Merge remote-tracking branch 'ggerganov/master'

* ggerganov/master: (44 commits)
  ruby : Add low-level methods to transcribe (ggml-org#2585)
  models : add `q8_0` models to `download-ggml-model.sh` (ggml-org#2589)
  ruby : Follow source tree change (ggml-org#2580)
  whisper : use backend registry (#0)
  ggml/sched : do not skip views in pre-assignments
  whisper : adapt to new ggml (wip)
  talk-llama : sync llama.cpp
  sync : ggml
  ggml : sync resolve (skip) (#0)
  Add required ggml-base and backend libs to cmake pkg (llama/10407)
  cuda : fix CUDA_FLAGS not being applied (llama/10403)
  sycl : Add option to set the SYCL architecture for all targets (llama/10266)
  vulkan: Optimize soft_max (llama/10301)
  sycl: Revert MUL_MAT_OP support changes (llama/10385)
  cuda : only use native when supported by cmake (llama/10389)
  vulkan: remove use of null initializer (llama/10372)
  metal : fox offset integer overflows in im2col (ggml/1015)
  Vulkan: Fix device info output format specifiers (llama/10366)
  metal : add `GGML_UNARY_OP_ELU` kernel (ggml/1018)
  CUDA: fix MMV kernel being used for FP16 src1 (llama/10357)
  ...

Author: bygreencn

.github/workflows/bindings-ruby.yml

@@ -3,61 +3,41 @@ on:
   push:
     paths:
       - bindings/ruby/**
-      - src/whisper.cpp
-      - include/whisper.h
-      - ggml/src/ggml.c
-      - ggml/src/ggml-impl.h
-      - ggml/src/ggml-aarch64.h
-      - ggml/src/ggml-aarch64.c
-      - ggml/src/ggml-alloc.c
-      - ggml/src/ggml-backend-impl.h
-      - ggml/src/ggml-backend.cpp
-      - ggml/src/ggml-common.h
-      - ggml/src/ggml-quants.h
-      - ggml/src/ggml-quants.c
-      - ggml/src/ggml-cpu-impl.h
-      - ggml/src/ggml-metal.m
-      - ggml/src/ggml-metal.metal
-      - ggml/src/ggml-blas.cpp
-      - ggml/include/ggml.h
-      - ggml/include/ggml-alloc.h
-      - ggml/include/ggml-backend.h
-      - ggml/include/ggml-cuda.h
-      - ggml/include/ggml-kompute.h
-      - ggml/include/ggml-metal.h
-      - ggml/include/ggml-sycl.h
-      - ggml/include/ggml-vulkan.h
-      - ggml/include/ggml-blas.h
+      - src/**/*.c
+      - src/**/*.cpp
+      - src/**/*.h
+      - src/**/*.m
+      - src/**/*.metal
+      - include/**/*.c
+      - include/**/*.cpp
+      - include/**/*.h
+      - include/**/*.m
+      - include/**/*.metal
+      - ggml/**/*.c
+      - ggml/**/*.cpp
+      - ggml/**/*.h
+      - ggml/**/*.m
+      - ggml/**/*.metal
       - scripts/get-flags.mk
       - examples/dr_wav.h
   pull_request:
     paths:
       - bindings/ruby/**
-      - src/whisper.cpp
-      - include/whisper.h
-      - ggml/src/ggml.c
-      - ggml/src/ggml-impl.h
-      - ggml/src/ggml-aarch64.h
-      - ggml/src/ggml-aarch64.c
-      - ggml/src/ggml-alloc.c
-      - ggml/src/ggml-backend-impl.h
-      - ggml/src/ggml-backend.cpp
-      - ggml/src/ggml-common.h
-      - ggml/src/ggml-quants.h
-      - ggml/src/ggml-quants.c
-      - ggml/src/ggml-cpu-impl.h
-      - ggml/src/ggml-metal.m
-      - ggml/src/ggml-metal.metal
-      - ggml/src/ggml-blas.cpp
-      - ggml/include/ggml.h
-      - ggml/include/ggml-alloc.h
-      - ggml/include/ggml-backend.h
-      - ggml/include/ggml-cuda.h
-      - ggml/include/ggml-kompute.h
-      - ggml/include/ggml-metal.h
-      - ggml/include/ggml-sycl.h
-      - ggml/include/ggml-vulkan.h
-      - ggml/include/ggml-blas.h
+      - src/**/*.c
+      - src/**/*.cpp
+      - src/**/*.h
+      - src/**/*.m
+      - src/**/*.metal
+      - include/**/*.c
+      - include/**/*.cpp
+      - include/**/*.h
+      - include/**/*.m
+      - include/**/*.metal
+      - ggml/**/*.c
+      - ggml/**/*.cpp
+      - ggml/**/*.h
+      - ggml/**/*.m
+      - ggml/**/*.metal
       - scripts/get-flags.mk
       - examples/dr_wav.h
 
@@ -70,6 +50,6 @@ jobs:
     steps:
       - uses: ruby/setup-ruby@v1
         with:
-          ruby-version: '3.0'
+          ruby-version: '3.1'
       - uses: actions/checkout@v4
       - run: rake test

.gitignore

@@ -1,5 +1,6 @@
 *.o
 *.a
+*.d
 .cache/
 .coreml/
 .test/
@@ -19,6 +20,9 @@ build-*/
 .swiftpm
 *.metallib
 
+ggml-metal-embed.metal
+ggml-metal-embed.metal.tmp
+
 /main
 /stream
 /command