Files
composable_kernel/include/ck_tile/ops/fmha.hpp
Dan Yao 2cab8d39e3 CK Tile FA Training kernels (#1286)
* FA fwd dropout

* FA bwd

* epilogue reuse

* CMakeLists update

* [CK_TILE] support alibi (#1269)

* add alibi support

* fix code

* update code based on comment

* Support more hdim

* fix fp8 bias

* support seqlen_k=0 case

* remove unused printf

* fix format

---------

Co-authored-by: rocking <ChunYu.Lai@amd.com>

* now fwd/bwd can build

* bwd alibi

* add bwd validation stream_config

* update generated filenames

* update bwd kernel launch

* CK_TILE_HOST_DEVICE in philox

* Transpose -> transpose

* format

* format

* format

* Generate the instance for FA required

* format

* fix error in WarpGemm

---------

Co-authored-by: danyao12 <danyao12>
Co-authored-by: carlushuang <carlus.huang@amd.com>
Co-authored-by: rocking <ChunYu.Lai@amd.com>
Co-authored-by: Po Yen Chen <PoYen.Chen@amd.com>
Co-authored-by: Jing Zhang <jizhan@amd.com>
2024-06-04 13:12:45 -05:00

38 lines
2.4 KiB
C++

// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#pragma once
#include "ck_tile/ops/fmha/block/block_attention_bias_enum.hpp"
#include "ck_tile/ops/fmha/block/block_dropout.hpp"
#include "ck_tile/ops/fmha/block/block_masking.hpp"
#include "ck_tile/ops/fmha/block/block_position_encoding.hpp"
#include "ck_tile/ops/fmha/kernel/fmha_bwd_kernel.hpp"
#include "ck_tile/ops/fmha/kernel/fmha_bwd_tile_partitioner.hpp"
#include "ck_tile/ops/fmha/kernel/fmha_fwd_kernel.hpp"
#include "ck_tile/ops/fmha/kernel/fmha_fwd_tile_partitioner.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_bwd_dot_do_o.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_bwd_dot_do_o_default_policy.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_bwd_dq_dk_dv_pipeline_ks_kts_vr.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_bwd_dq_dk_dv_pipeline_ks_kts_vr_default_policy.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_bwd_dq_dk_dv_pipeline_ks_vr.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_bwd_dq_dk_dv_pipeline_ks_vr_default_policy.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_bwd_dq_dk_dv_pipeline_qs_ks_vr_dos.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_bwd_dq_dk_dv_pipeline_qs_ks_vr_dos_default_policy.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_bwd_pipeline_default_policy.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_bwd_pipeline_enum.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_bwd_pipeline_problem.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_enum.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_problem.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_async.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_async_default_policy.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_default_policy.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qr_ks_vs_fp8.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qs_ks_vs.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qs_ks_vs_default_policy.hpp"
#include "ck_tile/ops/fmha/pipeline/block_fmha_pipeline_qx_ks_vs_custom_policy.hpp"
#include "ck_tile/ops/fmha/pipeline/tile_fmha_shape.hpp"
#include "ck_tile/ops/fmha/pipeline/tile_fmha_traits.hpp"
#include "ck_tile/ops/common/tensor_layout.hpp"