AA_David@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/attention/flash_attention_score/op_host/arch35
- ops/ops-transformer/attention/flash_attention_score/op_kernel/arch35
- ops/ops-transformer/attention/flash_attention_score/op_kernel/flash_attention_score_apt.cpp
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_div_cast.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_flashupdate_new.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_mul_sel_softmaxflashv2_cast_nz_dn.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_basic_block_aligned128_no_update.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_basic_block_aligned128_update.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_basic_block_unaligned64_no_update.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_basic_block_unaligned64_update.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_basic_block_unaligned128_no_update.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_basic_block_unaligned128_update.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_basic_block_unaligned256_no_update.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_basic_block_unaligned256_update.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_basic_block_utils.h
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/vf_mul_sel_softmaxflashv2_cast_nz.h
- ops/ops-transformer/attention/common/op_kernel/arch35/attenmask.h
- ops/ops-transformer/attention/common/op_kernel/arch35/dropmask.h
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_score_block_vec_train.h
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_score_kernel_base.h
- ops/ops-transformer/attention/common/op_kernel/arch35/infer_flash_attention_comm.h
- ops/ops-transformer/attention/common/op_kernel/arch35/pse.h
- ops/ops-transformer/attention/common/op_kernel/arch35/util_regbase.h
- ops/ops-transformer/attention/common/op_kernel/offset_calculator.h
opensource_style: null
kernel_style: null
unrelease:
test_code: null
non_delivery: null
open_source: null
AAG_David@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/attention/flash_attention_score_grad/op_host/arch35
- ops/ops-transformer/attention/flash_attention_score_grad/op_kernel/arch35
opensource_style: null
kernel_style: null
unrelease:
test_code: null
non_delivery: null
open_source: null
transformer-1@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/attention/incre_flash_attention/ophost/
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/ifa_public_define.h
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/incre_flash_attention.cpp
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/incre_flash_attention_allvec_new.h
- ops/ops-transformer/attention/incre_flash_attention/regbase/
- ops/ops-transformer/attention/common/op_host/arch22/
- ops/ops-transformer/attention/common/op_host/fia_tiling_info.h
- ops/ops-transformer/attention/common/op_host/fia_tiling_info.cpp
- ops/ops-transformer/attention/common/op_host/fia_tiling_base.h
- ops/ops-transformer/attention/common/op_host/fia_tiling_shape.cpp
- ops/ops-transformer/attention/common/op_host/fia_tiling_templates_registry.h
- ops/ops-transformer/attention/common/op_kernel/arch22/
- ops/ops-transformer/attention/common/op_kernel/buffers_policy.h
- ops/ops-transformer/attention/common/op_kernel/buffer_manager.h
- ops/ops-transformer/attention/common/op_kernel/buffer.h
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_kvsame_bn2gs1s2.h
- ops/ops-transformer/attention/common/op_kernel/arch35/infer_flash_attention_sparse.h
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_score_block_vec_infer.h
- ops/ops-transformer/attention/common/op_kernel/arch35/infer_flash_attention_kvcache.h
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_score_kernel_infer.h
opensource_style: null
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/attention/incre_flash_attention/README.md
- ops/ops-transformer/attention/incre_flash_attention/CMakeLists.txt
- ops/ops-transformer/attention/incre_flash_attention/tests/
- ops/ops-transformer/attention/incre_flash_attention/docs/
- ops/ops-transformer/attention/incre_flash_attention/examples/
non_delivery: null
open_source: null
transformer-2@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/attention/common/op_kernel/fia_public_define.h
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_score_common_regbase.h
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_score_tiling_regbase.h
- ops/ops-transformer/attention/common/op_kernel/vector_common.h
- ops/ops-transformer/attention/lightning_indexer/
- ops/ops-transformer/attention/sparse_flash_attention/
- ops/ops-transformer/attention/quant_lightning_indexer/
- ops/ops-transformer/attention/kv_quant_sparse_flash_attention/
opensource_style: null
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/attention/lightning_indexer/tests/
- ops/ops-transformer/attention/lightning_indexer/docs/
- ops/ops-transformer/attention/lightning_indexer/examples/
- ops/ops-transformer/attention/sparse_flash_attention/tests/
- ops/ops-transformer/attention/sparse_flash_attention/docs/
- ops/ops-transformer/attention/sparse_flash_attention/examples/
- ops/ops-transformer/attention/quant_lightning_indexer/tests/
- ops/ops-transformer/attention/quant_lightning_indexer/docs/
- ops/ops-transformer/attention/quant_lightning_indexer/examples/
- ops/ops-transformer/attention/kv_quant_sparse_flash_attention/tests/
- ops/ops-transformer/attention/kv_quant_sparse_flash_attention/docs/
- ops/ops-transformer/attention/kv_quant_sparse_flash_attention/examples/
non_delivery: null
open_source: null
transformer-pfa_unevolved@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/attention/prompt_flash_attention/op_host/
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/prompt_flash_attention.cpp
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_s1s2_bns1_mla.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_s1s2_bns1_x910_base.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_bnstilling_n_s_tail.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/kernel_data_copy_transpose.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/mla_custom_matmul_policy_d192.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_split_n_s_no_tail.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_var_len_score_sab.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/mla_common.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_bnstilling_n_s_tailWBNSD.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/kernel_operator_softmax_compute_nz.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_s1s2_bns1_mla_baseapi.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/mla_custom_matmul_policy_common.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_bnstilling_n_s_no_tailWBNSD.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/mla_custom_matmul_policy_d128.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_s1s2_bns1_x910.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_empty_tensor.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_var_len_score_sab_baseapi.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_bnstilling_n_s_no_tail.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/pfa_service_matmul_kv_nd.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/prompt_flash_attention_tilingkey.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch35/
- ops/ops-transformer/attention/prompt_flash_attention/README.md
- ops/ops-transformer/attention/prompt_flash_attention/CMakeLists.txt
opensource_style: null
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/attention/prompt_flash_attention/tests/
- ops/ops-transformer/attention/prompt_flash_attention/examples/
- ops/ops-transformer/attention/prompt_flash_attention/docs/
non_delivery: null
open_source: null
transformer-vf@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/attention/common/op_kernel/arch35/vf/
- ops/ops-transformer/attention/lightning_indexer/op_kernel/arch35/vf
- ops/ops-transformer/attention/quant_lightning_indexer/op_kernel/arch35/vf
opensource_style: null
kernel_style: null
unrelease:
test_code: null
non_delivery: null
open_source: null
transformer-3@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/attention/mla_prolog_v2/op_host
- ops/ops-transformer/attention/mla_prolog_v2/CMakeLists.txt
- ops/ops-transformer/attention/mla_prolog_v2/op_kernel
- ops/ops-transformer/attention/mla_prolog_v2/op_graph
- ops/ops-transformer/attention/mla_prolog/op_host
- ops/ops-transformer/attention/mla_prolog/CMakeLists.txt
- ops/ops-transformer/attention/mla_prolog/op_kernel
- ops/ops-transformer/attention/mla_prolog/op_graph
- ops/ops-transformer/attention/common/op_host/split_core.h
- ops/ops-transformer/attention/common/op_kernel/matmul.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/arch22
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/arch35
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/op_api
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/CMakeLists.txt
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/fallback_fused_infer_attention_score.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/fused_infer_attention_score_def.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/fused_infer_attention_score_infershape.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/fused_infer_attention_score_tiling.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/fused_infer_attention_score_tiling.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/fused_infer_attention_score_tiling_compile_info.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/fused_infer_attention_score_tiling_index.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/fused_infer_attention_score_tiling_register.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_kernel/arch22
- ops/ops-transformer/attention/fused_infer_attention_score/op_kernel/fused_infer_attention_score.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_kernel/fused_infer_attention_score_v3.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/CMakeLists.txt
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/arch22/incre_flash_attention_split_Bbn2s2_Us2.h
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/arch22/incre_flash_attention_preload_dd.h
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/arch22/incre_flash_attention_preload.h
opensource_style: null
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/attention/mla_prolog_v2/tests
- ops/ops-transformer/attention/mla_prolog_v2/examples
- ops/ops-transformer/attention/mla_prolog_v2/README.md
- ops/ops-transformer/attention/mla_prolog_v2/docs
- ops/ops-transformer/attention/mla_prolog/tests
- ops/ops-transformer/attention/mla_prolog/examples
- ops/ops-transformer/attention/mla_prolog/README.md
- ops/ops-transformer/attention/mla_prolog/docs
- ops/ops-transformer/attention/fused_infer_attention_score/tests
- ops/ops-transformer/attention/fused_infer_attention_score/README.md
- ops/ops-transformer/attention/fused_infer_attention_score/examples
- ops/ops-transformer/attention/fused_infer_attention_score/docs
non_delivery: null
open_source: null
transformer-5@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_score_antiquant_block_cube.h
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_score_antiquant_block_vec.h
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_score_antiquant_kernel.h
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_score_antiquant_processor.h
- ops/ops-transformer/attention/common/op_kernel/arch35/flash_attention_score_antiquant_baseapi.h
- ops/ops-transformer/attention/common/op_kernel/array.h
- ops/ops-transformer/attention/common/op_kernel/memory_copy.h
- ops/ops-transformer/attention/common/op_kernel/CopyInL1.h
- ops/ops-transformer/attention/common/op_kernel/FixpipeOut.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/arch35/fused_infer_attention_score_tiling_impl.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/arch35/fused_infer_attention_score_tiling_impl.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/arch35/fused_infer_attention_score_tiling_v4.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/arch35/fused_infer_attention_score_tiling_v4.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/base_checker.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/base_checker.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/dequant_checker.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/dequant_checker.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/mask_checker.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/mask_checker.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/paged_attention_checker.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/paged_attention_checker.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/post_quant_checker.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/post_quant_checker.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/rope_checker.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/rope_checker.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/softmax_lse_checker.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_host/checkers/softmax_lse_checker.h
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/arch22/incre_flash_attention_preload_mla.h
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/arch22/ifa_service_matmul_full_quant.h
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/arch35/service/antiquant_preprocessor_q.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch35/vf/vf_antiquant.h
opensource_style: null
kernel_style: null
unrelease:
test_code: null
non_delivery: null
open_source: null
ifa@ops-transformer:
src:
release:
huawei_style: null
opensource_style:
- ops/ops-transformer/attention/incre_flash_attention/op_host/op_api
- ops/ops-transformer/attention/incre_flash_attention/op_host/fallback_incre_flash_attention.cpp
- ops/ops-transformer/attention/incre_flash_attention/op_host/incre_flash_attention_def.cpp
- ops/ops-transformer/attention/incre_flash_attention/op_host/incre_flash_attention_infershape.cpp
- ops/ops-transformer/attention/incre_flash_attention/op_host/incre_flash_attention_tiling_base.h
- ops/ops-transformer/attention/incre_flash_attention/op_host/incre_flash_attention_tiling_check.cpp
- ops/ops-transformer/attention/incre_flash_attention/op_host/incre_flash_attention_tiling_context.h
- ops/ops-transformer/attention/incre_flash_attention/op_host/incre_flash_attention_tiling_impl.h
- ops/ops-transformer/attention/incre_flash_attention/op_host/incre_flash_attention_tiling_mla.h
- ops/ops-transformer/attention/incre_flash_attention/op_host/incre_flash_attention_tiling_register.cpp
- ops/ops-transformer/attention/incre_flash_attention/op_host/incre_flash_attention_tiling_struct.h
- ops/ops-transformer/attention/incre_flash_attention/op_host/incre_flash_attention_tiling.cpp
- ops/ops-transformer/attention/incre_flash_attention/op_host/incre_flash_attention_tiling.h
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/arch20
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/arch22/paged_attention_antiquantkv.h
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/incre_flash_attention_tilingkey.h
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/incre_flash_attention_obp.h
- ops/ops-transformer/attention/nsa_compress_with_cache/op_host
- ops/ops-transformer/attention/nsa_compress_with_cache/op_kernel
- ops/ops-transformer/attention/nsa_selected_attention_infer/op_host
- ops/ops-transformer/attention/nsa_selected_attention_infer/op_kernel
- ops/ops-transformer/attention/nsa_compress_attention_infer/op_host
- ops/ops-transformer/attention/nsa_compress_attention_infer/op_kernel
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/attention/incre_flash_attention/op_kernel/arch35
- ops/ops-transformer/attention/incre_flash_attention/examples/test_aclnn_incre_flash_attention.cpp
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/incre_flash_attention_fusion_pass_utils.cc
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/incre_flash_attention_fusion_pass_utils.h
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/incre_flash_attention_quant_deploy_pass.cc
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/incre_flash_attention_quant_deploy_pass.h
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/incre_flash_attention_quant_fusion_pass.h
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/incre_flash_attention_quant_fusion_pass.cc
non_delivery: null
open_source: null
pfa@ops-transformer:
src:
release:
huawei_style: null
opensource_style:
- ops/ops-transformer/attention/prompt_flash_attention/op_host
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/prompt_flash_attention_obp.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/prompt_flash_attention_310p.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/prompt_attention_prefill.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/unpad_flash_attention_common.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_base_api.h
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch22/prompt_flash_attention_base_common.h
- ops/ops-transformer/attention/prompt_flash_attention/examples/test_aclnn_prompt_flash_attention.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_kernel/attn_infra
- ops/ops-transformer/attention/fused_infer_attention_score/op_kernel/flash_attention_interface.cpp
- ops/ops-transformer/attention/fused_infer_attention_score/op_kernel/flash_attention_regular_decode.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_kernel/flash_attention_regular.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_kernel/fused_infer_attention_score_template_tiling_key.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_kernel/fused_infer_attention_score_tilingkey.h
- ops/ops-transformer/attention/fused_infer_attention_score/op_kernel/kernel_common.hpp
- ops/ops-transformer/attention/fused_infer_attention_score/examples/test_aclnn_fused_infer_attention_score.cpp
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch35
- ops/ops-transformer/attention/prompt_flash_attention/op_kernel/arch38
- ops/ops-transformer/attention/prompt_flash_attention/op_host/prompt_flash_attention_tiling_v2.cpp
- ops/ops-transformer/attention/prompt_flash_attention/op_host/prompt_flash_attention_tiling_arch38.cpp
- asl/ops/ops_adv/tests/ut/ops_test/src/transformer/prompt_flash_attention/
- asl/ops/ops_adv/tests/ut/ops_test/src/transformer/fused_infer_attention_score/
- asl/ops/cann/ops/built-in/tests/ut/op_tiling_test/test_fused_infer_attention_score_tiling.cpp
- asl/ops/cann/ops/built-in/op_tiling/runtime/fused_infer_attention_score_tiling_host.cpp
- asl/ops/cann/ops/built-in/framework/onnx_plugin/custom/npu_prompt_flash_attention_plugin.cc
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/prompt_flash_attention_quant_deploy_pass.cc
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/prompt_flash_attention_quant_deploy_pass.h
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/prompt_flash_attention_quant_fusion_pass.cc
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/prompt_flash_attention_quant_fusion_pass.h
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/fused_infer_attention_score_quant_deploy_pass.cc
- asl/ops/cann/ops/built-in/fusion_pass/graph_fusion/ai_core/fused_infer_attention_score_quant_fusion_pass.cc
non_delivery: null
open_source: null
gmm@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/gmm/grouped_matmul
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2
- ops/ops-transformer/gmm/quant_grouped_matmul_inplace_add
opensource_style: null
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/gmm/grouped_matmul/tests
- ops/ops-transformer/gmm/grouped_matmul/examples
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant/tests
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant/examples
- ops/ops-transformer/gmm/quant_grouped_matmul_inplace_add/tests
- ops/ops-transformer/gmm/quant_grouped_matmul_inplace_add/examples
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/tests
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/examples
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/examples_david
- ops/ops-transformer/gmm/grouped_matmul_add/tests
- ops/ops-transformer/gmm/grouped_matmul_add/examples
non_delivery: null
open_source: null
mc2_moe_advanced@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/mc2/moe_distribute_dispatch/op_host/
- ops/ops-transformer/mc2/moe_distribute_dispatch/op_api/
- ops/ops-transformer/mc2/moe_distribute_dispatch/op_graph/
- ops/ops-transformer/mc2/moe_distribute_dispatch/op_kernel/arch35/
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/arch35/
- ops/ops-transformer/mc2/moe_distribute_dispatch/op_kernel/check_winsize.h
- ops/ops-transformer/mc2/moe_distribute_dispatch/op_kernel/common.h
- ops/ops-transformer/mc2/moe_distribute_dispatch/op_kernel/moe_distribute_dispatch_tiling_key.h
- ops/ops-transformer/mc2/moe_distribute_dispatch/op_kernel/moe_distribute_dispatch_tiling.h
- ops/ops-transformer/mc2/moe_distribute_dispatch/op_kernel/moe_distribute_dispatch.cpp
- ops/ops-transformer/mc2/moe_distribute_dispatch/op_kernel/moe_distribute_dispatch.h
- ops/ops-transformer/mc2/moe_distribute_dispatch/op_kernel/quantize_functions.h
- ops/ops-transformer/mc2/moe_distribute_dispatch/CMakeLists.txt
- ops/ops-transformer/mc2/moe_distribute_dispatch/README.md
- ops/ops-transformer/mc2/moe_distribute_combine/op_host/
- ops/ops-transformer/mc2/moe_distribute_combine/op_api/
- ops/ops-transformer/mc2/moe_distribute_combine/op_graph/
- ops/ops-transformer/mc2/moe_distribute_combine/op_kernel/arch35/
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_kernel/arch35/
- ops/ops-transformer/mc2/moe_distribute_combine/op_kernel/moe_distribute_combine_tiling_key.h
- ops/ops-transformer/mc2/moe_distribute_combine/op_kernel/moe_distribute_combine_tiling.h
- ops/ops-transformer/mc2/moe_distribute_combine/op_kernel/moe_distribute_combine.cpp
- ops/ops-transformer/mc2/moe_distribute_combine/op_kernel/moe_distribute_combine.h
- ops/ops-transformer/mc2/moe_distribute_combine/CMakeLists.txt
- ops/ops-transformer/mc2/moe_distribute_combine/README.md
- ops/ops-transformer/mc2/CMakeLists.txt
opensource_style: null
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/mc2/moe_distribute_dispatch/tests/
- ops/ops-transformer/mc2/moe_distribute_dispatch/examples/
- ops/ops-transformer/mc2/moe_distribute_dispatch/docs/
- ops/ops-transformer/mc2/moe_distribute_combine/tests/
- ops/ops-transformer/mc2/moe_distribute_combine/examples/
- ops/ops-transformer/mc2/moe_distribute_combine/docs/
non_delivery: null
open_source:
- ops/ops-transformer/mc2/3rd/mat_mul_v3/
mc2_moe@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_host/config/
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_host/op_tiling/arch35/
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_host/op_tiling/moe_distribute_check_win_size.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_host/op_tiling/moe_distribute_dispatch_tiling_helper.cpp
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_host/op_tiling/moe_distribute_dispatch_tiling_helper.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_host/op_tiling/moe_distribute_dispatch_tiling_v2.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_host/op_tiling/moe_distribute_dispatch_v2_tiling.cpp
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_host/moe_distribute_dispatch_v2_def.cpp
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_host/moe_distribute_dispatch_v2_infershape.cpp
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_host/CMakeLists.txt
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_api/
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_graph/
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/CMakeLists.txt
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/README.md
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/common.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_tiling.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_v2_full_mesh.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_v2_host_kfc.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_v2_layered.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_v2_quant.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_v2_tiling_key.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_v2_tiling.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_v2.cpp
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_v2.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_elastic.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_v2_base.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/moe_distribute_dispatch_v2_constant.h
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/op_kernel/quantize_functions.h
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_host/config/
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_host/op_tiling/arch35/
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_host/op_tiling/moe_distribute_combine_tiling_helper.cpp
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_host/op_tiling/moe_distribute_combine_tiling_helper.h
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_host/op_tiling/moe_distribute_combine_v2_tiling.cpp
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_host/op_tiling/moe_distribute_combine_v2_tiling_base.h
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_host/moe_distribute_combine_v2_def.cpp
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_host/moe_distribute_combine_v2_infershape.cpp
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_host/CMakeLists.txt
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_api/
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_graph/
- ops/ops-transformer/mc2/moe_distribute_combine_v2/CMakeLists.txt
- ops/ops-transformer/mc2/moe_distribute_combine_v2/README.md
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_kernel/moe_distribute_combine_tiling.h
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_kernel/moe_distribute_combine_v2_layered.h
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_kernel/moe_distribute_combine_v2_quant.h
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_kernel/moe_distribute_combine_v2_tiling_key.h
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_kernel/moe_distribute_combine_v2_tiling.h
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_kernel/moe_distribute_combine_v2.cpp
- ops/ops-transformer/mc2/moe_distribute_combine_v2/op_kernel/moe_distribute_combine_v2.h
- ops/ops-transformer/mc2/moe_distribute_combine_add_rms_norm/op_host/
- ops/ops-transformer/mc2/moe_distribute_combine_add_rms_norm/op_api/
- ops/ops-transformer/mc2/moe_distribute_combine_add_rms_norm/op_graph/
- ops/ops-transformer/mc2/moe_distribute_combine_add_rms_norm/op_kernel/
- ops/ops-transformer/mc2/moe_distribute_combine_add_rms_norm/CMakeLists.txt
- ops/ops-transformer/mc2/moe_distribute_combine_add_rms_norm/README.md
- ops/ops-transformer/mc2/moe_distribute_buffer_reset/op_host/
- ops/ops-transformer/mc2/moe_distribute_buffer_reset/op_api/
- ops/ops-transformer/mc2/moe_distribute_buffer_reset/op_graph/
- ops/ops-transformer/mc2/moe_distribute_buffer_reset/op_kernel/
- ops/ops-transformer/mc2/moe_distribute_buffer_reset/CMakeLists.txt
- ops/ops-transformer/mc2/moe_distribute_buffer_reset/README.md
- ops/ops-transformer/mc2/moe_update_expert/op_host/
- ops/ops-transformer/mc2/moe_update_expert/op_api/
- ops/ops-transformer/mc2/moe_update_expert/op_graph/
- ops/ops-transformer/mc2/moe_update_expert/op_kernel/
- ops/ops-transformer/mc2/moe_update_expert/CMakeLists.txt
- ops/ops-transformer/mc2/moe_update_expert/README.md
- ops/ops-transformer/mc2/distribute_barrier/op_host/
- ops/ops-transformer/mc2/distribute_barrier/op_api/
- ops/ops-transformer/mc2/distribute_barrier/op_graph/
- ops/ops-transformer/mc2/distribute_barrier/op_kernel/
- ops/ops-transformer/mc2/distribute_barrier/CMakeLists.txt
- ops/ops-transformer/mc2/distribute_barrier/README.md
- ops/ops-transformer/mc2/elastic_receivable_test/op_host/
- ops/ops-transformer/mc2/elastic_receivable_test/op_api/
- ops/ops-transformer/mc2/elastic_receivable_test/op_graph/
- ops/ops-transformer/mc2/elastic_receivable_test/op_kernel/
- ops/ops-transformer/mc2/elastic_receivable_test/CMakeLists.txt
- ops/ops-transformer/mc2/elastic_receivable_test/README.md
- ops/ops-transformer/mc2/elastic_receivable_info_collect/op_host/
- ops/ops-transformer/mc2/elastic_receivable_info_collect/op_api/
- ops/ops-transformer/mc2/elastic_receivable_info_collect/op_graph/
- ops/ops-transformer/mc2/elastic_receivable_info_collect/op_kernel/
- ops/ops-transformer/mc2/elastic_receivable_info_collect/CMakeLists.txt
- ops/ops-transformer/mc2/elastic_receivable_info_collect/README.md
opensource_style: null
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/tests/
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/examples/
- ops/ops-transformer/mc2/moe_distribute_dispatch_v2/docs/
- ops/ops-transformer/mc2/moe_distribute_combine_v2/tests/
- ops/ops-transformer/mc2/moe_distribute_combine_v2/examples/
- ops/ops-transformer/mc2/moe_distribute_combine_v2/docs/
- ops/ops-transformer/mc2/moe_distribute_combine_add_rms_norm/tests/
- ops/ops-transformer/mc2/moe_distribute_combine_add_rms_norm/examples/
- ops/ops-transformer/mc2/moe_distribute_combine_add_rms_norm/docs/
- ops/ops-transformer/mc2/moe_distribute_buffer_reset/tests/
- ops/ops-transformer/mc2/moe_distribute_buffer_reset/examples/
- ops/ops-transformer/mc2/moe_distribute_buffer_reset/docs/
- ops/ops-transformer/mc2/moe_update_expert/tests/
- ops/ops-transformer/mc2/moe_update_expert/examples/
- ops/ops-transformer/mc2/moe_update_expert/docs/
- ops/ops-transformer/mc2/distribute_barrier/tests/
- ops/ops-transformer/mc2/distribute_barrier/examples/
- ops/ops-transformer/mc2/distribute_barrier/docs/
- ops/ops-transformer/mc2/elastic_receivable_test/tests/
- ops/ops-transformer/mc2/elastic_receivable_test/examples/
- ops/ops-transformer/mc2/elastic_receivable_test/docs/
- ops/ops-transformer/mc2/elastic_receivable_info_collect/tests/
- ops/ops-transformer/mc2/elastic_receivable_info_collect/examples/
- ops/ops-transformer/mc2/elastic_receivable_info_collect/docs/
non_delivery: null
open_source:
- ops/ops-transformer/mc2/3rd/batch_mat_mul_v3/
mc2_training@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/mc2/allto_all_all_gather_batch_mat_mul/op_host/
- ops/ops-transformer/mc2/allto_all_all_gather_batch_mat_mul/op_api/
- ops/ops-transformer/mc2/allto_all_all_gather_batch_mat_mul/op_graph/
- ops/ops-transformer/mc2/allto_all_all_gather_batch_mat_mul/op_kernel/
- ops/ops-transformer/mc2/allto_all_all_gather_batch_mat_mul/CMakeLists.txt
- ops/ops-transformer/mc2/allto_all_all_gather_batch_mat_mul/README.md
- ops/ops-transformer/mc2/batch_mat_mul_reduce_scatter_allto_all/op_host/
- ops/ops-transformer/mc2/batch_mat_mul_reduce_scatter_allto_all/op_api/
- ops/ops-transformer/mc2/batch_mat_mul_reduce_scatter_allto_all/op_graph/
- ops/ops-transformer/mc2/batch_mat_mul_reduce_scatter_allto_all/op_kernel/
- ops/ops-transformer/mc2/batch_mat_mul_reduce_scatter_allto_all/CMakeLists.txt
- ops/ops-transformer/mc2/batch_mat_mul_reduce_scatter_allto_all/README.md
- ops/ops-transformer/mc2/allto_allv_grouped_mat_mul/op_host/
- ops/ops-transformer/mc2/allto_allv_grouped_mat_mul/op_api/
- ops/ops-transformer/mc2/allto_allv_grouped_mat_mul/op_graph/
- ops/ops-transformer/mc2/allto_allv_grouped_mat_mul/op_kernel/
- ops/ops-transformer/mc2/allto_allv_grouped_mat_mul/CMakeLists.txt
- ops/ops-transformer/mc2/allto_allv_grouped_mat_mul/README.md
- ops/ops-transformer/mc2/grouped_mat_mul_allto_allv/op_host/
- ops/ops-transformer/mc2/grouped_mat_mul_allto_allv/op_api/
- ops/ops-transformer/mc2/grouped_mat_mul_allto_allv/op_graph/
- ops/ops-transformer/mc2/grouped_mat_mul_allto_allv/op_kernel/
- ops/ops-transformer/mc2/grouped_mat_mul_allto_allv/CMakeLists.txt
- ops/ops-transformer/mc2/grouped_mat_mul_allto_allv/README.md
opensource_style: null
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/mc2/allto_all_all_gather_batch_mat_mul/tests/
- ops/ops-transformer/mc2/allto_all_all_gather_batch_mat_mul/examples/
- ops/ops-transformer/mc2/allto_all_all_gather_batch_mat_mul/docs/
- ops/ops-transformer/mc2/batch_mat_mul_reduce_scatter_allto_all/tests/
- ops/ops-transformer/mc2/batch_mat_mul_reduce_scatter_allto_all/examples/
- ops/ops-transformer/mc2/batch_mat_mul_reduce_scatter_allto_all/docs/
- ops/ops-transformer/mc2/allto_allv_grouped_mat_mul/tests/
- ops/ops-transformer/mc2/allto_allv_grouped_mat_mul/examples/
- ops/ops-transformer/mc2/allto_allv_grouped_mat_mul/docs/
- ops/ops-transformer/mc2/grouped_mat_mul_allto_allv/tests/
- ops/ops-transformer/mc2/grouped_mat_mul_allto_allv/examples/
- ops/ops-transformer/mc2/grouped_mat_mul_allto_allv/docs/
non_delivery: null
open_source:
- ops/ops-transformer/mc2/3rd/quant_batch_matmul_v3/
mc2_infer@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/mc2/all_gather_matmul/op_host/
- ops/ops-transformer/mc2/all_gather_matmul/op_api/
- ops/ops-transformer/mc2/all_gather_matmul/op_graph/
- ops/ops-transformer/mc2/all_gather_matmul/op_kernel/
- ops/ops-transformer/mc2/all_gather_matmul/CMakeLists.txt
- ops/ops-transformer/mc2/all_gather_matmul/README.md
- ops/ops-transformer/mc2/matmul_reduce_scatter/op_host/
- ops/ops-transformer/mc2/matmul_reduce_scatter/op_api/
- ops/ops-transformer/mc2/matmul_reduce_scatter/op_graph/
- ops/ops-transformer/mc2/matmul_reduce_scatter/op_kernel/
- ops/ops-transformer/mc2/matmul_reduce_scatter/CMakeLists.txt
- ops/ops-transformer/mc2/matmul_reduce_scatter/README.md
- asl/ops/cann/ops/mc2/communication_and_computation/all_gather_matmul_v2/op_host/
- asl/ops/cann/ops/mc2/communication_and_computation/all_gather_matmul_v2/graph_plugin/
- asl/ops/cann/ops/mc2/communication_and_computation/all_gather_matmul_v2/op_kernel/
- ops/ops-transformer/mc2/matmul_reduce_scatter_v2/op_host/
- ops/ops-transformer/mc2/matmul_reduce_scatter_v2/op_api/
- ops/ops-transformer/mc2/matmul_reduce_scatter_v2/op_graph/
- ops/ops-transformer/mc2/matmul_reduce_scatter_v2/op_kernel/
- ops/ops-transformer/mc2/matmul_reduce_scatter_v2/CMakeLists.txt
- ops/ops-transformer/mc2/matmul_reduce_scatter_v2/README.md
- ops/ops-transformer/mc2/matmul_all_reduce/op_host/
- ops/ops-transformer/mc2/matmul_all_reduce/op_api/
- ops/ops-transformer/mc2/matmul_all_reduce/op_graph/
- ops/ops-transformer/mc2/matmul_all_reduce/op_kernel/
- ops/ops-transformer/mc2/matmul_all_reduce/CMakeLists.txt
- ops/ops-transformer/mc2/matmul_all_reduce/README.md
- ops/ops-transformer/mc2/quant_all_reduce/op_host/
- ops/ops-transformer/mc2/quant_all_reduce/op_api/
- ops/ops-transformer/mc2/quant_all_reduce/op_graph/
- ops/ops-transformer/mc2/quant_all_reduce/op_kernel/
- ops/ops-transformer/mc2/quant_all_reduce/CMakeLists.txt
- ops/ops-transformer/mc2/quant_all_reduce/README.md
- ops/ops-transformer/mc2/quant_reduce_scatter/op_host/
- ops/ops-transformer/mc2/quant_reduce_scatter/op_api/
- ops/ops-transformer/mc2/quant_reduce_scatter/op_graph/
- ops/ops-transformer/mc2/quant_reduce_scatter/op_kernel/
- ops/ops-transformer/mc2/quant_reduce_scatter/CMakeLists.txt
- ops/ops-transformer/mc2/quant_reduce_scatter/README.md
- ops/ops-transformer/mc2/3rd/common/
- ops/ops-transformer/mc2/3rd/ops_legacy/
- ops/ops-transformer/mc2/3rd/CMakeLists.txt
- ops/ops-transformer/mc2/common/inc/
- ops/ops-transformer/mc2/common/src/
opensource_style: null
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/mc2/all_gather_matmul/tests/
- ops/ops-transformer/mc2/all_gather_matmul/examples/
- ops/ops-transformer/mc2/all_gather_matmul/docs/
- ops/ops-transformer/mc2/matmul_reduce_scatter/tests/
- ops/ops-transformer/mc2/matmul_reduce_scatter/examples/
- ops/ops-transformer/mc2/matmul_reduce_scatter/docs/
- asl/ops/cann/ops/mc2/communication_and_computation/all_gather_matmul_v2/tests/
- asl/ops/cann/ops/mc2/communication_and_computation/all_gather_matmul_v2/examples/
- asl/ops/cann/ops/mc2/communication_and_computation/all_gather_matmul_v2/docs/
- ops/ops-transformer/mc2/matmul_reduce_scatter_v2/tests/
- ops/ops-transformer/mc2/matmul_reduce_scatter_v2/examples/
- ops/ops-transformer/mc2/matmul_reduce_scatter_v2/docs/
- ops/ops-transformer/mc2/matmul_all_reduce/tests/
- ops/ops-transformer/mc2/matmul_all_reduce/examples/
- ops/ops-transformer/mc2/matmul_all_reduce/docs/
- ops/ops-transformer/mc2/quant_all_reduce/tests/
- ops/ops-transformer/mc2/quant_all_reduce/examples/
- ops/ops-transformer/mc2/quant_all_reduce/docs/
- ops/ops-transformer/mc2/quant_reduce_scatter/tests/
- ops/ops-transformer/mc2/quant_reduce_scatter/examples/
- ops/ops-transformer/mc2/quant_reduce_scatter/docs/
non_delivery: null
open_source:
- ops/ops-transformer/mc2/3rd/weight_quant_batch_matmul_v2/
moe@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/moe/moe_compute_expert_tokens/op_host/moe_compute_expert_tokens_tiling.cpp
- ops/ops-transformer/moe/moe_compute_expert_tokens/op_host/moe_compute_expert_tokens_tiling.h
- ops/ops-transformer/moe/moe_compute_expert_tokens/op_kernel/
- ops/ops-transformer/moe/moe_finalize_routing_v2/op_host/moe_finalize_routing_v2_tiling_arch35.cpp
- ops/ops-transformer/moe/moe_finalize_routing_v2/op_host/moe_finalize_routing_v2_tiling_arch35.h
- ops/ops-transformer/moe/moe_finalize_routing_v2/op_host/moe_finalize_routing_v2_tiling.cpp
- ops/ops-transformer/moe/moe_finalize_routing_v2/op_host/moe_finalize_routing_v2_tiling.h
- ops/ops-transformer/moe/moe_finalize_routing_v2/op_host/moe_finalize_routing_v2_infershape.cpp
- ops/ops-transformer/moe/moe_finalize_routing_v2/op_kernel/arch35
- ops/ops-transformer/moe/moe_finalize_routing_v2/op_kernel/moe_finalize_routing_v2_apt.cpp
- ops/ops-transformer/moe/moe_finalize_routing_v2_grad/op_host/moe_finalize_routing_v2_grad_not_split_h_tiling_arch35.cpp
- ops/ops-transformer/moe/moe_finalize_routing_v2_grad/op_host/moe_finalize_routing_v2_grad_split_h_tiling_arch35.cpp
- ops/ops-transformer/moe/moe_finalize_routing_v2_grad/op_host/moe_finalize_routing_v2_grad_without_scale_tiling_arch35.cpp
- ops/ops-transformer/moe/moe_finalize_routing_v2_grad/op_host/moe_finalize_routing_v2_grad_tiling.cpp
- ops/ops-transformer/moe/moe_finalize_routing_v2_grad/op_host/moe_finalize_routing_v2_grad_tiling.h
- ops/ops-transformer/moe/moe_finalize_routing_v2_grad/op_host/moe_finalize_routing_v2_grad_infershape.cpp
- ops/ops-transformer/moe/moe_finalize_routing_v2_grad/op_kernel/arch35
- ops/ops-transformer/moe/moe_finalize_routing_v2_grad/op_kernel/moe_finalize_routing_v2_grad_apt.cpp
- ops/ops-transformer/moe/moe_gating_top_k/op_host/moe_gating_top_k_tiling_arch35.cpp
- ops/ops-transformer/moe/moe_gating_top_k/op_kernel/arch35
- ops/ops-transformer/moe/moe_gating_top_k/op_kernel/moe_gating_top_k_apt.cpp
- ops/ops-transformer/moe/moe_gating_top_k_softmax/op_host/moe_gating_top_k_softmax_tiling_arch35.cpp
- ops/ops-transformer/moe/moe_gating_top_k_softmax/op_kernel/arch35
- ops/ops-transformer/moe/moe_gating_top_k_softmax/op_kernel/moe_gating_top_k_softmax_apt.cpp
- ops/ops-transformer/moe/moe_gating_top_k_softmax_v2/op_host/moe_gating_top_k_softmax_v2_regbase_tiling_perf.cpp
- ops/ops-transformer/moe/moe_gating_top_k_softmax_v2/op_kernel/arch35
- ops/ops-transformer/moe/moe_gating_top_k_softmax_v2/op_kernel/moe_gating_top_k_softmax_v2_apt.cpp
- ops/ops-transformer/moe/moe_init_routing/op_host/moe_init_routing_infershape.cpp
- ops/ops-transformer/moe/moe_init_routing/op_host/moe_init_routing_tiling.cpp
- ops/ops-transformer/moe/moe_init_routing/op_host/moe_init_routing_tiling.h
- ops/ops-transformer/moe/moe_init_routing/op_kernel/arch35
- ops/ops-transformer/moe/moe_init_routing/op_kernel/moe_init_routing_apt.cpp
- ops/ops-transformer/moe/moe_init_routing_v2/op_host/moe_init_routing_v2_infershape.cpp
- ops/ops-transformer/moe/moe_init_routing_v2/op_host/moe_init_routing_v2_tiling_util.cpp
- ops/ops-transformer/moe/moe_init_routing_v2/op_host/moe_init_routing_v2_tiling.cpp
- ops/ops-transformer/moe/moe_init_routing_v2/op_host/moe_init_routing_v2_tiling.h
- ops/ops-transformer/moe/moe_init_routing_v2/op_kernel/arch35
- ops/ops-transformer/moe/moe_init_routing_v2/op_kernel/moe_init_routing_v2.cpp
- ops/ops-transformer/moe/moe_init_routing_quant_v2/op_host/moe_init_routing_quant_v2_infershape.cpp
- ops/ops-transformer/moe/moe_init_routing_quant_v2/op_host/moe_init_routing_quant_v2_tiling_base.cpp
- ops/ops-transformer/moe/moe_init_routing_quant_v2/op_host/moe_init_routing_quant_v2_tiling_base.h
- ops/ops-transformer/moe/moe_init_routing_quant_v2/op_host/moe_init_routing_quant_v2_tiling.cpp
- ops/ops-transformer/moe/moe_init_routing_quant_v2/op_host/moe_init_routing_quant_v2_tiling.h
- ops/ops-transformer/moe/moe_init_routing_quant_v2/op_kernel/arch35
- ops/ops-transformer/moe/moe_init_routing_quant_v2/op_kernel/moe_init_routing_quant_v2_apt.cpp
- ops/ops-transformer/moe/moe_init_routing_v2_grad/op_host/moe_init_routing_v2_grad_full_load_tiling_arch35.cpp
- ops/ops-transformer/moe/moe_init_routing_v2_grad/op_host/moe_init_routing_v2_grad_split_h_tiling_arch35.cpp
- ops/ops-transformer/moe/moe_init_routing_v2_grad/op_host/moe_init_routing_v2_grad_tiling_arch35.cpp
- ops/ops-transformer/moe/moe_init_routing_v2_grad/op_kernel/arch35
- ops/ops-transformer/moe/moe_init_routing_v2_grad/op_kernel/moe_init_routing_v2_grad_apt.cpp
- ops/ops-transformer/moe/moe_init_routing_v3/op_host/moe_init_routing_v3_infershape.cpp
- ops/ops-transformer/moe/moe_init_routing_v3/op_host/moe_init_routing_v3_tiling_arch35.cpp
- ops/ops-transformer/moe/moe_init_routing_v3/op_kernel/arch35
- ops/ops-transformer/moe/moe_init_routing_v3/op_kernel/moe_init_routing_v3_apt.cpp
- ops/ops-transformer/moe/moe_re_routing/op_host/moe_re_routing_infershape.cpp
- ops/ops-transformer/moe/moe_re_routing/op_host/moe_re_routing_membase_tiling.cpp
- ops/ops-transformer/moe/moe_re_routing/op_host/moe_re_routing_membase_tiling.h
- ops/ops-transformer/moe/moe_re_routing/op_host/moe_re_routing_r_tiling.cpp
- ops/ops-transformer/moe/moe_re_routing/op_host/moe_re_routing_r_tiling.h
- ops/ops-transformer/moe/moe_re_routing/op_host/moe_re_routing_re_tiling.cpp
- ops/ops-transformer/moe/moe_re_routing/op_host/moe_re_routing_re_tiling.h
- ops/ops-transformer/moe/moe_re_routing/op_host/moe_re_routing_tiling_base.cpp
- ops/ops-transformer/moe/moe_re_routing/op_host/moe_re_routing_tiling_base.h
- ops/ops-transformer/moe/moe_re_routing/op_host/moe_re_routing_tiling.cpp
- ops/ops-transformer/moe/moe_re_routing/op_host/moe_re_routing_tiling.h
- ops/ops-transformer/moe/moe_re_routing/op_kernel/arch35
- ops/ops-transformer/moe/moe_re_routing/op_kernel/moe_re_routing_apt.cpp
opensource_style: null
kernel_style: null
unrelease:
test_code:
- ops/ops-transformer/moe/moe_compute_expert_tokens/tests/
- ops/ops-transformer/moe/moe_compute_expert_tokens/examples/
- ops/ops-transformer/moe/moe_compute_expert_tokens/docs/
- ops/ops-transformer/moe/moe_finalize_routing/tests/
- ops/ops-transformer/moe/moe_finalize_routing/examples/
- ops/ops-transformer/moe/moe_finalize_routing/docs/
- ops/ops-transformer/moe/moe_finalize_routing_v2/tests/
- ops/ops-transformer/moe/moe_finalize_routing_v2/examples/
- ops/ops-transformer/moe/moe_finalize_routing_v2/docs/
- ops/ops-transformer/moe/moe_finalize_routing_v2_grad/tests/
- ops/ops-transformer/moe/moe_finalize_routing_v2_grad/examples/
- ops/ops-transformer/moe/moe_finalize_routing_v2_grad/docs/
- ops/ops-transformer/moe/moe_gating_top_k_softmax/tests/
- ops/ops-transformer/moe/moe_gating_top_k_softmax/examples/
- ops/ops-transformer/moe/moe_gating_top_k_softmax/docs/
- ops/ops-transformer/moe/moe_gating_top_k_softmax_v2/tests/
- ops/ops-transformer/moe/moe_gating_top_k_softmax_v2/examples/
- ops/ops-transformer/moe/moe_gating_top_k_softmax_v2/docs/
- ops/ops-transformer/moe/moe_init_routing/tests/
- ops/ops-transformer/moe/moe_init_routing/examples/
- ops/ops-transformer/moe/moe_init_routing/docs/
- ops/ops-transformer/moe/moe_init_routing_quant/tests/
- ops/ops-transformer/moe/moe_init_routing_quant/examples/
- ops/ops-transformer/moe/moe_init_routing_quant/docs/
- ops/ops-transformer/moe/moe_init_routing_quant_v2/tests/
- ops/ops-transformer/moe/moe_init_routing_quant_v2/examples/
- ops/ops-transformer/moe/moe_init_routing_quant_v2/docs/
- ops/ops-transformer/moe/moe_init_routing_v2/tests/
- ops/ops-transformer/moe/moe_init_routing_v2/examples/
- ops/ops-transformer/moe/moe_init_routing_v2/docs/
- ops/ops-transformer/moe/moe_init_routing_v2_grad/tests/
- ops/ops-transformer/moe/moe_init_routing_v2_grad/examples/
- ops/ops-transformer/moe/moe_init_routing_v2_grad/docs/
- ops/ops-transformer/moe/moe_token_permute_with_ep/tests/
- ops/ops-transformer/moe/moe_token_permute_with_ep/examples/
- ops/ops-transformer/moe/moe_token_permute_with_ep/docs/
- ops/ops-transformer/moe/moe_token_permute_with_ep_grad/tests/
- ops/ops-transformer/moe/moe_token_permute_with_ep_grad/examples/
- ops/ops-transformer/moe/moe_token_permute_with_ep_grad/docs/
- ops/ops-transformer/moe/moe_token_permute_with_routing_map/tests/
- ops/ops-transformer/moe/moe_token_permute_with_routing_map/examples/
- ops/ops-transformer/moe/moe_token_permute_with_routing_map/docs/
- ops/ops-transformer/moe/moe_token_permute_with_routing_map_grad/tests/
- ops/ops-transformer/moe/moe_token_permute_with_routing_map_grad/examples/
- ops/ops-transformer/moe/moe_token_permute_with_routing_map_grad/docs/
- ops/ops-transformer/moe/moe_token_unpermute/tests/
- ops/ops-transformer/moe/moe_token_unpermute/examples/
- ops/ops-transformer/moe/moe_token_unpermute/docs/
- ops/ops-transformer/moe/moe_token_unpermute_grad/tests/
- ops/ops-transformer/moe/moe_token_unpermute_grad/examples/
- ops/ops-transformer/moe/moe_token_unpermute_grad/docs/
- ops/ops-transformer/moe/moe_token_unpermute_with_ep/tests/
- ops/ops-transformer/moe/moe_token_unpermute_with_ep/examples/
- ops/ops-transformer/moe/moe_token_unpermute_with_ep/docs/
- ops/ops-transformer/moe/moe_token_unpermute_with_ep_grad/tests/
- ops/ops-transformer/moe/moe_token_unpermute_with_ep_grad/examples/
- ops/ops-transformer/moe/moe_token_unpermute_with_ep_grad/docs/
- ops/ops-transformer/moe/moe_token_unpermute_with_routing_map/tests/
- ops/ops-transformer/moe/moe_token_unpermute_with_routing_map/examples/
- ops/ops-transformer/moe/moe_token_unpermute_with_routing_map/docs/
- ops/ops-transformer/moe/moe_token_unpermute_with_routing_map_grad/tests/
- ops/ops-transformer/moe/moe_token_unpermute_with_routing_map_grad/examples/
- ops/ops-transformer/moe/moe_token_unpermute_with_routing_map_grad/docs/
non_delivery: null
open_source: null
HPTG2@ops-transformer:
src:
release:
huawei_style:
- ops/ops-transformer/posembedding/apply_rotary_pos_emb/op_host/apply_rotary_pos_emb_tiling_aba_and_ba_arch35.cpp
opensource_style: null
kernel_style: null
unrelease:
test_code: null
non_delivery: null
open_source: null
GMMSwigluQuant@ops-transformer:
src:
release:
huawei_style:
- asl/ops/ops_adv/src/transformer/mla_preprocess/ophost/
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/gmm_infra
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_a4w4.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_antiquant_a8w4_msd_new.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_antiquant_a8w4_msd_pre.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_antiquant_a8w4_msd.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_antiquant_a8w4_pre.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_antiquant_a16w8_msd.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_antiquant.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_autotiling_a8w4.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_fixaxismove_interface.cpp
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_fixaxismove_regular.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_pre_tiling.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_quant_mixcore.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_tiling_key.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_utils.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul_vector.h
- ops/ops-transformer/gmm/grouped_matmul/op_kernel/grouped_matmul.h
- ops/ops-transformer/gmm/grouped_matmul/op_host/op_tiling/grouped_matmul_tiling.cpp
- ops/ops-transformer/gmm/grouped_matmul/op_host/op_tiling/grouped_matmul_tiling.h
- ops/ops-transformer/gmm/grouped_matmul/op_api/aclnn_grouped_matmul.cpp
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_kernel/grouped_matmul_swiglu_quant_spilit_fusion.h
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_kernel/grouped_matmul_swiglu_quant_v2_a8w4_msd_mid.h
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_kernel/grouped_matmul_swiglu_quant_v2_a8w4_msd_post.h
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_kernel/grouped_matmul_swiglu_quant_v2_a8w4_msd_pre.h
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_kernel/grouped_matmul_swiglu_quant_v2_pipeline.h
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_kernel/grouped_matmul_swiglu_quant_v2_utils.h
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_host/op_tiling/grouped_matmul_swiglu_quant_base_tiling.cpp
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_host/op_tiling/grouped_matmul_swiglu_quant_base_tiling.h
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_host/op_tiling/grouped_matmul_swiglu_quant_fusion_tiling.cpp
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_host/op_tiling/grouped_matmul_swiglu_quant_fusion_tiling.h
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_host/op_tiling/grouped_matmul_swiglu_quant_tiling.cpp
- ops/ops-transformer/gmm/grouped_matmul_swiglu_quant_v2/op_host/op_tiling/grouped_matmul_swiglu_quant_tiling.h
opensource_style: null
kernel_style: null
unrelease:
test_code: null
non_delivery: null
open_source: null
