mirror of
https://github.com/ggml-org/whisper.cpp.git
synced 2026-10-17 20:03:45 +02:00
vitisai : add VitisAI Plugin for AMD Ryzen AI NPU encoder offload (#3608)
* Add VitisAI Plugin * Added VitisAI encoder module placeholder files * VitisAI build integration * VitisAI encoder offload functional * Clean up vitisai integration * Add c++17 requirement for Windows * Enabled preemption for windows runs * Add model cache override option * Remove vitisai premature log message * Add rai support through file mapping * Fixed flatbuffer loading * Fixed Windows file mapping issue * Update FlexmlRT resolution * Use Flexmlrt wheel pkg to build VitisAI plugin * Clean up * Remove prints * Change flexmlrt target from Shared to Interface * Add c++17 requirement for Windows * Enabled preemption for windows runs * Add rai support through file mapping * Fixed flatbuffer loading * Fixed Windows file mapping issue * Update FlexmlRT resolution * Use Flexmlrt wheel pkg to build VitisAI plugin * Clean up * Remove prints * Change flexmlrt target from Shared to Interface * Cleanup FlexmlRT integration * format fix * Adding AMD Licenses * Update CMakeLists.txt Co-authored-by: Kumawat, Sachin <sachin.kumawat@amd.com> * Update src/CMakeLists.txt Co-authored-by: Kumawat, Sachin <sachin.kumawat@amd.com> * Update whisper.cpp * Added VitisAI encoder readme section * Remove license headers from common files to whisper.cpp --------- Co-authored-by: Sachin Kumawat <sachink@amd.com> Co-authored-by: Jeff Lin <jeffylin@xilinx.com> Co-authored-by: Lin <jefflin@amd.com> Co-authored-by: Lin, Jeff (DCG-ENG) <jeff.lin@amd.com> Co-authored-by: Iswarya Alex <iswaryaalex96@gmail.com> Co-authored-by: Alex, Iswarya <Iswarya.Alex@amd.com> * Update README.md - RAI EULA Links - Updated for RAI Whisper instructions * Cleanup and add runtime print debug guard * turn off profiling * Add VitisAI Plugin * Added VitisAI encoder module placeholder files * VitisAI build integration * VitisAI encoder offload functional * Clean up vitisai integration * Add c++17 requirement for Windows * Enabled preemption for windows runs * Add model cache override option * Remove vitisai premature log message * Add rai support through file mapping * Fixed flatbuffer loading * Fixed Windows file mapping issue * Update FlexmlRT resolution * Use Flexmlrt wheel pkg to build VitisAI plugin * Clean up * Remove prints * Change flexmlrt target from Shared to Interface * Add c++17 requirement for Windows * Enabled preemption for windows runs * Add rai support through file mapping * Fixed flatbuffer loading * Fixed Windows file mapping issue * Update FlexmlRT resolution * Use Flexmlrt wheel pkg to build VitisAI plugin * Clean up * Remove prints * Change flexmlrt target from Shared to Interface * Cleanup FlexmlRT integration * format fix * Adding AMD Licenses * Update CMakeLists.txt Co-authored-by: Kumawat, Sachin <sachin.kumawat@amd.com> * Update src/CMakeLists.txt Co-authored-by: Kumawat, Sachin <sachin.kumawat@amd.com> * Update whisper.cpp * Added VitisAI encoder readme section * Remove license headers from common files to whisper.cpp --------- Co-authored-by: Sachin Kumawat <sachink@amd.com> Co-authored-by: Jeff Lin <jeffylin@xilinx.com> Co-authored-by: Lin <jefflin@amd.com> Co-authored-by: Lin, Jeff (DCG-ENG) <jeff.lin@amd.com> Co-authored-by: Iswarya Alex <iswaryaalex96@gmail.com> Co-authored-by: Alex, Iswarya <Iswarya.Alex@amd.com> * Cleanup and add runtime print debug guard * Update README.md - RAI EULA Links - Updated for RAI Whisper instructions * turn off profiling * Let flexmlrt detect device type * Add VitisAI model download scripts * Add encoder + cross projection layer offload * Add self hosted runner for amd npu * Update runner * Update workflow for linux * Update workflow for linux * Update flexmlrt packages for linux * Update flexmlrt packages for linux * Updated README * readme: clarify xrt * readme: clarify xrt * ci: update test config * Added supported plarform details with python 3.12 requirement for Linux * Use refactored helpers * Deprecate cross_proj .rai naming and cleanup * Remove stale function code * Fix: formatting --------- Co-authored-by: Jeff Lin <jeffylin@xilinx.com> Co-authored-by: Lin <jefflin@amd.com> Co-authored-by: Lin, Jeff (DCG-ENG) <jeff.lin@amd.com> Co-authored-by: Iswarya Alex <iswaryaalex96@gmail.com> Co-authored-by: Alex, Iswarya <Iswarya.Alex@amd.com> Co-authored-by: Iswarya Alex <47045679+iswaryaalex@users.noreply.github.com>
This commit is contained in:
Sachin Kumawat
co-authored by
Jeff Lin
Lin
Lin, Jeff (DCG-ENG)
Iswarya Alex
Alex, Iswarya
Iswarya Alex
Jeff Lin
Lin
Lin, Jeff (DCG-ENG)
Iswarya Alex
Alex, Iswarya
Iswarya Alex
parent
233fe1fc9b
commit
52dec9d889
@@ -48,6 +48,61 @@ if (WHISPER_OPENVINO)
|
||||
find_package(OpenVINO REQUIRED COMPONENTS Runtime)
|
||||
endif()
|
||||
|
||||
if (WHISPER_VITISAI)
|
||||
find_package(FlexmlRT REQUIRED)
|
||||
|
||||
# Legacy RAI overrides are required by FlexMLRT older than 1.8.0
|
||||
set(WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES_MODE "AUTO" CACHE STRING
|
||||
"Legacy RAI override mode for FlexMLRT (AUTO|ON|OFF)")
|
||||
set_property(CACHE WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES_MODE PROPERTY STRINGS AUTO ON OFF)
|
||||
|
||||
string(TOUPPER "${WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES_MODE}" _flexmlrt_legacy_mode)
|
||||
set(_flexmlrt_legacy_hint "Set -DWHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES_MODE=ON or OFF explicitly.")
|
||||
|
||||
if (NOT _flexmlrt_legacy_mode MATCHES "^(AUTO|ON|OFF)$")
|
||||
message(FATAL_ERROR
|
||||
"Invalid WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES_MODE='${WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES_MODE}'. "
|
||||
"Expected AUTO, ON, or OFF.")
|
||||
endif()
|
||||
|
||||
if (_flexmlrt_legacy_mode STREQUAL "AUTO")
|
||||
if (NOT FlexmlRT_DIR)
|
||||
message(FATAL_ERROR
|
||||
"FlexmlRT_DIR is unset after find_package(FlexmlRT). ${_flexmlrt_legacy_hint}")
|
||||
endif()
|
||||
|
||||
# FlexmlRT_DIR points to <pkg_root>/share/cmake/FlexmlRT.
|
||||
get_filename_component(_flexmlrt_init_py "${FlexmlRT_DIR}/../../../__init__.py" ABSOLUTE)
|
||||
if (NOT EXISTS "${_flexmlrt_init_py}")
|
||||
message(FATAL_ERROR
|
||||
"flexmlrt __init__.py not found at ${_flexmlrt_init_py}. ${_flexmlrt_legacy_hint}")
|
||||
endif()
|
||||
set_property(DIRECTORY APPEND PROPERTY CMAKE_CONFIGURE_DEPENDS "${_flexmlrt_init_py}")
|
||||
|
||||
file(STRINGS "${_flexmlrt_init_py}" _flexmlrt_version_lines
|
||||
REGEX "^VERSION[ \t]*=[ \t]*\"[0-9]+\\.[0-9]+\\.[0-9]+")
|
||||
if (NOT _flexmlrt_version_lines MATCHES "\"([0-9]+\\.[0-9]+\\.[0-9]+)")
|
||||
message(FATAL_ERROR
|
||||
"Could not parse flexmlrt VERSION from ${_flexmlrt_init_py}. ${_flexmlrt_legacy_hint}")
|
||||
endif()
|
||||
set(_flexmlrt_version "${CMAKE_MATCH_1}")
|
||||
|
||||
if (_flexmlrt_version VERSION_LESS "1.8.0")
|
||||
set(WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES 1)
|
||||
else()
|
||||
set(WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES 0)
|
||||
endif()
|
||||
message(STATUS "Detected flexmlrt VERSION=${_flexmlrt_version} from ${_flexmlrt_init_py} (legacy overrides=${WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES})")
|
||||
else()
|
||||
if (_flexmlrt_legacy_mode STREQUAL "ON")
|
||||
set(WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES 1)
|
||||
else()
|
||||
set(WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES 0)
|
||||
endif()
|
||||
message(STATUS "FlexMLRT legacy RAI overrides forced ${_flexmlrt_legacy_mode}")
|
||||
endif()
|
||||
endif()
|
||||
|
||||
#
|
||||
# libraries
|
||||
#
|
||||
@@ -101,6 +156,33 @@ if (WHISPER_OPENVINO)
|
||||
set_target_properties(${TARGET} PROPERTIES FOLDER "libs")
|
||||
endif()
|
||||
|
||||
if (WHISPER_VITISAI)
|
||||
set(TARGET whisper.vitisai)
|
||||
|
||||
add_library(${TARGET} OBJECT
|
||||
vitisai/whisper-vitisai-helpers.h
|
||||
vitisai/whisper-vitisai-helpers.cpp
|
||||
vitisai/whisper-vitisai-encoder.h
|
||||
vitisai/whisper-vitisai-encoder.cpp
|
||||
)
|
||||
|
||||
target_include_directories(${TARGET} PUBLIC .)
|
||||
|
||||
set_property(TARGET ${TARGET} PROPERTY POSITION_INDEPENDENT_CODE ON)
|
||||
set(WHISPER_EXTRA_FLAGS ${WHISPER_EXTRA_FLAGS} -DWHISPER_USE_VITISAI)
|
||||
|
||||
# FlexMLRT headers and this plugin require C++17. Keep it PRIVATE so the
|
||||
# C++11 requirement of the whisper target is not bumped.
|
||||
target_compile_features(${TARGET} PRIVATE cxx_std_17)
|
||||
|
||||
target_compile_definitions(${TARGET} PRIVATE
|
||||
WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES=${WHISPER_FLEXMLRT_LEGACY_RAI_OVERRIDES}
|
||||
)
|
||||
|
||||
target_link_libraries(${TARGET} PRIVATE ggml flexmlrt::flexmlrt)
|
||||
set_target_properties(${TARGET} PROPERTIES FOLDER "libs")
|
||||
endif()
|
||||
|
||||
# whisper
|
||||
|
||||
add_library(whisper
|
||||
@@ -159,6 +241,10 @@ if (WHISPER_OPENVINO)
|
||||
target_link_libraries(whisper PRIVATE whisper.openvino)
|
||||
endif()
|
||||
|
||||
if (WHISPER_VITISAI)
|
||||
target_link_libraries(whisper PRIVATE whisper.vitisai)
|
||||
endif()
|
||||
|
||||
if (WHISPER_MKL)
|
||||
target_link_libraries(whisper PRIVATE MKL::MKL)
|
||||
endif()
|
||||
|
||||
Reference in New Issue
Block a user