extern/whisper.cpp
1
0
Fork 1
mirror of https://github.com/ggerganov/whisper.cpp.git synced 2024-11-07 08:34:37 +01:00
Code Issues Packages Projects Releases Wiki Activity

whisper : disable CUDA mel + fix FFMPEG

Browse Source
This commit is contained in:
Georgi Gerganov 2024-06-26 20:11:38 +03:00
parent 3efedb9511
commit dc8cc2dd6f
No known key found for this signature in database
GPG Key ID: 449E073F9DC10735
6 changed files with 68 additions and 41 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
examples/CMakeLists.txt
View File

@ -11,7 +11,7 @@ if (WHISPER_SDL2)
string(STRIP "${SDL2_LIBRARIES}" SDL2_LIBRARIES) string(STRIP "${SDL2_LIBRARIES}" SDL2_LIBRARIES)
message(STATUS "SDL2_INCLUDE_DIRS = ${SDL2_INCLUDE_DIRS}") message(STATUS "SDL2_INCLUDE_DIRS = ${SDL2_INCLUDE_DIRS}")
message(STATUS "SDL2_LIBRARIES = ${SDL2_LIBRARIES}") message(STATUS "SDL2_LIBRARIES = ${SDL2_LIBRARIES}")
endif() endif()
if (WHISPER_CLBLAST) if (WHISPER_CLBLAST)
@ -22,10 +22,35 @@ endif()
set(TARGET common) set(TARGET common)
unset(COMMON_EXTRA_LIBS)
if (WHISPER_FFMPEG) if (WHISPER_FFMPEG)
# As of cmake 3.27, there is no official cmake support for FindFFmpeg.
# Consequnelty we added a FindFFmpeg.cmake script the cmake subfolder:
# whisper.cpp does not need the full ffmpeg libs, just AVFORMAT AVCODEC AVUTIL SWRESAMPLE
# libswresample performs highly optimized audio resampling, rematrixing and sample format conversion operations
# libavcodec provides a generic encoding/decoding framework and contains multiple decoders and encoders for audio, video and subtitle streams, and several bitstream filters.
# libavformat provides a generic framework for multiplexing and demultiplexing (muxing and demuxing) audio, video and subtitle streams.
find_package(FFmpeg REQUIRED)
if (NOT ${FFMPEG_FOUND})
message(FATAL_ERROR "Cannot find ffmpeg libs/headers")
endif()
message(STATUS "Found ffmpeg libs: ${FFMPEG_LIBRARIES}")
message(STATUS "Found ffmpeg headers in: ${FFMPEG_INCLUDE_DIRS}")
message(STATUS "ffmpeg definitions: ${FFMPEG_DEFINITIONS}")
message(STATUS "Found avformat ${AVFORMAT_VERSION}")
include_directories(${FFMPEG_INCLUDE_DIRS})
add_compile_definitions(WHISPER_FFMPEG)
list(APPEND COMMON_EXTRA_LIBS ${FFMPEG_LIBRARIES})
set(COMMON_SOURCES_FFMPEG ffmpeg-transcode.cpp) set(COMMON_SOURCES_FFMPEG ffmpeg-transcode.cpp)
endif() endif()
add_library(${TARGET} STATIC add_library(${TARGET} STATIC
common.h common.h
common.cpp common.cpp
@ -38,7 +63,7 @@ add_library(${TARGET} STATIC
include(DefaultTargetOptions) include(DefaultTargetOptions)
target_link_libraries(${TARGET} PRIVATE whisper) target_link_libraries(${TARGET} PRIVATE whisper ${COMMON_EXTRA_LIBS})
set_target_properties(${TARGET} PROPERTIES POSITION_INDEPENDENT_CODE ON) set_target_properties(${TARGET} PROPERTIES POSITION_INDEPENDENT_CODE ON)
set_target_properties(${TARGET} PROPERTIES FOLDER "libs") set_target_properties(${TARGET} PROPERTIES FOLDER "libs")
@ -55,8 +80,8 @@ if (WHISPER_SDL2)
include(DefaultTargetOptions) include(DefaultTargetOptions)
target_include_directories(${TARGET} PUBLIC ${SDL2_INCLUDE_DIRS}) target_include_directories(${TARGET} PUBLIC ${SDL2_INCLUDE_DIRS})
target_link_libraries(${TARGET} PRIVATE ${SDL2_LIBRARIES}) target_link_libraries (${TARGET} PRIVATE ${SDL2_LIBRARIES})
set_target_properties(${TARGET} PROPERTIES POSITION_INDEPENDENT_CODE ON) set_target_properties(${TARGET} PROPERTIES POSITION_INDEPENDENT_CODE ON)
set_target_properties(${TARGET} PROPERTIES FOLDER "libs") set_target_properties(${TARGET} PROPERTIES FOLDER "libs")
@ -108,7 +133,7 @@ if (WHISPER_SDL2)
set_target_properties(talk-llama PROPERTIES FOLDER "examples") set_target_properties(talk-llama PROPERTIES FOLDER "examples")
add_subdirectory(lsp) add_subdirectory(lsp)
set_target_properties(lsp PROPERTIES FOLDER "examples") set_target_properties(lsp PROPERTIES FOLDER "examples")
if (LLAMA_SYCL) if (GGML_SYCL)
add_subdirectory(sycl) add_subdirectory(sycl)
set_target_properties(sycl PROPERTIES FOLDER "examples") set_target_properties(sycl PROPERTIES FOLDER "examples")
endif() endif()

6
examples/common.cpp
View File

@ -30,7 +30,7 @@ extern bool ffmpeg_decode_audio(const std::string & ifname, std::vector<uint8_t>
#endif #endif
// Function to check if the next argument exists // Function to check if the next argument exists
std::string get_next_arg(int& i, int argc, char** argv, const std::string& flag, gpt_params& params) { static std::string get_next_arg(int& i, int argc, char** argv, const std::string& flag, gpt_params& params) {
if (i + 1 < argc && argv[i + 1][0] != '-') { if (i + 1 < argc && argv[i + 1][0] != '-') {
return argv[++i]; return argv[++i];
} else { } else {
@ -346,7 +346,7 @@ std::vector<gpt_vocab::id> gpt_tokenize(const gpt_vocab & vocab, const std::stri
return tokens; return tokens;
} }
std::vector<gpt_vocab::id> parse_tokens_from_string(const std::string& input, char delimiter) { static std::vector<gpt_vocab::id> parse_tokens_from_string(const std::string& input, char delimiter) {
std::vector<gpt_vocab::id> output; std::vector<gpt_vocab::id> output;
std::stringstream ss(input); std::stringstream ss(input);
std::string token; std::string token;
@ -358,7 +358,7 @@ std::vector<gpt_vocab::id> parse_tokens_from_string(const std::string& input, ch
return output; return output;
} }
std::map<std::string, std::vector<gpt_vocab::id>> extract_tests_from_file(const std::string & fpath_test){ static std::map<std::string, std::vector<gpt_vocab::id>> extract_tests_from_file(const std::string & fpath_test){
if (fpath_test.empty()){ if (fpath_test.empty()){
fprintf(stderr, "%s : No test file found.\n", __func__); fprintf(stderr, "%s : No test file found.\n", __func__);
return std::map<std::string, std::vector<gpt_vocab::id>>(); return std::map<std::string, std::vector<gpt_vocab::id>>();

8
scripts/build-info.sh
View File

@ -24,7 +24,7 @@ if out=$($CC -dumpmachine); then
build_target=$out build_target=$out
fi fi
echo "int LLAMA_BUILD_NUMBER = ${build_number};" echo "int WHISPER_BUILD_NUMBER = ${build_number};"
echo "char const *LLAMA_COMMIT = \"${build_commit}\";" echo "char const *WHISPER_COMMIT = \"${build_commit}\";"
echo "char const *LLAMA_COMPILER = \"${build_compiler}\";" echo "char const *WHISPER_COMPILER = \"${build_compiler}\";"
echo "char const *LLAMA_BUILD_TARGET = \"${build_target}\";" echo "char const *WHISPER_BUILD_TARGET = \"${build_target}\";"

53
src/CMakeLists.txt
View File

@ -77,27 +77,27 @@ if (WHISPER_OPENVINO)
set_target_properties(${TARGET} PROPERTIES FOLDER "libs") set_target_properties(${TARGET} PROPERTIES FOLDER "libs")
endif() endif()
if (GGML_CUDA) #if (GGML_CUDA)
cmake_minimum_required(VERSION 3.18) # for CMAKE_CUDA_ARCHITECTURES # cmake_minimum_required(VERSION 3.18) # for CMAKE_CUDA_ARCHITECTURES
#
find_package(CUDAToolkit) # find_package(CUDAToolkit)
if (CUDAToolkit_FOUND) # if (CUDAToolkit_FOUND)
message(STATUS "CUDA found") # message(STATUS "CUDA found")
#
if (NOT DEFINED CMAKE_CUDA_ARCHITECTURES) # if (NOT DEFINED CMAKE_CUDA_ARCHITECTURES)
# 52 == lowest CUDA 12 standard # # 52 == lowest CUDA 12 standard
# 60 == f16 CUDA intrinsics # # 60 == f16 CUDA intrinsics
# 61 == integer CUDA intrinsics # # 61 == integer CUDA intrinsics
# 70 == compute capability at which unrolling a loop in mul_mat_q kernels is faster # # 70 == compute capability at which unrolling a loop in mul_mat_q kernels is faster
set(CMAKE_CUDA_ARCHITECTURES "52;61;70") # lowest CUDA 12 standard + lowest for integer intrinsics # set(CMAKE_CUDA_ARCHITECTURES "52;61;70") # lowest CUDA 12 standard + lowest for integer intrinsics
endif() # endif()
message(STATUS "Using CUDA architectures: ${CMAKE_CUDA_ARCHITECTURES}") # message(STATUS "Using CUDA architectures: ${CMAKE_CUDA_ARCHITECTURES}")
#
enable_language(CUDA) # enable_language(CUDA)
else() # else()
message(WARNING "CUDA not found") # message(WARNING "CUDA not found")
endif() # endif()
endif() #endif()
# whisper # whisper
@ -107,11 +107,12 @@ add_library(whisper
whisper-mel.hpp whisper-mel.hpp
) )
if (GGML_CUDA) # TODO: disabled because it relies on ggml internals that are no longer accessible (ggml-backend-impl.h, ggml-cuda/common.cuh, ..)
target_sources(whisper PRIVATE whisper-mel-cuda.cu) #if (GGML_CUDA)
# target_sources(whisper PRIVATE whisper-mel-cuda.cu)
target_link_libraries(whisper PRIVATE CUDA::cufft) #
endif() # target_link_libraries(whisper PRIVATE CUDA::cufft)
#endif()
# Set the version numbers # Set the version numbers
set_target_properties(whisper PROPERTIES set_target_properties(whisper PROPERTIES

3
src/whisper-mel-cuda.cu
View File

@ -2,8 +2,7 @@
#include "whisper-mel-cuda.hpp" #include "whisper-mel-cuda.hpp"
#include "whisper.h" #include "whisper.h"
#include <ggml-cuda/common.cuh> #include <ggml-backend.h>
#include <ggml-backend-impl.h>
#include <cuda.h> #include <cuda.h>
#include <cuda_runtime.h> #include <cuda_runtime.h>

4
src/whisper.cpp
View File

@ -3215,7 +3215,9 @@ struct mel_calc_cpu : public whisper_mel_calc {
} }
static whisper_mel_calc * whisper_mel_calc_create(ggml_backend_t backend, const whisper_filters & filters) { static whisper_mel_calc * whisper_mel_calc_create(ggml_backend_t backend, const whisper_filters & filters) {
#if defined(GGML_USE_CUDA) && !defined(GGML_USE_HIPBLAS) // TODO: disabled because it relies on ggml internals that are no longer accessible (ggml-backend-impl.h, ggml-cuda/common.cuh, ..)
//#if defined(GGML_USE_CUDA) && !defined(GGML_USE_HIPBLAS)
#if 0
if (ggml_backend_is_cuda(backend)) { if (ggml_backend_is_cuda(backend)) {
auto ret = whisper_mel_calc_create_cuda(backend, filters); auto ret = whisper_mel_calc_create_cuda(backend, filters);
if (ret) { if (ret) {