wip : initial WASM port

Works but it is very slow because no SIMD is used.
For example, jfk.wav is processed in ~23 seconds using "tiny.en" model

CMakeLists.txt

@@ -7,20 +7,37 @@ set(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_PREFIX}/lib")
 
 if(CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR)
     set(WHISPER_STANDALONE ON)
+    include(cmake/GitVars.cmake)
+    include(cmake/BuildTypes.cmake)
 else()
     set(WHISPER_STANDALONE OFF)
 endif()
 
+if (EMSCRIPTEN)
+    set(BUILD_SHARED_LIBS_DEFAULT OFF)
+
+    option(WHISPER_WASM_SINGLE_FILE "whisper: embed WASM inside the generated whisper.js" ON)
+else()
+    if (MINGW)
+        set(BUILD_SHARED_LIBS_DEFAULT OFF)
+    else()
+        set(BUILD_SHARED_LIBS_DEFAULT ON)
+    endif()
+endif()
+
 # options
 
+option(BUILD_SHARED_LIBS               "whisper: build shared libs" ${BUILD_SHARED_LIBS_DEFAULT})
+
 option(WHISPER_ALL_WARNINGS            "whisper: enable all compiler warnings"                   ON)
 option(WHISPER_ALL_WARNINGS_3RD_PARTY  "whisper: enable all compiler warnings in 3rd party libs" OFF)
 
 option(WHISPER_SANITIZE_THREAD         "whisper: enable thread sanitizer"    OFF)
 option(WHISPER_SANITIZE_ADDRESS        "whisper: enable address sanitizer"   OFF)
 option(WHISPER_SANITIZE_UNDEFINED      "whisper: enable undefined sanitizer" OFF)
 
-option(WHISPER_BUILD_TESTS             "whisper: build tests" ${WHISPER_STANDALONE})
+option(WHISPER_BUILD_TESTS             "whisper: build tests"    ${WHISPER_STANDALONE})
+option(WHISPER_BUILD_EXAMPLES          "whisper: build examples" ${WHISPER_STANDALONE})
 
 option(WHISPER_SUPPORT_SDL2            "whisper: support for libSDL2" OFF)
 
@@ -69,16 +86,6 @@ if (APPLE AND NOT WHISPER_NO_ACCELERATE)
     endif()
 endif()
 
-if (WHISPER_SUPPORT_SDL2)
-    # SDL2
-    find_package(SDL2 REQUIRED)
-
-    string(STRIP "${SDL2_LIBRARIES}" SDL2_LIBRARIES)
-
-    message(STATUS "SDL2_INCLUDE_DIRS = ${SDL2_INCLUDE_DIRS}")
-    message(STATUS "SDL2_LIBRARIES = ${SDL2_LIBRARIES}")
-endif()
-
 # compiler flags
 
 if (NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
@@ -120,6 +127,11 @@ else()
     endif()
 endif()
 
+if (EMSCRIPTEN)
+    set(CMAKE_C_FLAGS   "${CMAKE_C_FLAGS}   -pthread -msimd128")
+    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -pthread")
+endif()
+
 # whisper - this is the main library of the project
 
 set(TARGET whisper)
@@ -154,24 +166,46 @@ install(TARGETS ${TARGET}
     ARCHIVE DESTINATION lib/static
     )
 
+# bindings
+
+add_subdirectory(bindings)
+
 # programs, examples and tests
 
 if (WHISPER_STANDALONE)
-    # main
-    set(TARGET main)
-    add_executable(${TARGET} main.cpp)
-    target_link_libraries(${TARGET} PRIVATE whisper ${CMAKE_THREAD_LIBS_INIT})
-
-    if (WHISPER_SUPPORT_SDL2)
-        # stream
-        set(TARGET stream)
-        add_executable(${TARGET} stream.cpp)
-        target_include_directories(${TARGET} PRIVATE ${SDL2_INCLUDE_DIRS})
-        target_link_libraries(${TARGET} PRIVATE whisper ${SDL2_LIBRARIES} ${CMAKE_THREAD_LIBS_INIT})
-    endif ()
+    if (NOT EMSCRIPTEN)
+        # TODO: move to examples
+        # main
+        set(TARGET main)
+        add_executable(${TARGET} main.cpp)
+        target_link_libraries(${TARGET} PRIVATE whisper ${CMAKE_THREAD_LIBS_INIT})
+
+        # TODO: move to examples
+        if (WHISPER_SUPPORT_SDL2)
+            if (WHISPER_SUPPORT_SDL2)
+                # SDL2
+                find_package(SDL2 REQUIRED)
+
+                string(STRIP "${SDL2_LIBRARIES}" SDL2_LIBRARIES)
+
+                message(STATUS "SDL2_INCLUDE_DIRS = ${SDL2_INCLUDE_DIRS}")
+                message(STATUS "SDL2_LIBRARIES = ${SDL2_LIBRARIES}")
+            endif()
+
+            # stream
+            set(TARGET stream)
+            add_executable(${TARGET} stream.cpp)
+            target_include_directories(${TARGET} PRIVATE ${SDL2_INCLUDE_DIRS})
+            target_link_libraries(${TARGET} PRIVATE whisper ${SDL2_LIBRARIES} ${CMAKE_THREAD_LIBS_INIT})
+        endif ()
+    endif()
 
     if (WHISPER_BUILD_TESTS)
         enable_testing()
         add_subdirectory(tests)
     endif ()
+
+    if (WHISPER_BUILD_EXAMPLES)
+        add_subdirectory(examples)
+    endif()
 endif ()

bindings/CMakeLists.txt

@@ -0,0 +1,3 @@
+if (EMSCRIPTEN)
+    add_subdirectory(javascript)
+endif()

bindings/javascript/.gitignore

@@ -0,0 +1 @@
+publish.log

bindings/javascript/CMakeLists.txt

@@ -0,0 +1,34 @@
+set(TARGET libwhisper)
+
+add_executable(${TARGET}
+    emscripten.cpp
+    )
+
+target_link_libraries(${TARGET} PRIVATE
+    whisper
+    )
+
+unset(EXTRA_FLAGS)
+if (WHISPER_WASM_SINGLE_FILE)
+    set(EXTRA_FLAGS "-s SINGLE_FILE=1")
+    message(STATUS "Embedding WASM inside whisper.js")
+
+    add_custom_command(
+        TARGET libwhisper POST_BUILD
+        COMMAND ${CMAKE_COMMAND} -E copy
+        ${CMAKE_BINARY_DIR}/bin/libwhisper.js
+        ${CMAKE_CURRENT_SOURCE_DIR}/whisper.js
+        )
+endif()
+
+set_target_properties(${TARGET} PROPERTIES LINK_FLAGS " \
+    --bind \
+    -s MODULARIZE=1 \
+    -s ASSERTIONS=1 \
+    -s USE_PTHREADS=1 \
+    -s PTHREAD_POOL_SIZE=8 \
+    -s TOTAL_MEMORY=536870912 \
+    -s FORCE_FILESYSTEM=1 \
+    -s EXPORT_NAME=\"'whisper_factory'\" \
+    ${EXTRA_FLAGS} \
+    ")

bindings/javascript/emscripten.cpp

@@ -0,0 +1,70 @@
+#include "whisper.h"
+
+#include <emscripten.h>
+#include <emscripten/bind.h>
+
+#include <vector>
+
+std::vector<struct whisper_context *> g_contexts(4, nullptr);
+
+EMSCRIPTEN_BINDINGS(whisper) {
+    emscripten::function("init", emscripten::optional_override([](const std::string & path_model) {
+        for (size_t i = 0; i < g_contexts.size(); ++i) {
+            if (g_contexts[i] == nullptr) {
+                g_contexts[i] = whisper_init(path_model.c_str());
+                return i + 1;
+            }
+        }
+
+        return (size_t) 0;
+    }));
+
+    emscripten::function("free", emscripten::optional_override([](size_t index) {
+        --index;
+
+        if (index < g_contexts.size()) {
+            whisper_free(g_contexts[index]);
+            g_contexts[index] = nullptr;
+        }
+    }));
+
+    emscripten::function("full_default", emscripten::optional_override([](size_t index, const emscripten::val & audio) {
+        --index;
+
+        if (index >= g_contexts.size()) {
+            return -1;
+        }
+
+        if (g_contexts[index] == nullptr) {
+            return -2;
+        }
+
+        struct whisper_full_params params = whisper_full_default_params(whisper_sampling_strategy::WHISPER_SAMPLING_GREEDY);
+
+        params.print_realtime       = true;
+        params.print_progress       = false;
+        params.print_timestamps     = true;
+        params.print_special_tokens = false;
+        params.translate            = false;
+        params.language             = "en";
+        params.n_threads            = 4;
+        params.offset_ms            = 0;
+
+        std::vector<float> pcmf32;
+        const int n = audio["length"].as<int>();
+
+        emscripten::val heap = emscripten::val::module_property("HEAPU8");
+        emscripten::val memory = heap["buffer"];
+
+        pcmf32.resize(n);
+
+        emscripten::val memoryView = audio["constructor"].new_(memory, reinterpret_cast<uintptr_t>(pcmf32.data()), n);
+        memoryView.call<void>("set", audio);
+
+        int ret = whisper_full(g_contexts[index], params, pcmf32.data(), pcmf32.size());
+
+        whisper_print_timings(g_contexts[index]);
+
+        return ret;
+    }));
+}

cmake/BuildTypes.cmake

@@ -0,0 +1,54 @@
+# Add new build types
+
+# ReleaseGG - Release with enabled asserts
+
+SET(CMAKE_CXX_FLAGS_RELEASEGG
+    "-O3"
+    CACHE STRING "Flags used by the c++ compiler during release builds with enabled asserts."
+    FORCE )
+SET(CMAKE_C_FLAGS_RELEASEGG
+    "-O3"
+    CACHE STRING "Flags used by the compiler during release builds with enabled asserts."
+    FORCE )
+SET(CMAKE_EXE_LINKER_FLAGS_RELEASEGG
+    ""
+    CACHE STRING "Flags used for linking binaries during release builds with enabled asserts."
+    FORCE )
+SET(CMAKE_SHARED_LINKER_FLAGS_RELEASEGG
+    ""
+    CACHE STRING "Flags used by the shared libraries linker during release builds with enabled asserts."
+    FORCE )
+MARK_AS_ADVANCED(
+    CMAKE_CXX_FLAGS_RELEASEGG
+    CMAKE_C_FLAGS_RELEASEGG
+    CMAKE_EXE_LINKER_FLAGS_RELEASEGG
+    CMAKE_SHARED_LINKER_FLAGS_RELEASEGG )
+
+# RelWithDebInfoGG - RelWithDebInfo with enabled asserts
+
+SET(CMAKE_CXX_FLAGS_RELWITHDEBINFOGG
+    "-O2 -g"
+    CACHE STRING "Flags used by the c++ compiler during release builds with debug symbols and enabled asserts."
+    FORCE )
+SET(CMAKE_C_FLAGS_RELWITHDEBINFOGG
+    "-O2 -g"
+    CACHE STRING "Flags used by the compiler during release builds with debug symbols and enabled asserts."
+    FORCE )
+SET(CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFOGG
+    ""
+    CACHE STRING "Flags used for linking binaries during release builds with debug symbols and enabled asserts."
+    FORCE )
+SET(CMAKE_SHARED_LINKER_FLAGS_RELWITHDEBINFOGG
+    ""
+    CACHE STRING "Flags used by the shared libraries linker during release builds with debug symbols and enabled asserts."
+    FORCE )
+MARK_AS_ADVANCED(
+    CMAKE_CXX_FLAGS_RELWITHDEBINFOGG
+    CMAKE_C_FLAGS_RELWITHDEBINFOGG
+    CMAKE_EXE_LINKER_FLAGS_RELWITHDEBINFOGG
+    CMAKE_SHARED_LINKER_FLAGS_RELWITHDEBINFOGG )
+
+if (NOT XCODE AND NOT MSVC AND NOT CMAKE_BUILD_TYPE)
+    set(CMAKE_BUILD_TYPE Release CACHE STRING "Build type" FORCE)
+    set_property(CACHE CMAKE_BUILD_TYPE PROPERTY STRINGS "Debug" "Release" "MinSizeRel" "RelWithDebInfo" "ReleaseGG" "RelWithDebInfoGG")
+endif()

cmake/GitVars.cmake

@@ -0,0 +1,22 @@
+find_package(Git)
+
+# the commit's SHA1
+execute_process(COMMAND
+    "${GIT_EXECUTABLE}" describe --match=NeVeRmAtCh --always --abbrev=8
+    WORKING_DIRECTORY "${CMAKE_SOURCE_DIR}"
+    OUTPUT_VARIABLE GIT_SHA1
+    ERROR_QUIET OUTPUT_STRIP_TRAILING_WHITESPACE)
+
+# the date of the commit
+execute_process(COMMAND
+    "${GIT_EXECUTABLE}" log -1 --format=%ad --date=local
+    WORKING_DIRECTORY "${CMAKE_SOURCE_DIR}"
+    OUTPUT_VARIABLE GIT_DATE
+    ERROR_QUIET OUTPUT_STRIP_TRAILING_WHITESPACE)
+
+# the subject of the commit
+execute_process(COMMAND
+    "${GIT_EXECUTABLE}" log -1 --format=%s
+    WORKING_DIRECTORY "${CMAKE_SOURCE_DIR}"
+    OUTPUT_VARIABLE GIT_COMMIT_SUBJECT
+    ERROR_QUIET OUTPUT_STRIP_TRAILING_WHITESPACE)

examples/CMakeLists.txt

@@ -0,0 +1,14 @@
+# dependencies
+
+find_package(Threads REQUIRED)
+
+# third-party
+
+#add_subdirectory(third-party)
+
+# examples
+
+if (EMSCRIPTEN)
+    add_subdirectory(whisper.wasm)
+else()
+endif()

examples/whisper.wasm/CMakeLists.txt

@@ -0,0 +1,4 @@
+set(TARGET whisper.wasm)
+
+configure_file(${CMAKE_CURRENT_SOURCE_DIR}/index-tmpl.html        ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/${TARGET}/index.html @ONLY)
+configure_file(${CMAKE_SOURCE_DIR}/bindings/javascript/whisper.js ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/${TARGET}/whisper.js  COPYONLY)

examples/whisper.wasm/README.md

@@ -0,0 +1,3 @@
+# whisper.wasm
+
+Live demo: https://whisper.ggerganov.com