model_executor/layers/tilelang_ops/swa_index.py
deepinfer/model_executor/layers/tilelang_ops/asc/__init__.py
deepinfer/model_executor/layers/tilelang_ops/asc/candidate_blocks.py
deepinfer/model_executor/layers/tilelang_ops/asc/compress.py
deepinfer/model_executor/layers/tilelang_ops/asc/compute_seqlens_k_zos.py
deepinfer/model_executor/layers/tilelang_ops/asc/dflash.py
deepinfer/model_executor/layers/tilelang_ops/asc/dflash_copy.py
deepinfer/model_executor/layers/tilelang_ops/asc/flash_mla_utils.py
deepinfer/model_executor/layers/tilelang_ops/asc/fused_sample.py
deepinfer/model_executor/layers/tilelang_ops/asc/fused_transform.py
deepinfer/model_executor/layers/tilelang_ops/asc/gather.py
deepinfer/model_executor/layers/tilelang_ops/asc/gather_engram_tokens_zos.py
deepinfer/model_executor/layers/tilelang_ops/asc/gather_input_ids_zos.py
deepinfer/model_executor/layers/tilelang_ops/asc/get_kv_index.py
deepinfer/model_executor/layers/tilelang_ops/asc/image_embed_ops.py
deepinfer/model_executor/layers/tilelang_ops/asc/kv_index_cat.py
deepinfer/model_executor/layers/tilelang_ops/asc/kv_quant.py
deepinfer/model_executor/layers/tilelang_ops/asc/load_paged_kv_cache.py
deepinfer/model_executor/layers/tilelang_ops/asc/norm.py
deepinfer/model_executor/layers/tilelang_ops/asc/prepare_cp_mirror.py
deepinfer/model_executor/layers/tilelang_ops/asc/prepare_v4_decode_metadata.py
deepinfer/model_executor/layers/tilelang_ops/asc/random_sampling.py
deepinfer/model_executor/layers/tilelang_ops/asc/rstd_1ulp.py
deepinfer/model_executor/layers/tilelang_ops/asc/sampler.py
deepinfer/model_executor/layers/tilelang_ops/asc/schedule_mtp_varlen.py
deepinfer/model_executor/layers/tilelang_ops/asc/state_cache.py
deepinfer/model_executor/layers/tilelang_ops/asc/store_kv_cache.py
deepinfer/model_executor/layers/tilelang_ops/asc/store_mtp_info.py
deepinfer/model_executor/layers/tilelang_ops/asc/structure_output.py
deepinfer/model_executor/layers/tilelang_ops/asc/sum_smaller_probs.py
deepinfer/model_executor/layers/tilelang_ops/asc/swa_index.py
deepinfer/model_executor/layers/tilelang_ops/cuda/__init__.py
deepinfer/model_executor/layers/tilelang_ops/cuda/candidate_blocks.py
deepinfer/model_executor/layers/tilelang_ops/cuda/compress.py
deepinfer/model_executor/layers/tilelang_ops/cuda/compute_seqlens_k_zos.py
deepinfer/model_executor/layers/tilelang_ops/cuda/dflash.py
deepinfer/model_executor/layers/tilelang_ops/cuda/dflash_copy.py
deepinfer/model_executor/layers/tilelang_ops/cuda/flash_mla_utils.py
deepinfer/model_executor/layers/tilelang_ops/cuda/fused_sample.py
deepinfer/model_executor/layers/tilelang_ops/cuda/fused_transform.py
deepinfer/model_executor/layers/tilelang_ops/cuda/gather.py
deepinfer/model_executor/layers/tilelang_ops/cuda/gather_engram_tokens_zos.py
deepinfer/model_executor/layers/tilelang_ops/cuda/gather_input_ids_zos.py
deepinfer/model_executor/layers/tilelang_ops/cuda/get_kv_index.py
deepinfer/model_executor/layers/tilelang_ops/cuda/image_embed_ops.py
deepinfer/model_executor/layers/tilelang_ops/cuda/kv_index_cat.py
deepinfer/model_executor/layers/tilelang_ops/cuda/load_paged_kv_cache.py
deepinfer/model_executor/layers/tilelang_ops/cuda/norm.py
deepinfer/model_executor/layers/tilelang_ops/cuda/outlier_detection.py
deepinfer/model_executor/layers/tilelang_ops/cuda/prepare_cp_mirror.py
deepinfer/model_executor/layers/tilelang_ops/cuda/prepare_decode_metadata.py
deepinfer/model_executor/layers/tilelang_ops/cuda/quant.py
deepinfer/model_executor/layers/tilelang_ops/cuda/random_sampling.py
deepinfer/model_executor/layers/tilelang_ops/cuda/sampler.py
deepinfer/model_executor/layers/tilelang_ops/cuda/schedule_mtp_varlen.py
deepinfer/model_executor/layers/tilelang_ops/cuda/state_cache.py
deepinfer/model_executor/layers/tilelang_ops/cuda/store_kv_cache.py
deepinfer/model_executor/layers/tilelang_ops/cuda/store_mtp_info.py
deepinfer/model_executor/layers/tilelang_ops/cuda/structure_output.py
deepinfer/model_executor/layers/tilelang_ops/cuda/sum_smaller_probs.py
deepinfer/model_executor/layers/tilelang_ops/cuda/swa_index.py
deepinfer/model_executor/models/__init__.py
deepinfer/model_executor/models/decoder_layer.py
deepinfer/model_executor/models/deepseek.py
deepinfer/model_executor/models/deepseek_vl.py
deepinfer/model_executor/models/mtp/__init__.py
deepinfer/model_executor/models/mtp/dflash.py
deepinfer/model_executor/models/v4_modules/__init__.py
deepinfer/model_executor/models/v4_modules/attn.py
deepinfer/model_executor/models/v4_modules/builder.py
deepinfer/model_executor/models/v4_modules/compress.py
deepinfer/model_executor/models/v4_modules/containers.py
deepinfer/model_executor/models/v4_modules/cross_layer.py
deepinfer/model_executor/models/v4_modules/engram.py
deepinfer/model_executor/models/v4_modules/engram_metadata.py
deepinfer/model_executor/models/v4_modules/engram_parameters.py
deepinfer/model_executor/models/v4_modules/hc.py
deepinfer/model_executor/models/v4_modules/hc_blocks.py
deepinfer/model_executor/models/v4_modules/hc_lazy.py
deepinfer/model_executor/models/v4_modules/heads.py
deepinfer/model_executor/models/v4_modules/index_attn.py
deepinfer/model_executor/models/v4_modules/linear.py
deepinfer/model_executor/models/v4_modules/mhc_modeling.py
deepinfer/model_executor/models/v4_modules/mhc_ops.py
deepinfer/model_executor/models/v4_modules/mla_attn.py
deepinfer/model_executor/models/v4_modules/mlp.py
deepinfer/model_executor/models/v4_modules/moe.py
deepinfer/model_executor/models/v4_modules/norm.py
deepinfer/model_executor/models/v4_modules/rope.py
deepinfer/model_executor/models/v4_modules/utils.py
deepinfer/model_executor/models/v4_modules/yoco.py
deepinfer/model_executor/models/vision_encoder/__init__.py
deepinfer/model_executor/models/vision_encoder/mlp_projector.py
deepinfer/model_executor/models/vision_encoder/vit.py
deepinfer/model_executor/models/vision_encoder/vl_aligner.py
deepinfer/model_executor/parallel_utils/__init__.py
deepinfer/model_executor/parallel_utils/communication_op.py
deepinfer/model_executor/parallel_utils/cp.py
deepinfer/model_executor/parallel_utils/nccl_memory.py
deepinfer/model_executor/parallel_utils/parallel_state.py
deepinfer/model_executor/parallel_utils/userbuffer.py
deepinfer/npu_ops/__init__.py
deepinfer/npu_ops/deep_select.py
deepinfer/npu_ops/patch_imports.py
deepinfer/npu_ops/vit_attn.py
deepinfer/structure_output/__init__.py
deepinfer/structure_output/compiler.py
deepinfer/structure_output/mask.py
deepinfer/structure_output/matcher.py
deepinfer/structure_output/so_matcher_builder.py
deepinfer/ve/__init__.py
deepinfer/ve/multimodal_processor.py
deepinfer/ve/prepare_tensor_input_vl.py
deepinfer/ve/schema.py
deepinfer/ve/processor/__init__.py
deepinfer/ve/processor/common.py
deepinfer/ve/processor/image_processor.py
deepinfer/ve/processor/mm_processor.py
deepinfer/ve/processor/text_processor.py
deepinfer/worker/__init__.py
deepinfer/worker/base.py
deepinfer/worker/cache_engine.py
deepinfer/worker/cuda_graph_runner.py
deepinfer/worker/decode_perf_cache.py
deepinfer/worker/dflash_cuda_graph_runner.py
deepinfer/worker/model_runner.py
deepinfer/worker/runtime_config.py
deepinfer/worker/worker.py
deepinfer/worker/zos_status.py
tests/test_cuda_sync_guards.py