whispercpp 1.3.1 → 1.3.2

data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/rope_common.comp

@@ -0,0 +1,71 @@
+#include "common.comp"
+
+#define GGML_ROPE_TYPE_NEOX 2
+
+// TODO: use a local size of 32 or more (Metal uses 1024)
+layout(local_size_x = 1) in;
+
+layout (push_constant) uniform parameter {
+    uint inAOff;
+    uint inBOff;
+    uint inCOff;
+    uint outOff;
+    int n_dims;
+    int mode;
+    int n_ctx_orig;
+    float freq_base;
+    float freq_scale;
+    bool has_freq_factors;
+    float ext_factor;
+    float attn_factor;
+    float beta_fast;
+    float beta_slow;
+    uint nb00;
+    uint nb01;
+    uint nb02;
+    uint nb03;
+    int ne0;
+    uint nb0;
+    uint nb1;
+    uint nb2;
+    uint nb3;
+} pcs;
+
+float rope_yarn_ramp(const float low, const float high, const float i0) {
+    const float y = (i0 / 2 - low) / max(0.001f, high - low);
+    return 1.0f - min(1.0f, max(0.0f, y));
+}
+
+// YaRN algorithm based on LlamaYaRNScaledRotaryEmbedding.py from https://github.com/jquesnelle/yarn
+// MIT licensed. Copyright (c) 2023 Jeffrey Quesnelle and Bowen Peng.
+void rope_yarn(
+    float theta_extrap, float freq_scale, float corr_dims[2], float i0, float ext_factor, float mscale,
+    out float cos_theta, out float sin_theta
+) {
+    // Get n-d rotational scaling corrected for extrapolation
+    float theta_interp = freq_scale * theta_extrap;
+    float theta = theta_interp;
+    if (ext_factor != 0.0f) {
+        float ramp_mix = rope_yarn_ramp(corr_dims[0], corr_dims[1], i0) * ext_factor;
+        theta = theta_interp * (1 - ramp_mix) + theta_extrap * ramp_mix;
+
+        // Get n-d magnitude scaling corrected for interpolation
+        mscale *= 1.0f + 0.1f * log(1.0f / freq_scale);
+    }
+    cos_theta = cos(theta) * mscale;
+    sin_theta = sin(theta) * mscale;
+}
+
+// Apparently solving `n_rot = 2pi * x * base^((2 * max_pos_emb) / n_dims)` for x, we get
+// `corr_fac(n_rot) = n_dims * log(max_pos_emb / (n_rot * 2pi)) / (2 * log(base))`
+float rope_yarn_corr_factor(int n_dims, int n_ctx_orig, float n_rot, float base) {
+    return n_dims * log(n_ctx_orig / (n_rot * TWOPI_F)) / (2 * log(base));
+}
+
+void rope_yarn_corr_dims(
+    int n_dims, int n_ctx_orig, float freq_base, float beta_fast, float beta_slow, out float dims[2]
+) {
+    // start and end correction dims
+    dims[0] = max(0.0f,         floor(rope_yarn_corr_factor(n_dims, n_ctx_orig, beta_fast, freq_base)));
+    dims[1] = min(n_dims - 1.0f, ceil(rope_yarn_corr_factor(n_dims, n_ctx_orig, beta_slow, freq_base)));
+}

data/ext/sources/ggml/src/ggml-metal/CMakeLists.txt

@@ -0,0 +1,120 @@
+find_library(FOUNDATION_LIBRARY Foundation REQUIRED)
+find_library(METAL_FRAMEWORK    Metal      REQUIRED)
+find_library(METALKIT_FRAMEWORK MetalKit   REQUIRED)
+
+message(STATUS "Metal framework found")
+
+ggml_add_backend_library(ggml-metal
+                         ggml-metal.m
+                        )
+
+target_link_libraries(ggml-metal PRIVATE
+                      ${FOUNDATION_LIBRARY}
+                      ${METAL_FRAMEWORK}
+                      ${METALKIT_FRAMEWORK}
+                      )
+
+if (GGML_METAL_NDEBUG)
+    add_compile_definitions(GGML_METAL_NDEBUG)
+endif()
+
+if (GGML_METAL_USE_BF16)
+    add_compile_definitions(GGML_METAL_USE_BF16)
+endif()
+
+# copy metal files to bin directory
+configure_file(../ggml-common.h  ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-common.h     COPYONLY)
+configure_file(ggml-metal.metal  ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.metal  COPYONLY)
+configure_file(ggml-metal-impl.h ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal-impl.h COPYONLY)
+
+set(METALLIB_COMMON "${CMAKE_CURRENT_SOURCE_DIR}/../ggml-common.h")
+if (GGML_METAL_EMBED_LIBRARY)
+    enable_language(ASM)
+
+    add_compile_definitions(GGML_METAL_EMBED_LIBRARY)
+
+    set(METALLIB_SOURCE "${CMAKE_CURRENT_SOURCE_DIR}/ggml-metal.metal")
+    set(METALLIB_IMPL   "${CMAKE_CURRENT_SOURCE_DIR}/ggml-metal-impl.h")
+
+    file(MAKE_DIRECTORY "${CMAKE_BINARY_DIR}/autogenerated")
+
+    # merge ggml-common.h and ggml-metal.metal into a single file
+    set(METALLIB_EMBED_ASM        "${CMAKE_BINARY_DIR}/autogenerated/ggml-metal-embed.s")
+    set(METALLIB_SOURCE_EMBED     "${CMAKE_BINARY_DIR}/autogenerated/ggml-metal-embed.metal")
+    set(METALLIB_SOURCE_EMBED_TMP "${CMAKE_BINARY_DIR}/autogenerated/ggml-metal-embed.metal.tmp")
+
+    add_custom_command(
+        OUTPUT ${METALLIB_EMBED_ASM}
+        COMMAND echo "Embedding Metal library"
+        COMMAND sed -e '/__embed_ggml-common.h__/r         ${METALLIB_COMMON}' -e '/__embed_ggml-common.h__/d'         < ${METALLIB_SOURCE}           > ${METALLIB_SOURCE_EMBED_TMP}
+        COMMAND sed -e '/\#include \"ggml-metal-impl.h\"/r ${METALLIB_IMPL}'   -e '/\#include \"ggml-metal-impl.h\"/d' < ${METALLIB_SOURCE_EMBED_TMP} > ${METALLIB_SOURCE_EMBED}
+        COMMAND echo ".section __DATA,__ggml_metallib"          >  ${METALLIB_EMBED_ASM}
+        COMMAND echo ".globl _ggml_metallib_start"              >> ${METALLIB_EMBED_ASM}
+        COMMAND echo "_ggml_metallib_start:"                    >> ${METALLIB_EMBED_ASM}
+        COMMAND echo ".incbin \\\"${METALLIB_SOURCE_EMBED}\\\"" >> ${METALLIB_EMBED_ASM}
+        COMMAND echo ".globl _ggml_metallib_end"                >> ${METALLIB_EMBED_ASM}
+        COMMAND echo "_ggml_metallib_end:"                      >> ${METALLIB_EMBED_ASM}
+        DEPENDS ../ggml-common.h ggml-metal.metal ggml-metal-impl.h
+        COMMENT "Generate assembly for embedded Metal library"
+    )
+
+    target_sources(ggml-metal PRIVATE ${METALLIB_EMBED_ASM})
+else()
+    if (GGML_METAL_SHADER_DEBUG)
+        # custom command to do the following:
+        #   xcrun -sdk macosx metal    -fno-fast-math -c ggml-metal.metal -o ggml-metal.air
+        #   xcrun -sdk macosx metallib                   ggml-metal.air   -o default.metallib
+        #
+        # note: this is the only way I found to disable fast-math in Metal. it's ugly, but at least it works
+        #       disabling fast math is needed in order to pass tests/test-backend-ops
+        # note: adding -fno-inline fixes the tests when using MTL_SHADER_VALIDATION=1
+        # note: unfortunately, we have to call it default.metallib instead of ggml.metallib
+        #       ref: https://github.com/ggerganov/whisper.cpp/issues/1720
+        set(XC_FLAGS -fno-fast-math -fno-inline -g)
+    else()
+        set(XC_FLAGS -O3)
+    endif()
+
+    # Append macOS metal versioning flags
+    if (GGML_METAL_MACOSX_VERSION_MIN)
+        message(STATUS "Adding  -mmacosx-version-min=${GGML_METAL_MACOSX_VERSION_MIN} flag to metal compilation")
+        list   (APPEND XC_FLAGS -mmacosx-version-min=${GGML_METAL_MACOSX_VERSION_MIN})
+    endif()
+
+    if (GGML_METAL_STD)
+        message(STATUS "Adding  -std=${GGML_METAL_STD} flag to metal compilation")
+        list   (APPEND XC_FLAGS -std=${GGML_METAL_STD})
+    endif()
+
+    add_custom_command(
+        OUTPUT ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib
+        COMMAND xcrun -sdk macosx metal ${XC_FLAGS} -c ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.metal -o - |
+            xcrun -sdk macosx metallib - -o ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib
+        COMMAND rm -f ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-common.h
+        COMMAND rm -f ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/ggml-metal.metal
+        DEPENDS ggml-metal.metal ${METALLIB_COMMON}
+        COMMENT "Compiling Metal kernels"
+        )
+
+    # FIXME: only add to the ggml-metal target?
+    add_custom_target(
+        ggml-metal-lib ALL
+        DEPENDS ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib
+        )
+endif() # GGML_METAL_EMBED_LIBRARY
+
+if (NOT GGML_METAL_EMBED_LIBRARY)
+    install(
+        FILES src/ggml-metal/ggml-metal.metal
+        PERMISSIONS
+            OWNER_READ
+            OWNER_WRITE
+            GROUP_READ
+            WORLD_READ
+        DESTINATION ${CMAKE_INSTALL_BINDIR})
+
+        install(
+            FILES ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/default.metallib
+            DESTINATION ${CMAKE_INSTALL_BINDIR}
+        )
+endif()