whispercpp 1.3.1 → 1.3.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (797) hide show
  1. checksums.yaml +4 -4
  2. data/.gitignore +4 -3
  3. data/README.md +92 -31
  4. data/Rakefile +26 -7
  5. data/ext/.gitignore +5 -7
  6. data/ext/dependencies.rb +61 -0
  7. data/ext/extconf.rb +21 -198
  8. data/ext/options.rb +221 -0
  9. data/ext/ruby_whisper.c +159 -0
  10. data/ext/ruby_whisper.h +17 -2
  11. data/ext/ruby_whisper_context.c +641 -0
  12. data/ext/ruby_whisper_error.c +52 -0
  13. data/ext/ruby_whisper_model.c +232 -0
  14. data/ext/ruby_whisper_params.c +1301 -0
  15. data/ext/ruby_whisper_segment.c +143 -0
  16. data/ext/ruby_whisper_transcribe.cpp +87 -0
  17. data/ext/ruby_whisper_vad_params.c +288 -0
  18. data/ext/sources/.dockerignore +3 -0
  19. data/ext/sources/.github/workflows/bindings-ruby.yml +21 -0
  20. data/ext/sources/CMakeGraphVizOptions.cmake +8 -0
  21. data/ext/sources/CMakeLists.txt +251 -0
  22. data/ext/sources/bindings/javascript/CMakeLists.txt +41 -0
  23. data/ext/sources/bindings/javascript/emscripten.cpp +93 -0
  24. data/ext/sources/bindings/javascript/libwhisper.worker.js +1 -0
  25. data/ext/sources/bindings/javascript/package-tmpl.json +26 -0
  26. data/ext/sources/bindings/javascript/package.json +26 -0
  27. data/ext/sources/bindings/javascript/whisper.js +19 -0
  28. data/ext/sources/build-xcframework.sh +547 -0
  29. data/ext/sources/ci/run.sh +336 -0
  30. data/ext/sources/close-issue.yml +28 -0
  31. data/ext/sources/cmake/DefaultTargetOptions.cmake +16 -0
  32. data/ext/sources/cmake/FindFFmpeg.cmake +163 -0
  33. data/ext/sources/cmake/build-info.cmake +60 -0
  34. data/ext/sources/cmake/git-vars.cmake +22 -0
  35. data/ext/sources/cmake/whisper-config.cmake.in +65 -0
  36. data/ext/sources/cmake/whisper.pc.in +10 -0
  37. data/ext/sources/examples/CMakeLists.txt +124 -0
  38. data/ext/sources/examples/addon.node/CMakeLists.txt +31 -0
  39. data/ext/sources/examples/addon.node/__test__/whisper.spec.js +37 -0
  40. data/ext/sources/examples/addon.node/addon.cpp +438 -0
  41. data/ext/sources/examples/addon.node/index.js +54 -0
  42. data/ext/sources/examples/addon.node/package.json +16 -0
  43. data/ext/sources/examples/bench/CMakeLists.txt +8 -0
  44. data/ext/sources/examples/bench/bench.cpp +175 -0
  45. data/ext/sources/examples/bench.wasm/CMakeLists.txt +49 -0
  46. data/ext/sources/examples/bench.wasm/emscripten.cpp +87 -0
  47. data/ext/sources/examples/bench.wasm/index-tmpl.html +284 -0
  48. data/ext/sources/examples/cli/CMakeLists.txt +8 -0
  49. data/ext/sources/examples/cli/cli.cpp +1294 -0
  50. data/ext/sources/examples/coi-serviceworker.js +146 -0
  51. data/ext/sources/examples/command/CMakeLists.txt +10 -0
  52. data/ext/sources/examples/command/command.cpp +776 -0
  53. data/ext/sources/examples/command/commands.txt +9 -0
  54. data/ext/sources/examples/command.wasm/CMakeLists.txt +50 -0
  55. data/ext/sources/examples/command.wasm/emscripten.cpp +327 -0
  56. data/ext/sources/examples/command.wasm/index-tmpl.html +414 -0
  57. data/ext/sources/examples/common-ggml.cpp +238 -0
  58. data/ext/sources/examples/common-ggml.h +18 -0
  59. data/ext/sources/examples/common-sdl.cpp +227 -0
  60. data/ext/sources/examples/common-sdl.h +49 -0
  61. data/ext/sources/examples/common-whisper.cpp +168 -0
  62. data/ext/sources/examples/common-whisper.h +24 -0
  63. data/ext/sources/examples/common.cpp +675 -0
  64. data/ext/sources/examples/common.h +322 -0
  65. data/ext/sources/examples/deprecation-warning/CMakeLists.txt +6 -0
  66. data/ext/sources/examples/deprecation-warning/deprecation-warning.cpp +38 -0
  67. data/ext/sources/examples/ffmpeg-transcode.cpp +368 -0
  68. data/ext/sources/examples/generate-karaoke.sh +57 -0
  69. data/ext/sources/examples/grammar-parser.cpp +423 -0
  70. data/ext/sources/examples/grammar-parser.h +29 -0
  71. data/ext/sources/examples/helpers.js +191 -0
  72. data/ext/sources/examples/json.hpp +24596 -0
  73. data/ext/sources/examples/livestream.sh +112 -0
  74. data/ext/sources/examples/lsp/CMakeLists.txt +9 -0
  75. data/ext/sources/examples/lsp/lsp.cpp +467 -0
  76. data/ext/sources/examples/lsp/whisper.vim +362 -0
  77. data/ext/sources/examples/miniaudio.h +93468 -0
  78. data/ext/sources/examples/python/test_whisper_processor.py +7 -0
  79. data/ext/sources/examples/python/whisper_processor.py +54 -0
  80. data/ext/sources/examples/quantize/CMakeLists.txt +6 -0
  81. data/ext/sources/examples/quantize/quantize.cpp +223 -0
  82. data/ext/sources/examples/server/CMakeLists.txt +12 -0
  83. data/ext/sources/examples/server/bench.js +29 -0
  84. data/ext/sources/examples/server/httplib.h +10497 -0
  85. data/ext/sources/examples/server/server.cpp +1091 -0
  86. data/ext/sources/examples/server.py +115 -0
  87. data/ext/sources/examples/stb_vorbis.c +5584 -0
  88. data/ext/sources/examples/stream/CMakeLists.txt +10 -0
  89. data/ext/sources/examples/stream/stream.cpp +429 -0
  90. data/ext/sources/examples/stream.wasm/CMakeLists.txt +49 -0
  91. data/ext/sources/examples/stream.wasm/emscripten.cpp +216 -0
  92. data/ext/sources/examples/stream.wasm/index-tmpl.html +414 -0
  93. data/ext/sources/examples/sycl/CMakeLists.txt +9 -0
  94. data/ext/sources/examples/sycl/build.sh +22 -0
  95. data/ext/sources/examples/sycl/ls-sycl-device.cpp +11 -0
  96. data/ext/sources/examples/sycl/run-whisper.sh +17 -0
  97. data/ext/sources/examples/talk-llama/CMakeLists.txt +40 -0
  98. data/ext/sources/examples/talk-llama/eleven-labs.py +80 -0
  99. data/ext/sources/examples/talk-llama/llama-adapter.cpp +388 -0
  100. data/ext/sources/examples/talk-llama/llama-adapter.h +76 -0
  101. data/ext/sources/examples/talk-llama/llama-arch.cpp +1746 -0
  102. data/ext/sources/examples/talk-llama/llama-arch.h +437 -0
  103. data/ext/sources/examples/talk-llama/llama-batch.cpp +374 -0
  104. data/ext/sources/examples/talk-llama/llama-batch.h +89 -0
  105. data/ext/sources/examples/talk-llama/llama-chat.cpp +663 -0
  106. data/ext/sources/examples/talk-llama/llama-chat.h +58 -0
  107. data/ext/sources/examples/talk-llama/llama-context.cpp +2676 -0
  108. data/ext/sources/examples/talk-llama/llama-context.h +276 -0
  109. data/ext/sources/examples/talk-llama/llama-cparams.cpp +5 -0
  110. data/ext/sources/examples/talk-llama/llama-cparams.h +41 -0
  111. data/ext/sources/examples/talk-llama/llama-grammar.cpp +1229 -0
  112. data/ext/sources/examples/talk-llama/llama-grammar.h +173 -0
  113. data/ext/sources/examples/talk-llama/llama-graph.cpp +1618 -0
  114. data/ext/sources/examples/talk-llama/llama-graph.h +640 -0
  115. data/ext/sources/examples/talk-llama/llama-hparams.cpp +95 -0
  116. data/ext/sources/examples/talk-llama/llama-hparams.h +190 -0
  117. data/ext/sources/examples/talk-llama/llama-impl.cpp +167 -0
  118. data/ext/sources/examples/talk-llama/llama-impl.h +61 -0
  119. data/ext/sources/examples/talk-llama/llama-io.cpp +15 -0
  120. data/ext/sources/examples/talk-llama/llama-io.h +35 -0
  121. data/ext/sources/examples/talk-llama/llama-kv-cache.cpp +2739 -0
  122. data/ext/sources/examples/talk-llama/llama-kv-cache.h +502 -0
  123. data/ext/sources/examples/talk-llama/llama-kv-cells.h +379 -0
  124. data/ext/sources/examples/talk-llama/llama-memory.cpp +1 -0
  125. data/ext/sources/examples/talk-llama/llama-memory.h +32 -0
  126. data/ext/sources/examples/talk-llama/llama-mmap.cpp +600 -0
  127. data/ext/sources/examples/talk-llama/llama-mmap.h +68 -0
  128. data/ext/sources/examples/talk-llama/llama-model-loader.cpp +1138 -0
  129. data/ext/sources/examples/talk-llama/llama-model-loader.h +169 -0
  130. data/ext/sources/examples/talk-llama/llama-model-saver.cpp +281 -0
  131. data/ext/sources/examples/talk-llama/llama-model-saver.h +37 -0
  132. data/ext/sources/examples/talk-llama/llama-model.cpp +13814 -0
  133. data/ext/sources/examples/talk-llama/llama-model.h +425 -0
  134. data/ext/sources/examples/talk-llama/llama-quant.cpp +966 -0
  135. data/ext/sources/examples/talk-llama/llama-quant.h +1 -0
  136. data/ext/sources/examples/talk-llama/llama-sampling.cpp +2575 -0
  137. data/ext/sources/examples/talk-llama/llama-sampling.h +32 -0
  138. data/ext/sources/examples/talk-llama/llama-vocab.cpp +3340 -0
  139. data/ext/sources/examples/talk-llama/llama-vocab.h +131 -0
  140. data/ext/sources/examples/talk-llama/llama.cpp +354 -0
  141. data/ext/sources/examples/talk-llama/llama.h +1377 -0
  142. data/ext/sources/examples/talk-llama/prompts/talk-alpaca.txt +23 -0
  143. data/ext/sources/examples/talk-llama/speak +40 -0
  144. data/ext/sources/examples/talk-llama/speak.bat +1 -0
  145. data/ext/sources/examples/talk-llama/speak.ps1 +14 -0
  146. data/ext/sources/examples/talk-llama/talk-llama.cpp +808 -0
  147. data/ext/sources/examples/talk-llama/unicode-data.cpp +7034 -0
  148. data/ext/sources/examples/talk-llama/unicode-data.h +20 -0
  149. data/ext/sources/examples/talk-llama/unicode.cpp +849 -0
  150. data/ext/sources/examples/talk-llama/unicode.h +66 -0
  151. data/ext/sources/examples/vad-speech-segments/CMakeLists.txt +8 -0
  152. data/ext/sources/examples/vad-speech-segments/speech.cpp +143 -0
  153. data/ext/sources/examples/wchess/CMakeLists.txt +10 -0
  154. data/ext/sources/examples/wchess/libwchess/CMakeLists.txt +19 -0
  155. data/ext/sources/examples/wchess/libwchess/Chessboard.cpp +803 -0
  156. data/ext/sources/examples/wchess/libwchess/Chessboard.h +33 -0
  157. data/ext/sources/examples/wchess/libwchess/WChess.cpp +193 -0
  158. data/ext/sources/examples/wchess/libwchess/WChess.h +63 -0
  159. data/ext/sources/examples/wchess/libwchess/test-chessboard.cpp +117 -0
  160. data/ext/sources/examples/wchess/wchess.cmd/CMakeLists.txt +8 -0
  161. data/ext/sources/examples/wchess/wchess.cmd/wchess.cmd.cpp +249 -0
  162. data/ext/sources/examples/whisper.wasm/CMakeLists.txt +50 -0
  163. data/ext/sources/examples/whisper.wasm/emscripten.cpp +118 -0
  164. data/ext/sources/examples/whisper.wasm/index-tmpl.html +658 -0
  165. data/ext/sources/ggml/CMakeLists.txt +390 -0
  166. data/ext/sources/ggml/cmake/BuildTypes.cmake +54 -0
  167. data/ext/sources/ggml/cmake/GitVars.cmake +22 -0
  168. data/ext/sources/ggml/cmake/common.cmake +26 -0
  169. data/ext/sources/ggml/cmake/ggml-config.cmake.in +152 -0
  170. data/ext/{ggml → sources/ggml}/include/ggml-alloc.h +1 -1
  171. data/ext/{ggml → sources/ggml}/include/ggml-backend.h +9 -7
  172. data/ext/{ggml → sources/ggml}/include/ggml-cpp.h +2 -1
  173. data/ext/{ggml → sources/ggml}/include/ggml-cpu.h +9 -1
  174. data/ext/{ggml → sources/ggml}/include/ggml-metal.h +1 -1
  175. data/ext/{ggml → sources/ggml}/include/ggml-opt.h +49 -28
  176. data/ext/{ggml → sources/ggml}/include/ggml-rpc.h +6 -1
  177. data/ext/{ggml → sources/ggml}/include/ggml-vulkan.h +0 -2
  178. data/ext/{ggml → sources/ggml}/include/ggml.h +182 -265
  179. data/ext/sources/ggml/include/gguf.h +202 -0
  180. data/ext/sources/ggml/src/CMakeLists.txt +346 -0
  181. data/ext/{ggml → sources/ggml}/src/ggml-alloc.c +34 -29
  182. data/ext/sources/ggml/src/ggml-amx/CMakeLists.txt +107 -0
  183. data/ext/{ggml → sources/ggml}/src/ggml-backend-impl.h +1 -2
  184. data/ext/{ggml → sources/ggml}/src/ggml-backend-reg.cpp +87 -53
  185. data/ext/{ggml → sources/ggml}/src/ggml-backend.cpp +26 -14
  186. data/ext/sources/ggml/src/ggml-blas/CMakeLists.txt +87 -0
  187. data/ext/sources/ggml/src/ggml-cann/CMakeLists.txt +74 -0
  188. data/ext/sources/ggml/src/ggml-cann/Doxyfile +2579 -0
  189. data/ext/{ggml → sources/ggml}/src/ggml-cann/acl_tensor.cpp +10 -4
  190. data/ext/{ggml → sources/ggml}/src/ggml-cann/acl_tensor.h +5 -5
  191. data/ext/{ggml → sources/ggml}/src/ggml-cann/aclnn_ops.cpp +1272 -1506
  192. data/ext/sources/ggml/src/ggml-cann/aclnn_ops.h +1125 -0
  193. data/ext/{ggml → sources/ggml}/src/ggml-cann/common.h +135 -1
  194. data/ext/{ggml → sources/ggml}/src/ggml-cann/ggml-cann.cpp +564 -146
  195. data/ext/sources/ggml/src/ggml-cann/kernels/CMakeLists.txt +30 -0
  196. data/ext/{ggml → sources/ggml}/src/ggml-cann/kernels/dup.cpp +3 -5
  197. data/ext/{ggml → sources/ggml}/src/ggml-common.h +12 -8
  198. data/ext/sources/ggml/src/ggml-cpu/CMakeLists.txt +504 -0
  199. data/ext/{ggml → sources/ggml}/src/ggml-cpu/amx/amx.cpp +2 -1
  200. data/ext/sources/ggml/src/ggml-cpu/binary-ops.cpp +158 -0
  201. data/ext/sources/ggml/src/ggml-cpu/binary-ops.h +16 -0
  202. data/ext/sources/ggml/src/ggml-cpu/cmake/FindSIMD.cmake +100 -0
  203. data/ext/sources/ggml/src/ggml-cpu/common.h +72 -0
  204. data/ext/{ggml → sources/ggml}/src/ggml-cpu/cpu-feats-x86.cpp +5 -1
  205. data/ext/sources/ggml/src/ggml-cpu/ggml-cpu-aarch64.cpp +6431 -0
  206. data/ext/{ggml → sources/ggml}/src/ggml-cpu/ggml-cpu-impl.h +163 -41
  207. data/ext/{ggml → sources/ggml}/src/ggml-cpu/ggml-cpu-quants.c +4029 -1117
  208. data/ext/sources/ggml/src/ggml-cpu/ggml-cpu.c +3510 -0
  209. data/ext/{ggml → sources/ggml}/src/ggml-cpu/ggml-cpu.cpp +67 -18
  210. data/ext/sources/ggml/src/ggml-cpu/kleidiai/kernels.cpp +337 -0
  211. data/ext/sources/ggml/src/ggml-cpu/kleidiai/kernels.h +95 -0
  212. data/ext/sources/ggml/src/ggml-cpu/kleidiai/kleidiai.cpp +482 -0
  213. data/ext/sources/ggml/src/ggml-cpu/kleidiai/kleidiai.h +17 -0
  214. data/ext/sources/ggml/src/ggml-cpu/llamafile/sgemm.cpp +3544 -0
  215. data/ext/sources/ggml/src/ggml-cpu/llamafile/sgemm.h +14 -0
  216. data/ext/sources/ggml/src/ggml-cpu/ops.cpp +8903 -0
  217. data/ext/sources/ggml/src/ggml-cpu/ops.h +110 -0
  218. data/ext/sources/ggml/src/ggml-cpu/simd-mappings.h +892 -0
  219. data/ext/sources/ggml/src/ggml-cpu/unary-ops.cpp +186 -0
  220. data/ext/sources/ggml/src/ggml-cpu/unary-ops.h +28 -0
  221. data/ext/sources/ggml/src/ggml-cpu/vec.cpp +252 -0
  222. data/ext/sources/ggml/src/ggml-cpu/vec.h +818 -0
  223. data/ext/sources/ggml/src/ggml-cuda/CMakeLists.txt +184 -0
  224. data/ext/sources/ggml/src/ggml-cuda/acc.cu +61 -0
  225. data/ext/sources/ggml/src/ggml-cuda/acc.cuh +5 -0
  226. data/ext/sources/ggml/src/ggml-cuda/arange.cu +34 -0
  227. data/ext/sources/ggml/src/ggml-cuda/arange.cuh +5 -0
  228. data/ext/sources/ggml/src/ggml-cuda/argmax.cu +91 -0
  229. data/ext/sources/ggml/src/ggml-cuda/argmax.cuh +3 -0
  230. data/ext/sources/ggml/src/ggml-cuda/argsort.cu +104 -0
  231. data/ext/sources/ggml/src/ggml-cuda/argsort.cuh +3 -0
  232. data/ext/sources/ggml/src/ggml-cuda/binbcast.cu +363 -0
  233. data/ext/sources/ggml/src/ggml-cuda/binbcast.cuh +9 -0
  234. data/ext/sources/ggml/src/ggml-cuda/clamp.cu +45 -0
  235. data/ext/sources/ggml/src/ggml-cuda/clamp.cuh +5 -0
  236. data/ext/sources/ggml/src/ggml-cuda/common.cuh +828 -0
  237. data/ext/sources/ggml/src/ggml-cuda/concat.cu +221 -0
  238. data/ext/sources/ggml/src/ggml-cuda/concat.cuh +5 -0
  239. data/ext/sources/ggml/src/ggml-cuda/conv-transpose-1d.cu +89 -0
  240. data/ext/sources/ggml/src/ggml-cuda/conv-transpose-1d.cuh +5 -0
  241. data/ext/sources/ggml/src/ggml-cuda/convert.cu +730 -0
  242. data/ext/sources/ggml/src/ggml-cuda/convert.cuh +26 -0
  243. data/ext/sources/ggml/src/ggml-cuda/count-equal.cu +64 -0
  244. data/ext/sources/ggml/src/ggml-cuda/count-equal.cuh +5 -0
  245. data/ext/sources/ggml/src/ggml-cuda/cp-async.cuh +57 -0
  246. data/ext/sources/ggml/src/ggml-cuda/cpy.cu +705 -0
  247. data/ext/sources/ggml/src/ggml-cuda/cpy.cuh +11 -0
  248. data/ext/sources/ggml/src/ggml-cuda/cross-entropy-loss.cu +189 -0
  249. data/ext/sources/ggml/src/ggml-cuda/cross-entropy-loss.cuh +7 -0
  250. data/ext/sources/ggml/src/ggml-cuda/dequantize.cuh +103 -0
  251. data/ext/sources/ggml/src/ggml-cuda/diagmask.cu +40 -0
  252. data/ext/sources/ggml/src/ggml-cuda/diagmask.cuh +5 -0
  253. data/ext/sources/ggml/src/ggml-cuda/fattn-common.cuh +881 -0
  254. data/ext/sources/ggml/src/ggml-cuda/fattn-mma-f16.cuh +1471 -0
  255. data/ext/sources/ggml/src/ggml-cuda/fattn-tile-f16.cu +357 -0
  256. data/ext/sources/ggml/src/ggml-cuda/fattn-tile-f16.cuh +3 -0
  257. data/ext/sources/ggml/src/ggml-cuda/fattn-tile-f32.cu +365 -0
  258. data/ext/sources/ggml/src/ggml-cuda/fattn-tile-f32.cuh +3 -0
  259. data/ext/sources/ggml/src/ggml-cuda/fattn-vec-f16.cuh +482 -0
  260. data/ext/sources/ggml/src/ggml-cuda/fattn-vec-f32.cuh +472 -0
  261. data/ext/sources/ggml/src/ggml-cuda/fattn-wmma-f16.cu +634 -0
  262. data/ext/sources/ggml/src/ggml-cuda/fattn-wmma-f16.cuh +3 -0
  263. data/ext/sources/ggml/src/ggml-cuda/fattn.cu +346 -0
  264. data/ext/sources/ggml/src/ggml-cuda/fattn.cuh +3 -0
  265. data/ext/sources/ggml/src/ggml-cuda/getrows.cu +275 -0
  266. data/ext/sources/ggml/src/ggml-cuda/getrows.cuh +15 -0
  267. data/ext/sources/ggml/src/ggml-cuda/ggml-cuda.cu +3505 -0
  268. data/ext/sources/ggml/src/ggml-cuda/gla.cu +93 -0
  269. data/ext/sources/ggml/src/ggml-cuda/gla.cuh +3 -0
  270. data/ext/sources/ggml/src/ggml-cuda/im2col.cu +103 -0
  271. data/ext/sources/ggml/src/ggml-cuda/im2col.cuh +5 -0
  272. data/ext/sources/ggml/src/ggml-cuda/mma.cuh +396 -0
  273. data/ext/sources/ggml/src/ggml-cuda/mmq.cu +324 -0
  274. data/ext/sources/ggml/src/ggml-cuda/mmq.cuh +3217 -0
  275. data/ext/sources/ggml/src/ggml-cuda/mmv.cu +336 -0
  276. data/ext/sources/ggml/src/ggml-cuda/mmv.cuh +12 -0
  277. data/ext/sources/ggml/src/ggml-cuda/mmvq.cu +595 -0
  278. data/ext/sources/ggml/src/ggml-cuda/mmvq.cuh +12 -0
  279. data/ext/sources/ggml/src/ggml-cuda/norm.cu +458 -0
  280. data/ext/sources/ggml/src/ggml-cuda/norm.cuh +11 -0
  281. data/ext/sources/ggml/src/ggml-cuda/opt-step-adamw.cu +78 -0
  282. data/ext/sources/ggml/src/ggml-cuda/opt-step-adamw.cuh +5 -0
  283. data/ext/sources/ggml/src/ggml-cuda/out-prod.cu +68 -0
  284. data/ext/sources/ggml/src/ggml-cuda/out-prod.cuh +3 -0
  285. data/ext/sources/ggml/src/ggml-cuda/pad.cu +49 -0
  286. data/ext/sources/ggml/src/ggml-cuda/pad.cuh +5 -0
  287. data/ext/sources/ggml/src/ggml-cuda/pool2d.cu +94 -0
  288. data/ext/sources/ggml/src/ggml-cuda/pool2d.cuh +5 -0
  289. data/ext/sources/ggml/src/ggml-cuda/quantize.cu +190 -0
  290. data/ext/sources/ggml/src/ggml-cuda/quantize.cuh +27 -0
  291. data/ext/sources/ggml/src/ggml-cuda/rope.cu +456 -0
  292. data/ext/sources/ggml/src/ggml-cuda/rope.cuh +7 -0
  293. data/ext/sources/ggml/src/ggml-cuda/scale.cu +31 -0
  294. data/ext/sources/ggml/src/ggml-cuda/scale.cuh +5 -0
  295. data/ext/sources/ggml/src/ggml-cuda/softmax.cu +283 -0
  296. data/ext/sources/ggml/src/ggml-cuda/softmax.cuh +7 -0
  297. data/ext/sources/ggml/src/ggml-cuda/ssm-conv.cu +148 -0
  298. data/ext/sources/ggml/src/ggml-cuda/ssm-conv.cuh +3 -0
  299. data/ext/sources/ggml/src/ggml-cuda/ssm-scan.cu +153 -0
  300. data/ext/sources/ggml/src/ggml-cuda/ssm-scan.cuh +3 -0
  301. data/ext/sources/ggml/src/ggml-cuda/sum.cu +45 -0
  302. data/ext/sources/ggml/src/ggml-cuda/sum.cuh +5 -0
  303. data/ext/sources/ggml/src/ggml-cuda/sumrows.cu +39 -0
  304. data/ext/sources/ggml/src/ggml-cuda/sumrows.cuh +5 -0
  305. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_1-ncols2_16.cu +5 -0
  306. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_1-ncols2_8.cu +10 -0
  307. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_1.cu +10 -0
  308. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_2.cu +10 -0
  309. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_4.cu +10 -0
  310. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_16.cu +5 -0
  311. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_4.cu +10 -0
  312. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_8.cu +10 -0
  313. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_32-ncols2_1.cu +10 -0
  314. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_32-ncols2_2.cu +10 -0
  315. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_16.cu +5 -0
  316. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_2.cu +10 -0
  317. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_4.cu +10 -0
  318. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_8.cu +10 -0
  319. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_64-ncols2_1.cu +10 -0
  320. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_1.cu +10 -0
  321. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_2.cu +10 -0
  322. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_4.cu +10 -0
  323. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_8.cu +10 -0
  324. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-f16.cu +5 -0
  325. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q4_0.cu +5 -0
  326. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q4_1.cu +5 -0
  327. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q5_0.cu +5 -0
  328. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q5_1.cu +5 -0
  329. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q8_0.cu +5 -0
  330. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-f16.cu +5 -0
  331. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q4_0.cu +5 -0
  332. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q4_1.cu +5 -0
  333. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q5_0.cu +5 -0
  334. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q5_1.cu +5 -0
  335. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q8_0.cu +5 -0
  336. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-f16.cu +5 -0
  337. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q4_0.cu +5 -0
  338. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q4_1.cu +5 -0
  339. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q5_0.cu +5 -0
  340. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q5_1.cu +5 -0
  341. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q8_0.cu +5 -0
  342. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-f16.cu +5 -0
  343. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q4_0.cu +5 -0
  344. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q4_1.cu +5 -0
  345. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q5_0.cu +5 -0
  346. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q5_1.cu +5 -0
  347. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q8_0.cu +5 -0
  348. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-f16.cu +5 -0
  349. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q4_0.cu +5 -0
  350. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q4_1.cu +5 -0
  351. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q5_0.cu +5 -0
  352. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q5_1.cu +5 -0
  353. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q8_0.cu +5 -0
  354. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-f16.cu +5 -0
  355. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q4_0.cu +5 -0
  356. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q4_1.cu +5 -0
  357. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q5_0.cu +5 -0
  358. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q5_1.cu +5 -0
  359. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q8_0.cu +5 -0
  360. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs256-f16-f16.cu +5 -0
  361. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-f16.cu +5 -0
  362. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q4_0.cu +5 -0
  363. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q4_1.cu +5 -0
  364. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q5_0.cu +5 -0
  365. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q5_1.cu +5 -0
  366. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q8_0.cu +5 -0
  367. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-f16.cu +5 -0
  368. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q4_0.cu +5 -0
  369. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q4_1.cu +5 -0
  370. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q5_0.cu +5 -0
  371. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q5_1.cu +5 -0
  372. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q8_0.cu +5 -0
  373. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-f16.cu +5 -0
  374. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q4_0.cu +5 -0
  375. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q4_1.cu +5 -0
  376. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q5_0.cu +5 -0
  377. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q5_1.cu +5 -0
  378. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q8_0.cu +5 -0
  379. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-f16.cu +5 -0
  380. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q4_0.cu +5 -0
  381. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q4_1.cu +5 -0
  382. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q5_0.cu +5 -0
  383. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q5_1.cu +5 -0
  384. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q8_0.cu +5 -0
  385. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-f16.cu +5 -0
  386. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q4_0.cu +5 -0
  387. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q4_1.cu +5 -0
  388. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q5_0.cu +5 -0
  389. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q5_1.cu +5 -0
  390. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q8_0.cu +5 -0
  391. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-f16.cu +5 -0
  392. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q4_0.cu +5 -0
  393. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q4_1.cu +5 -0
  394. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q5_0.cu +5 -0
  395. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q5_1.cu +5 -0
  396. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q8_0.cu +5 -0
  397. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-f16.cu +5 -0
  398. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q4_0.cu +5 -0
  399. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q4_1.cu +5 -0
  400. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q5_0.cu +5 -0
  401. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q5_1.cu +5 -0
  402. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q8_0.cu +5 -0
  403. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs256-f16-f16.cu +5 -0
  404. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-f16.cu +5 -0
  405. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q4_0.cu +5 -0
  406. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q4_1.cu +5 -0
  407. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q5_0.cu +5 -0
  408. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q5_1.cu +5 -0
  409. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q8_0.cu +5 -0
  410. data/ext/sources/ggml/src/ggml-cuda/template-instances/generate_cu_files.py +78 -0
  411. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq1_s.cu +5 -0
  412. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq2_s.cu +5 -0
  413. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq2_xs.cu +5 -0
  414. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq2_xxs.cu +5 -0
  415. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq3_s.cu +5 -0
  416. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq3_xxs.cu +5 -0
  417. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq4_nl.cu +5 -0
  418. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq4_xs.cu +5 -0
  419. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q2_k.cu +5 -0
  420. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q3_k.cu +5 -0
  421. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q4_0.cu +5 -0
  422. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q4_1.cu +5 -0
  423. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q4_k.cu +5 -0
  424. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q5_0.cu +5 -0
  425. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q5_1.cu +5 -0
  426. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q5_k.cu +5 -0
  427. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q6_k.cu +5 -0
  428. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q8_0.cu +5 -0
  429. data/ext/sources/ggml/src/ggml-cuda/tsembd.cu +47 -0
  430. data/ext/sources/ggml/src/ggml-cuda/tsembd.cuh +5 -0
  431. data/ext/sources/ggml/src/ggml-cuda/unary.cu +289 -0
  432. data/ext/sources/ggml/src/ggml-cuda/unary.cuh +59 -0
  433. data/ext/sources/ggml/src/ggml-cuda/upscale.cu +51 -0
  434. data/ext/sources/ggml/src/ggml-cuda/upscale.cuh +5 -0
  435. data/ext/sources/ggml/src/ggml-cuda/vecdotq.cuh +1135 -0
  436. data/ext/{ggml → sources/ggml}/src/ggml-cuda/vendors/cuda.h +1 -0
  437. data/ext/{ggml → sources/ggml}/src/ggml-cuda/vendors/hip.h +57 -0
  438. data/ext/{ggml → sources/ggml}/src/ggml-cuda/vendors/musa.h +7 -1
  439. data/ext/sources/ggml/src/ggml-cuda/wkv.cu +199 -0
  440. data/ext/sources/ggml/src/ggml-cuda/wkv.cuh +7 -0
  441. data/ext/sources/ggml/src/ggml-hip/CMakeLists.txt +131 -0
  442. data/ext/{ggml → sources/ggml}/src/ggml-impl.h +64 -19
  443. data/ext/sources/ggml/src/ggml-kompute/CMakeLists.txt +166 -0
  444. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/common.comp +112 -0
  445. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_add.comp +58 -0
  446. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_addrow.comp +25 -0
  447. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f16_f16.comp +52 -0
  448. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f16_f32.comp +52 -0
  449. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f32_f16.comp +52 -0
  450. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f32_f32.comp +52 -0
  451. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_diagmask.comp +30 -0
  452. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_gelu.comp +22 -0
  453. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows.comp +17 -0
  454. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows_f16.comp +31 -0
  455. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows_f32.comp +31 -0
  456. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q4_0.comp +38 -0
  457. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q4_1.comp +39 -0
  458. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q6_k.comp +44 -0
  459. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul.comp +52 -0
  460. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_f16.comp +69 -0
  461. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_mat_f32.comp +51 -0
  462. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_0.comp +33 -0
  463. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_1.comp +35 -0
  464. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_k.comp +140 -0
  465. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q6_k.comp +106 -0
  466. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q8_0.comp +73 -0
  467. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n.comp +52 -0
  468. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n_pre.comp +28 -0
  469. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_norm.comp +84 -0
  470. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_relu.comp +21 -0
  471. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_rmsnorm.comp +53 -0
  472. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f16.comp +52 -0
  473. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f32.comp +52 -0
  474. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f16.comp +52 -0
  475. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f32.comp +52 -0
  476. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_scale.comp +19 -0
  477. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_scale_8.comp +23 -0
  478. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_silu.comp +22 -0
  479. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_softmax.comp +72 -0
  480. data/ext/sources/ggml/src/ggml-kompute/kompute-shaders/rope_common.comp +71 -0
  481. data/ext/sources/ggml/src/ggml-metal/CMakeLists.txt +120 -0
  482. data/ext/sources/ggml/src/ggml-metal/ggml-metal-impl.h +622 -0
  483. data/ext/{ggml → sources/ggml}/src/ggml-metal/ggml-metal.m +2178 -1064
  484. data/ext/{ggml → sources/ggml}/src/ggml-metal/ggml-metal.metal +1575 -1218
  485. data/ext/sources/ggml/src/ggml-musa/CMakeLists.txt +113 -0
  486. data/ext/sources/ggml/src/ggml-musa/mudnn.cu +112 -0
  487. data/ext/sources/ggml/src/ggml-musa/mudnn.cuh +12 -0
  488. data/ext/sources/ggml/src/ggml-opencl/CMakeLists.txt +96 -0
  489. data/ext/sources/ggml/src/ggml-opencl/ggml-opencl.cpp +5124 -0
  490. data/ext/sources/ggml/src/ggml-opencl/kernels/add.cl +83 -0
  491. data/ext/sources/ggml/src/ggml-opencl/kernels/clamp.cl +20 -0
  492. data/ext/sources/ggml/src/ggml-opencl/kernels/cpy.cl +184 -0
  493. data/ext/sources/ggml/src/ggml-opencl/kernels/cvt.cl +118 -0
  494. data/ext/sources/ggml/src/ggml-opencl/kernels/diag_mask_inf.cl +58 -0
  495. data/ext/sources/ggml/src/ggml-opencl/kernels/embed_kernel.py +26 -0
  496. data/ext/sources/ggml/src/ggml-opencl/kernels/gelu.cl +62 -0
  497. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle.cl +268 -0
  498. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_general.cl +274 -0
  499. data/ext/sources/ggml/src/ggml-opencl/kernels/get_rows.cl +163 -0
  500. data/ext/sources/ggml/src/ggml-opencl/kernels/im2col_f16.cl +57 -0
  501. data/ext/sources/ggml/src/ggml-opencl/kernels/im2col_f32.cl +57 -0
  502. data/ext/sources/ggml/src/ggml-opencl/kernels/mul.cl +79 -0
  503. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mat_Ab_Bi_8x4.cl +139 -0
  504. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f16_f16.cl +118 -0
  505. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f16_f32.cl +118 -0
  506. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f16_f32_1row.cl +94 -0
  507. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f16_f32_l4.cl +84 -0
  508. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f32_f32.cl +118 -0
  509. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32.cl +192 -0
  510. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32_1d_16x_flat.cl +307 -0
  511. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32_1d_8x_flat.cl +265 -0
  512. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32_8x_flat.cl +272 -0
  513. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32_v.cl +254 -0
  514. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q6_k.cl +190 -0
  515. data/ext/sources/ggml/src/ggml-opencl/kernels/norm.cl +81 -0
  516. data/ext/sources/ggml/src/ggml-opencl/kernels/relu.cl +16 -0
  517. data/ext/sources/ggml/src/ggml-opencl/kernels/rms_norm.cl +96 -0
  518. data/ext/sources/ggml/src/ggml-opencl/kernels/rope.cl +721 -0
  519. data/ext/sources/ggml/src/ggml-opencl/kernels/scale.cl +16 -0
  520. data/ext/sources/ggml/src/ggml-opencl/kernels/silu.cl +30 -0
  521. data/ext/sources/ggml/src/ggml-opencl/kernels/softmax_4_f16.cl +87 -0
  522. data/ext/sources/ggml/src/ggml-opencl/kernels/softmax_4_f32.cl +87 -0
  523. data/ext/sources/ggml/src/ggml-opencl/kernels/softmax_f16.cl +86 -0
  524. data/ext/sources/ggml/src/ggml-opencl/kernels/softmax_f32.cl +86 -0
  525. data/ext/sources/ggml/src/ggml-opencl/kernels/transpose.cl +84 -0
  526. data/ext/{ggml → sources/ggml}/src/ggml-opt.cpp +373 -190
  527. data/ext/{ggml → sources/ggml}/src/ggml-quants.c +114 -120
  528. data/ext/sources/ggml/src/ggml-rpc/CMakeLists.txt +9 -0
  529. data/ext/{ggml → sources/ggml}/src/ggml-rpc/ggml-rpc.cpp +480 -73
  530. data/ext/sources/ggml/src/ggml-sycl/CMakeLists.txt +189 -0
  531. data/ext/sources/ggml/src/ggml-sycl/backend.hpp +37 -0
  532. data/ext/sources/ggml/src/ggml-sycl/binbcast.cpp +345 -0
  533. data/ext/sources/ggml/src/ggml-sycl/binbcast.hpp +39 -0
  534. data/ext/{ggml → sources/ggml}/src/ggml-sycl/common.cpp +20 -32
  535. data/ext/sources/ggml/src/ggml-sycl/common.hpp +589 -0
  536. data/ext/{ggml → sources/ggml}/src/ggml-sycl/concat.cpp +32 -33
  537. data/ext/sources/ggml/src/ggml-sycl/concat.hpp +20 -0
  538. data/ext/{ggml → sources/ggml}/src/ggml-sycl/conv.cpp +4 -2
  539. data/ext/sources/ggml/src/ggml-sycl/conv.hpp +20 -0
  540. data/ext/{ggml → sources/ggml}/src/ggml-sycl/convert.cpp +104 -28
  541. data/ext/sources/ggml/src/ggml-sycl/convert.hpp +34 -0
  542. data/ext/sources/ggml/src/ggml-sycl/cpy.cpp +700 -0
  543. data/ext/sources/ggml/src/ggml-sycl/cpy.hpp +11 -0
  544. data/ext/sources/ggml/src/ggml-sycl/dequantize.hpp +791 -0
  545. data/ext/{ggml → sources/ggml}/src/ggml-sycl/dmmv.cpp +156 -17
  546. data/ext/sources/ggml/src/ggml-sycl/dmmv.hpp +27 -0
  547. data/ext/sources/ggml/src/ggml-sycl/dpct/helper.hpp +2957 -0
  548. data/ext/sources/ggml/src/ggml-sycl/element_wise.cpp +1511 -0
  549. data/ext/sources/ggml/src/ggml-sycl/element_wise.hpp +75 -0
  550. data/ext/sources/ggml/src/ggml-sycl/gemm.hpp +99 -0
  551. data/ext/sources/ggml/src/ggml-sycl/getrows.cpp +309 -0
  552. data/ext/sources/ggml/src/ggml-sycl/getrows.hpp +20 -0
  553. data/ext/{ggml → sources/ggml}/src/ggml-sycl/ggml-sycl.cpp +1004 -1240
  554. data/ext/sources/ggml/src/ggml-sycl/gla.cpp +106 -0
  555. data/ext/sources/ggml/src/ggml-sycl/gla.hpp +8 -0
  556. data/ext/sources/ggml/src/ggml-sycl/im2col.cpp +136 -0
  557. data/ext/sources/ggml/src/ggml-sycl/im2col.hpp +21 -0
  558. data/ext/{ggml → sources/ggml}/src/ggml-sycl/mmq.cpp +0 -1
  559. data/ext/sources/ggml/src/ggml-sycl/mmq.hpp +33 -0
  560. data/ext/{ggml → sources/ggml}/src/ggml-sycl/mmvq.cpp +261 -166
  561. data/ext/sources/ggml/src/ggml-sycl/mmvq.hpp +27 -0
  562. data/ext/{ggml → sources/ggml}/src/ggml-sycl/norm.cpp +204 -81
  563. data/ext/sources/ggml/src/ggml-sycl/norm.hpp +26 -0
  564. data/ext/{ggml → sources/ggml}/src/ggml-sycl/outprod.cpp +8 -17
  565. data/ext/sources/ggml/src/ggml-sycl/outprod.hpp +10 -0
  566. data/ext/sources/ggml/src/ggml-sycl/presets.hpp +74 -0
  567. data/ext/sources/ggml/src/ggml-sycl/quants.hpp +83 -0
  568. data/ext/sources/ggml/src/ggml-sycl/rope.cpp +361 -0
  569. data/ext/sources/ggml/src/ggml-sycl/rope.hpp +20 -0
  570. data/ext/{ggml → sources/ggml}/src/ggml-sycl/softmax.cpp +35 -25
  571. data/ext/sources/ggml/src/ggml-sycl/softmax.hpp +20 -0
  572. data/ext/sources/ggml/src/ggml-sycl/sycl_hw.cpp +13 -0
  573. data/ext/sources/ggml/src/ggml-sycl/sycl_hw.hpp +23 -0
  574. data/ext/{ggml → sources/ggml}/src/ggml-sycl/tsembd.cpp +3 -3
  575. data/ext/sources/ggml/src/ggml-sycl/tsembd.hpp +20 -0
  576. data/ext/sources/ggml/src/ggml-sycl/vecdotq.hpp +1215 -0
  577. data/ext/sources/ggml/src/ggml-sycl/wkv.cpp +293 -0
  578. data/ext/sources/ggml/src/ggml-sycl/wkv.hpp +10 -0
  579. data/ext/sources/ggml/src/ggml-vulkan/CMakeLists.txt +196 -0
  580. data/ext/sources/ggml/src/ggml-vulkan/cmake/host-toolchain.cmake.in +15 -0
  581. data/ext/{ggml → sources/ggml}/src/ggml-vulkan/ggml-vulkan.cpp +3130 -1087
  582. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt +39 -0
  583. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/acc.comp +29 -0
  584. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/add.comp +29 -0
  585. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/argmax.comp +51 -0
  586. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/argsort.comp +69 -0
  587. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/clamp.comp +17 -0
  588. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/concat.comp +41 -0
  589. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/contig_copy.comp +49 -0
  590. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_dw.comp +105 -0
  591. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/copy.comp +23 -0
  592. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/copy_from_quant.comp +51 -0
  593. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/copy_to_quant.comp +242 -0
  594. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/cos.comp +17 -0
  595. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/count_equal.comp +31 -0
  596. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_f32.comp +20 -0
  597. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.comp +462 -0
  598. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.comp +699 -0
  599. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_head.comp +13 -0
  600. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_m.comp +42 -0
  601. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_s.comp +35 -0
  602. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_s.comp +44 -0
  603. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xs.comp +43 -0
  604. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xxs.comp +48 -0
  605. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_s.comp +39 -0
  606. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_xxs.comp +49 -0
  607. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp +32 -0
  608. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_xs.comp +34 -0
  609. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q2_k.comp +34 -0
  610. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q3_k.comp +42 -0
  611. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_0.comp +30 -0
  612. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_1.comp +32 -0
  613. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_k.comp +68 -0
  614. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_0.comp +34 -0
  615. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_1.comp +35 -0
  616. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_k.comp +70 -0
  617. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q6_k.comp +33 -0
  618. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q8_0.comp +31 -0
  619. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/diag_mask_inf.comp +34 -0
  620. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/div.comp +27 -0
  621. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp +337 -0
  622. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_base.comp +162 -0
  623. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp +360 -0
  624. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp +267 -0
  625. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_split_k_reduce.comp +59 -0
  626. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/gelu.comp +25 -0
  627. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/gelu_quick.comp +23 -0
  628. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.comp +64 -0
  629. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/generic_head.comp +9 -0
  630. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/generic_unary_head.comp +76 -0
  631. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/get_rows.comp +33 -0
  632. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/get_rows_quant.comp +41 -0
  633. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/group_norm.comp +66 -0
  634. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/im2col.comp +100 -0
  635. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/l2_norm.comp +41 -0
  636. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/leaky_relu.comp +22 -0
  637. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul.comp +27 -0
  638. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_split_k_reduce.comp +48 -0
  639. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp +169 -0
  640. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.comp +118 -0
  641. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_m.comp +82 -0
  642. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_s.comp +79 -0
  643. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_s.comp +90 -0
  644. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xs.comp +87 -0
  645. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xxs.comp +87 -0
  646. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_s.comp +90 -0
  647. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_xxs.comp +88 -0
  648. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_nc.comp +118 -0
  649. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_p021.comp +154 -0
  650. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp +130 -0
  651. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp +132 -0
  652. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp +136 -0
  653. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp +167 -0
  654. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp +130 -0
  655. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp +868 -0
  656. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp +441 -0
  657. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq.comp +442 -0
  658. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.comp +99 -0
  659. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/norm.comp +44 -0
  660. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_adamw.comp +42 -0
  661. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/pad.comp +28 -0
  662. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/pool2d.comp +74 -0
  663. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/quantize_q8_1.comp +77 -0
  664. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/relu.comp +21 -0
  665. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/repeat.comp +26 -0
  666. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/repeat_back.comp +37 -0
  667. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp +52 -0
  668. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_back.comp +55 -0
  669. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.comp +58 -0
  670. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_multi.comp +60 -0
  671. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_neox.comp +43 -0
  672. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_norm.comp +43 -0
  673. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_vision.comp +47 -0
  674. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp +24 -0
  675. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sigmoid.comp +20 -0
  676. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/silu.comp +22 -0
  677. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/silu_back.comp +26 -0
  678. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp +17 -0
  679. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp +173 -0
  680. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/soft_max_back.comp +50 -0
  681. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/square.comp +17 -0
  682. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sub.comp +29 -0
  683. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sum_rows.comp +37 -0
  684. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/tanh.comp +20 -0
  685. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/test_bfloat16_support.comp +7 -0
  686. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat2_support.comp +7 -0
  687. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat_support.comp +7 -0
  688. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/test_integer_dot_support.comp +7 -0
  689. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/timestep_embedding.comp +41 -0
  690. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/types.comp +1373 -0
  691. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/upscale.comp +36 -0
  692. data/ext/{ggml → sources/ggml}/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp +193 -35
  693. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/wkv6.comp +87 -0
  694. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/wkv7.comp +91 -0
  695. data/ext/{ggml → sources/ggml}/src/ggml.c +676 -1820
  696. data/ext/sources/ggml/src/gguf.cpp +1330 -0
  697. data/ext/{include → sources/include}/whisper.h +68 -2
  698. data/ext/sources/src/CMakeLists.txt +143 -0
  699. data/ext/{src → sources/src}/coreml/whisper-decoder-impl.h +27 -15
  700. data/ext/{src → sources/src}/coreml/whisper-decoder-impl.m +35 -10
  701. data/ext/{src → sources/src}/coreml/whisper-encoder-impl.h +21 -9
  702. data/ext/{src → sources/src}/coreml/whisper-encoder-impl.m +28 -3
  703. data/ext/sources/src/coreml/whisper-encoder.mm +73 -0
  704. data/ext/sources/src/whisper-arch.h +197 -0
  705. data/ext/{src → sources/src}/whisper.cpp +1905 -374
  706. data/ext/sources/tests/CMakeLists.txt +105 -0
  707. data/ext/sources/tests/earnings21/eval.mk +58 -0
  708. data/ext/sources/tests/earnings21/eval.py +68 -0
  709. data/ext/sources/tests/earnings21/normalizers/__init__.py +2 -0
  710. data/ext/sources/tests/earnings21/normalizers/basic.py +80 -0
  711. data/ext/sources/tests/earnings21/normalizers/english.json +1741 -0
  712. data/ext/sources/tests/earnings21/normalizers/english.py +550 -0
  713. data/ext/sources/tests/earnings21/requirements.txt +6 -0
  714. data/ext/sources/tests/en-0-ref.txt +1 -0
  715. data/ext/sources/tests/en-1-ref.txt +1 -0
  716. data/ext/sources/tests/en-2-ref.txt +1 -0
  717. data/ext/sources/tests/es-0-ref.txt +1 -0
  718. data/ext/sources/tests/librispeech/eval.mk +39 -0
  719. data/ext/sources/tests/librispeech/eval.py +47 -0
  720. data/ext/sources/tests/librispeech/normalizers/__init__.py +2 -0
  721. data/ext/sources/tests/librispeech/normalizers/basic.py +80 -0
  722. data/ext/sources/tests/librispeech/normalizers/english.json +1741 -0
  723. data/ext/sources/tests/librispeech/normalizers/english.py +550 -0
  724. data/ext/sources/tests/librispeech/requirements.txt +6 -0
  725. data/ext/sources/tests/run-tests.sh +130 -0
  726. data/ext/sources/tests/test-c.c +3 -0
  727. data/ext/sources/tests/test-vad-full.cpp +54 -0
  728. data/ext/sources/tests/test-vad.cpp +83 -0
  729. data/ext/sources/tests/test-whisper.js +58 -0
  730. data/extsources.rb +33 -5
  731. data/lib/whisper/model/uri.rb +149 -128
  732. data/sig/whisper.rbs +480 -0
  733. data/tests/helper.rb +28 -0
  734. data/tests/test_callback.rb +45 -3
  735. data/tests/test_error.rb +2 -2
  736. data/tests/test_model.rb +38 -0
  737. data/tests/test_package.rb +18 -3
  738. data/tests/test_params.rb +145 -8
  739. data/tests/test_segment.rb +10 -19
  740. data/tests/test_vad.rb +19 -0
  741. data/tests/test_vad_params.rb +103 -0
  742. data/tests/test_whisper.rb +37 -37
  743. data/whispercpp.gemspec +5 -4
  744. metadata +766 -111
  745. data/ext/cpu.mk +0 -9
  746. data/ext/examples/dr_wav.h +0 -8815
  747. data/ext/ggml/src/ggml-cann/aclnn_ops.h +0 -592
  748. data/ext/ggml/src/ggml-cpu/ggml-cpu-aarch64.cpp +0 -4262
  749. data/ext/ggml/src/ggml-cpu/ggml-cpu.c +0 -14123
  750. data/ext/ggml/src/ggml-cpu/llamafile/sgemm.cpp +0 -1884
  751. data/ext/ggml/src/ggml-cpu/llamafile/sgemm.h +0 -14
  752. data/ext/ggml/src/ggml-metal/ggml-metal-impl.h +0 -288
  753. data/ext/ggml/src/ggml-sycl/element_wise.cpp +0 -1030
  754. data/ext/ggml/src/ggml-sycl/im2col.cpp +0 -126
  755. data/ext/ggml/src/ggml-sycl/rope.cpp +0 -276
  756. data/ext/ggml/src/ggml-sycl/wkv6.cpp +0 -141
  757. data/ext/metal-embed.mk +0 -17
  758. data/ext/metal.mk +0 -6
  759. data/ext/ruby_whisper.cpp +0 -1909
  760. data/ext/scripts/get-flags.mk +0 -38
  761. data/lib/whisper.rb +0 -2
  762. /data/ext/{ggml → sources/ggml}/include/ggml-blas.h +0 -0
  763. /data/ext/{ggml → sources/ggml}/include/ggml-cann.h +0 -0
  764. /data/ext/{ggml → sources/ggml}/include/ggml-cuda.h +0 -0
  765. /data/ext/{ggml → sources/ggml}/include/ggml-kompute.h +0 -0
  766. /data/ext/{ggml → sources/ggml}/include/ggml-opencl.h +0 -0
  767. /data/ext/{ggml → sources/ggml}/include/ggml-sycl.h +0 -0
  768. /data/ext/{ggml → sources/ggml}/src/ggml-amx/common.h +0 -0
  769. /data/ext/{ggml → sources/ggml}/src/ggml-amx/ggml-amx.cpp +0 -0
  770. /data/ext/{ggml → sources/ggml}/src/ggml-amx/mmq.cpp +0 -0
  771. /data/ext/{ggml → sources/ggml}/src/ggml-amx/mmq.h +0 -0
  772. /data/ext/{ggml → sources/ggml}/src/ggml-blas/ggml-blas.cpp +0 -0
  773. /data/ext/{ggml → sources/ggml}/src/ggml-cann/kernels/ascendc_kernels.h +0 -0
  774. /data/ext/{ggml → sources/ggml}/src/ggml-cann/kernels/get_row_f16.cpp +0 -0
  775. /data/ext/{ggml → sources/ggml}/src/ggml-cann/kernels/get_row_f32.cpp +0 -0
  776. /data/ext/{ggml → sources/ggml}/src/ggml-cann/kernels/get_row_q4_0.cpp +0 -0
  777. /data/ext/{ggml → sources/ggml}/src/ggml-cann/kernels/get_row_q8_0.cpp +0 -0
  778. /data/ext/{ggml → sources/ggml}/src/ggml-cann/kernels/quantize_f16_q8_0.cpp +0 -0
  779. /data/ext/{ggml → sources/ggml}/src/ggml-cann/kernels/quantize_f32_q8_0.cpp +0 -0
  780. /data/ext/{ggml → sources/ggml}/src/ggml-cann/kernels/quantize_float_to_q4_0.cpp +0 -0
  781. /data/ext/{ggml → sources/ggml}/src/ggml-cpu/amx/amx.h +0 -0
  782. /data/ext/{ggml → sources/ggml}/src/ggml-cpu/amx/common.h +0 -0
  783. /data/ext/{ggml → sources/ggml}/src/ggml-cpu/amx/mmq.cpp +0 -0
  784. /data/ext/{ggml → sources/ggml}/src/ggml-cpu/amx/mmq.h +0 -0
  785. /data/ext/{ggml → sources/ggml}/src/ggml-cpu/ggml-cpu-aarch64.h +0 -0
  786. /data/ext/{ggml → sources/ggml}/src/ggml-cpu/ggml-cpu-hbm.cpp +0 -0
  787. /data/ext/{ggml → sources/ggml}/src/ggml-cpu/ggml-cpu-hbm.h +0 -0
  788. /data/ext/{ggml → sources/ggml}/src/ggml-cpu/ggml-cpu-quants.h +0 -0
  789. /data/ext/{ggml → sources/ggml}/src/ggml-cpu/ggml-cpu-traits.cpp +0 -0
  790. /data/ext/{ggml → sources/ggml}/src/ggml-cpu/ggml-cpu-traits.h +0 -0
  791. /data/ext/{ggml → sources/ggml}/src/ggml-kompute/ggml-kompute.cpp +0 -0
  792. /data/ext/{ggml → sources/ggml}/src/ggml-quants.h +0 -0
  793. /data/ext/{ggml → sources/ggml}/src/ggml-threading.cpp +0 -0
  794. /data/ext/{ggml → sources/ggml}/src/ggml-threading.h +0 -0
  795. /data/ext/{src → sources/src}/coreml/whisper-encoder.h +0 -0
  796. /data/ext/{src → sources/src}/openvino/whisper-openvino-encoder.cpp +0 -0
  797. /data/ext/{src → sources/src}/openvino/whisper-openvino-encoder.h +0 -0
metadata CHANGED
@@ -1,7 +1,7 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: whispercpp
3
3
  version: !ruby/object:Gem::Version
4
- version: 1.3.1
4
+ version: 1.3.2
5
5
  platform: ruby
6
6
  authors:
7
7
  - Georgi Gerganov
@@ -9,7 +9,7 @@ authors:
9
9
  autorequire:
10
10
  bindir: bin
11
11
  cert_chain: []
12
- date: 2024-12-19 00:00:00.000000000 Z
12
+ date: 2025-05-11 00:00:00.000000000 Z
13
13
  dependencies: []
14
14
  description: High-performance inference of OpenAI's Whisper automatic speech recognition
15
15
  (ASR) model via Ruby
@@ -26,118 +26,769 @@ files:
26
26
  - README.md
27
27
  - Rakefile
28
28
  - ext/.gitignore
29
- - ext/cpu.mk
30
- - ext/examples/dr_wav.h
29
+ - ext/dependencies.rb
31
30
  - ext/extconf.rb
32
- - ext/ggml/include/ggml-alloc.h
33
- - ext/ggml/include/ggml-backend.h
34
- - ext/ggml/include/ggml-blas.h
35
- - ext/ggml/include/ggml-cann.h
36
- - ext/ggml/include/ggml-cpp.h
37
- - ext/ggml/include/ggml-cpu.h
38
- - ext/ggml/include/ggml-cuda.h
39
- - ext/ggml/include/ggml-kompute.h
40
- - ext/ggml/include/ggml-metal.h
41
- - ext/ggml/include/ggml-opencl.h
42
- - ext/ggml/include/ggml-opt.h
43
- - ext/ggml/include/ggml-rpc.h
44
- - ext/ggml/include/ggml-sycl.h
45
- - ext/ggml/include/ggml-vulkan.h
46
- - ext/ggml/include/ggml.h
47
- - ext/ggml/src/ggml-alloc.c
48
- - ext/ggml/src/ggml-amx/common.h
49
- - ext/ggml/src/ggml-amx/ggml-amx.cpp
50
- - ext/ggml/src/ggml-amx/mmq.cpp
51
- - ext/ggml/src/ggml-amx/mmq.h
52
- - ext/ggml/src/ggml-backend-impl.h
53
- - ext/ggml/src/ggml-backend-reg.cpp
54
- - ext/ggml/src/ggml-backend.cpp
55
- - ext/ggml/src/ggml-blas/ggml-blas.cpp
56
- - ext/ggml/src/ggml-cann/acl_tensor.cpp
57
- - ext/ggml/src/ggml-cann/acl_tensor.h
58
- - ext/ggml/src/ggml-cann/aclnn_ops.cpp
59
- - ext/ggml/src/ggml-cann/aclnn_ops.h
60
- - ext/ggml/src/ggml-cann/common.h
61
- - ext/ggml/src/ggml-cann/ggml-cann.cpp
62
- - ext/ggml/src/ggml-cann/kernels/ascendc_kernels.h
63
- - ext/ggml/src/ggml-cann/kernels/dup.cpp
64
- - ext/ggml/src/ggml-cann/kernels/get_row_f16.cpp
65
- - ext/ggml/src/ggml-cann/kernels/get_row_f32.cpp
66
- - ext/ggml/src/ggml-cann/kernels/get_row_q4_0.cpp
67
- - ext/ggml/src/ggml-cann/kernels/get_row_q8_0.cpp
68
- - ext/ggml/src/ggml-cann/kernels/quantize_f16_q8_0.cpp
69
- - ext/ggml/src/ggml-cann/kernels/quantize_f32_q8_0.cpp
70
- - ext/ggml/src/ggml-cann/kernels/quantize_float_to_q4_0.cpp
71
- - ext/ggml/src/ggml-common.h
72
- - ext/ggml/src/ggml-cpu/amx/amx.cpp
73
- - ext/ggml/src/ggml-cpu/amx/amx.h
74
- - ext/ggml/src/ggml-cpu/amx/common.h
75
- - ext/ggml/src/ggml-cpu/amx/mmq.cpp
76
- - ext/ggml/src/ggml-cpu/amx/mmq.h
77
- - ext/ggml/src/ggml-cpu/cpu-feats-x86.cpp
78
- - ext/ggml/src/ggml-cpu/ggml-cpu-aarch64.cpp
79
- - ext/ggml/src/ggml-cpu/ggml-cpu-aarch64.h
80
- - ext/ggml/src/ggml-cpu/ggml-cpu-hbm.cpp
81
- - ext/ggml/src/ggml-cpu/ggml-cpu-hbm.h
82
- - ext/ggml/src/ggml-cpu/ggml-cpu-impl.h
83
- - ext/ggml/src/ggml-cpu/ggml-cpu-quants.c
84
- - ext/ggml/src/ggml-cpu/ggml-cpu-quants.h
85
- - ext/ggml/src/ggml-cpu/ggml-cpu-traits.cpp
86
- - ext/ggml/src/ggml-cpu/ggml-cpu-traits.h
87
- - ext/ggml/src/ggml-cpu/ggml-cpu.c
88
- - ext/ggml/src/ggml-cpu/ggml-cpu.cpp
89
- - ext/ggml/src/ggml-cpu/llamafile/sgemm.cpp
90
- - ext/ggml/src/ggml-cpu/llamafile/sgemm.h
91
- - ext/ggml/src/ggml-cuda/vendors/cuda.h
92
- - ext/ggml/src/ggml-cuda/vendors/hip.h
93
- - ext/ggml/src/ggml-cuda/vendors/musa.h
94
- - ext/ggml/src/ggml-impl.h
95
- - ext/ggml/src/ggml-kompute/ggml-kompute.cpp
96
- - ext/ggml/src/ggml-metal/ggml-metal-impl.h
97
- - ext/ggml/src/ggml-metal/ggml-metal.m
98
- - ext/ggml/src/ggml-metal/ggml-metal.metal
99
- - ext/ggml/src/ggml-opt.cpp
100
- - ext/ggml/src/ggml-quants.c
101
- - ext/ggml/src/ggml-quants.h
102
- - ext/ggml/src/ggml-rpc/ggml-rpc.cpp
103
- - ext/ggml/src/ggml-sycl/common.cpp
104
- - ext/ggml/src/ggml-sycl/concat.cpp
105
- - ext/ggml/src/ggml-sycl/conv.cpp
106
- - ext/ggml/src/ggml-sycl/convert.cpp
107
- - ext/ggml/src/ggml-sycl/dmmv.cpp
108
- - ext/ggml/src/ggml-sycl/element_wise.cpp
109
- - ext/ggml/src/ggml-sycl/ggml-sycl.cpp
110
- - ext/ggml/src/ggml-sycl/im2col.cpp
111
- - ext/ggml/src/ggml-sycl/mmq.cpp
112
- - ext/ggml/src/ggml-sycl/mmvq.cpp
113
- - ext/ggml/src/ggml-sycl/norm.cpp
114
- - ext/ggml/src/ggml-sycl/outprod.cpp
115
- - ext/ggml/src/ggml-sycl/rope.cpp
116
- - ext/ggml/src/ggml-sycl/softmax.cpp
117
- - ext/ggml/src/ggml-sycl/tsembd.cpp
118
- - ext/ggml/src/ggml-sycl/wkv6.cpp
119
- - ext/ggml/src/ggml-threading.cpp
120
- - ext/ggml/src/ggml-threading.h
121
- - ext/ggml/src/ggml-vulkan/ggml-vulkan.cpp
122
- - ext/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp
123
- - ext/ggml/src/ggml.c
124
- - ext/include/whisper.h
125
- - ext/metal-embed.mk
126
- - ext/metal.mk
127
- - ext/ruby_whisper.cpp
31
+ - ext/options.rb
32
+ - ext/ruby_whisper.c
128
33
  - ext/ruby_whisper.h
129
- - ext/scripts/get-flags.mk
130
- - ext/src/coreml/whisper-decoder-impl.h
131
- - ext/src/coreml/whisper-decoder-impl.m
132
- - ext/src/coreml/whisper-encoder-impl.h
133
- - ext/src/coreml/whisper-encoder-impl.m
134
- - ext/src/coreml/whisper-encoder.h
135
- - ext/src/openvino/whisper-openvino-encoder.cpp
136
- - ext/src/openvino/whisper-openvino-encoder.h
137
- - ext/src/whisper.cpp
34
+ - ext/ruby_whisper_context.c
35
+ - ext/ruby_whisper_error.c
36
+ - ext/ruby_whisper_model.c
37
+ - ext/ruby_whisper_params.c
38
+ - ext/ruby_whisper_segment.c
39
+ - ext/ruby_whisper_transcribe.cpp
40
+ - ext/ruby_whisper_vad_params.c
41
+ - ext/sources/.dockerignore
42
+ - ext/sources/.github/workflows/bindings-ruby.yml
43
+ - ext/sources/CMakeGraphVizOptions.cmake
44
+ - ext/sources/CMakeLists.txt
45
+ - ext/sources/bindings/javascript/CMakeLists.txt
46
+ - ext/sources/bindings/javascript/emscripten.cpp
47
+ - ext/sources/bindings/javascript/libwhisper.worker.js
48
+ - ext/sources/bindings/javascript/package-tmpl.json
49
+ - ext/sources/bindings/javascript/package.json
50
+ - ext/sources/bindings/javascript/whisper.js
51
+ - ext/sources/build-xcframework.sh
52
+ - ext/sources/ci/run.sh
53
+ - ext/sources/close-issue.yml
54
+ - ext/sources/cmake/DefaultTargetOptions.cmake
55
+ - ext/sources/cmake/FindFFmpeg.cmake
56
+ - ext/sources/cmake/build-info.cmake
57
+ - ext/sources/cmake/git-vars.cmake
58
+ - ext/sources/cmake/whisper-config.cmake.in
59
+ - ext/sources/cmake/whisper.pc.in
60
+ - ext/sources/examples/CMakeLists.txt
61
+ - ext/sources/examples/addon.node/CMakeLists.txt
62
+ - ext/sources/examples/addon.node/__test__/whisper.spec.js
63
+ - ext/sources/examples/addon.node/addon.cpp
64
+ - ext/sources/examples/addon.node/index.js
65
+ - ext/sources/examples/addon.node/package.json
66
+ - ext/sources/examples/bench.wasm/CMakeLists.txt
67
+ - ext/sources/examples/bench.wasm/emscripten.cpp
68
+ - ext/sources/examples/bench.wasm/index-tmpl.html
69
+ - ext/sources/examples/bench/CMakeLists.txt
70
+ - ext/sources/examples/bench/bench.cpp
71
+ - ext/sources/examples/cli/CMakeLists.txt
72
+ - ext/sources/examples/cli/cli.cpp
73
+ - ext/sources/examples/coi-serviceworker.js
74
+ - ext/sources/examples/command.wasm/CMakeLists.txt
75
+ - ext/sources/examples/command.wasm/emscripten.cpp
76
+ - ext/sources/examples/command.wasm/index-tmpl.html
77
+ - ext/sources/examples/command/CMakeLists.txt
78
+ - ext/sources/examples/command/command.cpp
79
+ - ext/sources/examples/command/commands.txt
80
+ - ext/sources/examples/common-ggml.cpp
81
+ - ext/sources/examples/common-ggml.h
82
+ - ext/sources/examples/common-sdl.cpp
83
+ - ext/sources/examples/common-sdl.h
84
+ - ext/sources/examples/common-whisper.cpp
85
+ - ext/sources/examples/common-whisper.h
86
+ - ext/sources/examples/common.cpp
87
+ - ext/sources/examples/common.h
88
+ - ext/sources/examples/deprecation-warning/CMakeLists.txt
89
+ - ext/sources/examples/deprecation-warning/deprecation-warning.cpp
90
+ - ext/sources/examples/ffmpeg-transcode.cpp
91
+ - ext/sources/examples/generate-karaoke.sh
92
+ - ext/sources/examples/grammar-parser.cpp
93
+ - ext/sources/examples/grammar-parser.h
94
+ - ext/sources/examples/helpers.js
95
+ - ext/sources/examples/json.hpp
96
+ - ext/sources/examples/livestream.sh
97
+ - ext/sources/examples/lsp/CMakeLists.txt
98
+ - ext/sources/examples/lsp/lsp.cpp
99
+ - ext/sources/examples/lsp/whisper.vim
100
+ - ext/sources/examples/miniaudio.h
101
+ - ext/sources/examples/python/test_whisper_processor.py
102
+ - ext/sources/examples/python/whisper_processor.py
103
+ - ext/sources/examples/quantize/CMakeLists.txt
104
+ - ext/sources/examples/quantize/quantize.cpp
105
+ - ext/sources/examples/server.py
106
+ - ext/sources/examples/server/CMakeLists.txt
107
+ - ext/sources/examples/server/bench.js
108
+ - ext/sources/examples/server/httplib.h
109
+ - ext/sources/examples/server/server.cpp
110
+ - ext/sources/examples/stb_vorbis.c
111
+ - ext/sources/examples/stream.wasm/CMakeLists.txt
112
+ - ext/sources/examples/stream.wasm/emscripten.cpp
113
+ - ext/sources/examples/stream.wasm/index-tmpl.html
114
+ - ext/sources/examples/stream/CMakeLists.txt
115
+ - ext/sources/examples/stream/stream.cpp
116
+ - ext/sources/examples/sycl/CMakeLists.txt
117
+ - ext/sources/examples/sycl/build.sh
118
+ - ext/sources/examples/sycl/ls-sycl-device.cpp
119
+ - ext/sources/examples/sycl/run-whisper.sh
120
+ - ext/sources/examples/talk-llama/CMakeLists.txt
121
+ - ext/sources/examples/talk-llama/eleven-labs.py
122
+ - ext/sources/examples/talk-llama/llama-adapter.cpp
123
+ - ext/sources/examples/talk-llama/llama-adapter.h
124
+ - ext/sources/examples/talk-llama/llama-arch.cpp
125
+ - ext/sources/examples/talk-llama/llama-arch.h
126
+ - ext/sources/examples/talk-llama/llama-batch.cpp
127
+ - ext/sources/examples/talk-llama/llama-batch.h
128
+ - ext/sources/examples/talk-llama/llama-chat.cpp
129
+ - ext/sources/examples/talk-llama/llama-chat.h
130
+ - ext/sources/examples/talk-llama/llama-context.cpp
131
+ - ext/sources/examples/talk-llama/llama-context.h
132
+ - ext/sources/examples/talk-llama/llama-cparams.cpp
133
+ - ext/sources/examples/talk-llama/llama-cparams.h
134
+ - ext/sources/examples/talk-llama/llama-grammar.cpp
135
+ - ext/sources/examples/talk-llama/llama-grammar.h
136
+ - ext/sources/examples/talk-llama/llama-graph.cpp
137
+ - ext/sources/examples/talk-llama/llama-graph.h
138
+ - ext/sources/examples/talk-llama/llama-hparams.cpp
139
+ - ext/sources/examples/talk-llama/llama-hparams.h
140
+ - ext/sources/examples/talk-llama/llama-impl.cpp
141
+ - ext/sources/examples/talk-llama/llama-impl.h
142
+ - ext/sources/examples/talk-llama/llama-io.cpp
143
+ - ext/sources/examples/talk-llama/llama-io.h
144
+ - ext/sources/examples/talk-llama/llama-kv-cache.cpp
145
+ - ext/sources/examples/talk-llama/llama-kv-cache.h
146
+ - ext/sources/examples/talk-llama/llama-kv-cells.h
147
+ - ext/sources/examples/talk-llama/llama-memory.cpp
148
+ - ext/sources/examples/talk-llama/llama-memory.h
149
+ - ext/sources/examples/talk-llama/llama-mmap.cpp
150
+ - ext/sources/examples/talk-llama/llama-mmap.h
151
+ - ext/sources/examples/talk-llama/llama-model-loader.cpp
152
+ - ext/sources/examples/talk-llama/llama-model-loader.h
153
+ - ext/sources/examples/talk-llama/llama-model-saver.cpp
154
+ - ext/sources/examples/talk-llama/llama-model-saver.h
155
+ - ext/sources/examples/talk-llama/llama-model.cpp
156
+ - ext/sources/examples/talk-llama/llama-model.h
157
+ - ext/sources/examples/talk-llama/llama-quant.cpp
158
+ - ext/sources/examples/talk-llama/llama-quant.h
159
+ - ext/sources/examples/talk-llama/llama-sampling.cpp
160
+ - ext/sources/examples/talk-llama/llama-sampling.h
161
+ - ext/sources/examples/talk-llama/llama-vocab.cpp
162
+ - ext/sources/examples/talk-llama/llama-vocab.h
163
+ - ext/sources/examples/talk-llama/llama.cpp
164
+ - ext/sources/examples/talk-llama/llama.h
165
+ - ext/sources/examples/talk-llama/prompts/talk-alpaca.txt
166
+ - ext/sources/examples/talk-llama/speak
167
+ - ext/sources/examples/talk-llama/speak.bat
168
+ - ext/sources/examples/talk-llama/speak.ps1
169
+ - ext/sources/examples/talk-llama/talk-llama.cpp
170
+ - ext/sources/examples/talk-llama/unicode-data.cpp
171
+ - ext/sources/examples/talk-llama/unicode-data.h
172
+ - ext/sources/examples/talk-llama/unicode.cpp
173
+ - ext/sources/examples/talk-llama/unicode.h
174
+ - ext/sources/examples/vad-speech-segments/CMakeLists.txt
175
+ - ext/sources/examples/vad-speech-segments/speech.cpp
176
+ - ext/sources/examples/wchess/CMakeLists.txt
177
+ - ext/sources/examples/wchess/libwchess/CMakeLists.txt
178
+ - ext/sources/examples/wchess/libwchess/Chessboard.cpp
179
+ - ext/sources/examples/wchess/libwchess/Chessboard.h
180
+ - ext/sources/examples/wchess/libwchess/WChess.cpp
181
+ - ext/sources/examples/wchess/libwchess/WChess.h
182
+ - ext/sources/examples/wchess/libwchess/test-chessboard.cpp
183
+ - ext/sources/examples/wchess/wchess.cmd/CMakeLists.txt
184
+ - ext/sources/examples/wchess/wchess.cmd/wchess.cmd.cpp
185
+ - ext/sources/examples/whisper.wasm/CMakeLists.txt
186
+ - ext/sources/examples/whisper.wasm/emscripten.cpp
187
+ - ext/sources/examples/whisper.wasm/index-tmpl.html
188
+ - ext/sources/ggml/CMakeLists.txt
189
+ - ext/sources/ggml/cmake/BuildTypes.cmake
190
+ - ext/sources/ggml/cmake/GitVars.cmake
191
+ - ext/sources/ggml/cmake/common.cmake
192
+ - ext/sources/ggml/cmake/ggml-config.cmake.in
193
+ - ext/sources/ggml/include/ggml-alloc.h
194
+ - ext/sources/ggml/include/ggml-backend.h
195
+ - ext/sources/ggml/include/ggml-blas.h
196
+ - ext/sources/ggml/include/ggml-cann.h
197
+ - ext/sources/ggml/include/ggml-cpp.h
198
+ - ext/sources/ggml/include/ggml-cpu.h
199
+ - ext/sources/ggml/include/ggml-cuda.h
200
+ - ext/sources/ggml/include/ggml-kompute.h
201
+ - ext/sources/ggml/include/ggml-metal.h
202
+ - ext/sources/ggml/include/ggml-opencl.h
203
+ - ext/sources/ggml/include/ggml-opt.h
204
+ - ext/sources/ggml/include/ggml-rpc.h
205
+ - ext/sources/ggml/include/ggml-sycl.h
206
+ - ext/sources/ggml/include/ggml-vulkan.h
207
+ - ext/sources/ggml/include/ggml.h
208
+ - ext/sources/ggml/include/gguf.h
209
+ - ext/sources/ggml/src/CMakeLists.txt
210
+ - ext/sources/ggml/src/ggml-alloc.c
211
+ - ext/sources/ggml/src/ggml-amx/CMakeLists.txt
212
+ - ext/sources/ggml/src/ggml-amx/common.h
213
+ - ext/sources/ggml/src/ggml-amx/ggml-amx.cpp
214
+ - ext/sources/ggml/src/ggml-amx/mmq.cpp
215
+ - ext/sources/ggml/src/ggml-amx/mmq.h
216
+ - ext/sources/ggml/src/ggml-backend-impl.h
217
+ - ext/sources/ggml/src/ggml-backend-reg.cpp
218
+ - ext/sources/ggml/src/ggml-backend.cpp
219
+ - ext/sources/ggml/src/ggml-blas/CMakeLists.txt
220
+ - ext/sources/ggml/src/ggml-blas/ggml-blas.cpp
221
+ - ext/sources/ggml/src/ggml-cann/CMakeLists.txt
222
+ - ext/sources/ggml/src/ggml-cann/Doxyfile
223
+ - ext/sources/ggml/src/ggml-cann/acl_tensor.cpp
224
+ - ext/sources/ggml/src/ggml-cann/acl_tensor.h
225
+ - ext/sources/ggml/src/ggml-cann/aclnn_ops.cpp
226
+ - ext/sources/ggml/src/ggml-cann/aclnn_ops.h
227
+ - ext/sources/ggml/src/ggml-cann/common.h
228
+ - ext/sources/ggml/src/ggml-cann/ggml-cann.cpp
229
+ - ext/sources/ggml/src/ggml-cann/kernels/CMakeLists.txt
230
+ - ext/sources/ggml/src/ggml-cann/kernels/ascendc_kernels.h
231
+ - ext/sources/ggml/src/ggml-cann/kernels/dup.cpp
232
+ - ext/sources/ggml/src/ggml-cann/kernels/get_row_f16.cpp
233
+ - ext/sources/ggml/src/ggml-cann/kernels/get_row_f32.cpp
234
+ - ext/sources/ggml/src/ggml-cann/kernels/get_row_q4_0.cpp
235
+ - ext/sources/ggml/src/ggml-cann/kernels/get_row_q8_0.cpp
236
+ - ext/sources/ggml/src/ggml-cann/kernels/quantize_f16_q8_0.cpp
237
+ - ext/sources/ggml/src/ggml-cann/kernels/quantize_f32_q8_0.cpp
238
+ - ext/sources/ggml/src/ggml-cann/kernels/quantize_float_to_q4_0.cpp
239
+ - ext/sources/ggml/src/ggml-common.h
240
+ - ext/sources/ggml/src/ggml-cpu/CMakeLists.txt
241
+ - ext/sources/ggml/src/ggml-cpu/amx/amx.cpp
242
+ - ext/sources/ggml/src/ggml-cpu/amx/amx.h
243
+ - ext/sources/ggml/src/ggml-cpu/amx/common.h
244
+ - ext/sources/ggml/src/ggml-cpu/amx/mmq.cpp
245
+ - ext/sources/ggml/src/ggml-cpu/amx/mmq.h
246
+ - ext/sources/ggml/src/ggml-cpu/binary-ops.cpp
247
+ - ext/sources/ggml/src/ggml-cpu/binary-ops.h
248
+ - ext/sources/ggml/src/ggml-cpu/cmake/FindSIMD.cmake
249
+ - ext/sources/ggml/src/ggml-cpu/common.h
250
+ - ext/sources/ggml/src/ggml-cpu/cpu-feats-x86.cpp
251
+ - ext/sources/ggml/src/ggml-cpu/ggml-cpu-aarch64.cpp
252
+ - ext/sources/ggml/src/ggml-cpu/ggml-cpu-aarch64.h
253
+ - ext/sources/ggml/src/ggml-cpu/ggml-cpu-hbm.cpp
254
+ - ext/sources/ggml/src/ggml-cpu/ggml-cpu-hbm.h
255
+ - ext/sources/ggml/src/ggml-cpu/ggml-cpu-impl.h
256
+ - ext/sources/ggml/src/ggml-cpu/ggml-cpu-quants.c
257
+ - ext/sources/ggml/src/ggml-cpu/ggml-cpu-quants.h
258
+ - ext/sources/ggml/src/ggml-cpu/ggml-cpu-traits.cpp
259
+ - ext/sources/ggml/src/ggml-cpu/ggml-cpu-traits.h
260
+ - ext/sources/ggml/src/ggml-cpu/ggml-cpu.c
261
+ - ext/sources/ggml/src/ggml-cpu/ggml-cpu.cpp
262
+ - ext/sources/ggml/src/ggml-cpu/kleidiai/kernels.cpp
263
+ - ext/sources/ggml/src/ggml-cpu/kleidiai/kernels.h
264
+ - ext/sources/ggml/src/ggml-cpu/kleidiai/kleidiai.cpp
265
+ - ext/sources/ggml/src/ggml-cpu/kleidiai/kleidiai.h
266
+ - ext/sources/ggml/src/ggml-cpu/llamafile/sgemm.cpp
267
+ - ext/sources/ggml/src/ggml-cpu/llamafile/sgemm.h
268
+ - ext/sources/ggml/src/ggml-cpu/ops.cpp
269
+ - ext/sources/ggml/src/ggml-cpu/ops.h
270
+ - ext/sources/ggml/src/ggml-cpu/simd-mappings.h
271
+ - ext/sources/ggml/src/ggml-cpu/unary-ops.cpp
272
+ - ext/sources/ggml/src/ggml-cpu/unary-ops.h
273
+ - ext/sources/ggml/src/ggml-cpu/vec.cpp
274
+ - ext/sources/ggml/src/ggml-cpu/vec.h
275
+ - ext/sources/ggml/src/ggml-cuda/CMakeLists.txt
276
+ - ext/sources/ggml/src/ggml-cuda/acc.cu
277
+ - ext/sources/ggml/src/ggml-cuda/acc.cuh
278
+ - ext/sources/ggml/src/ggml-cuda/arange.cu
279
+ - ext/sources/ggml/src/ggml-cuda/arange.cuh
280
+ - ext/sources/ggml/src/ggml-cuda/argmax.cu
281
+ - ext/sources/ggml/src/ggml-cuda/argmax.cuh
282
+ - ext/sources/ggml/src/ggml-cuda/argsort.cu
283
+ - ext/sources/ggml/src/ggml-cuda/argsort.cuh
284
+ - ext/sources/ggml/src/ggml-cuda/binbcast.cu
285
+ - ext/sources/ggml/src/ggml-cuda/binbcast.cuh
286
+ - ext/sources/ggml/src/ggml-cuda/clamp.cu
287
+ - ext/sources/ggml/src/ggml-cuda/clamp.cuh
288
+ - ext/sources/ggml/src/ggml-cuda/common.cuh
289
+ - ext/sources/ggml/src/ggml-cuda/concat.cu
290
+ - ext/sources/ggml/src/ggml-cuda/concat.cuh
291
+ - ext/sources/ggml/src/ggml-cuda/conv-transpose-1d.cu
292
+ - ext/sources/ggml/src/ggml-cuda/conv-transpose-1d.cuh
293
+ - ext/sources/ggml/src/ggml-cuda/convert.cu
294
+ - ext/sources/ggml/src/ggml-cuda/convert.cuh
295
+ - ext/sources/ggml/src/ggml-cuda/count-equal.cu
296
+ - ext/sources/ggml/src/ggml-cuda/count-equal.cuh
297
+ - ext/sources/ggml/src/ggml-cuda/cp-async.cuh
298
+ - ext/sources/ggml/src/ggml-cuda/cpy.cu
299
+ - ext/sources/ggml/src/ggml-cuda/cpy.cuh
300
+ - ext/sources/ggml/src/ggml-cuda/cross-entropy-loss.cu
301
+ - ext/sources/ggml/src/ggml-cuda/cross-entropy-loss.cuh
302
+ - ext/sources/ggml/src/ggml-cuda/dequantize.cuh
303
+ - ext/sources/ggml/src/ggml-cuda/diagmask.cu
304
+ - ext/sources/ggml/src/ggml-cuda/diagmask.cuh
305
+ - ext/sources/ggml/src/ggml-cuda/fattn-common.cuh
306
+ - ext/sources/ggml/src/ggml-cuda/fattn-mma-f16.cuh
307
+ - ext/sources/ggml/src/ggml-cuda/fattn-tile-f16.cu
308
+ - ext/sources/ggml/src/ggml-cuda/fattn-tile-f16.cuh
309
+ - ext/sources/ggml/src/ggml-cuda/fattn-tile-f32.cu
310
+ - ext/sources/ggml/src/ggml-cuda/fattn-tile-f32.cuh
311
+ - ext/sources/ggml/src/ggml-cuda/fattn-vec-f16.cuh
312
+ - ext/sources/ggml/src/ggml-cuda/fattn-vec-f32.cuh
313
+ - ext/sources/ggml/src/ggml-cuda/fattn-wmma-f16.cu
314
+ - ext/sources/ggml/src/ggml-cuda/fattn-wmma-f16.cuh
315
+ - ext/sources/ggml/src/ggml-cuda/fattn.cu
316
+ - ext/sources/ggml/src/ggml-cuda/fattn.cuh
317
+ - ext/sources/ggml/src/ggml-cuda/getrows.cu
318
+ - ext/sources/ggml/src/ggml-cuda/getrows.cuh
319
+ - ext/sources/ggml/src/ggml-cuda/ggml-cuda.cu
320
+ - ext/sources/ggml/src/ggml-cuda/gla.cu
321
+ - ext/sources/ggml/src/ggml-cuda/gla.cuh
322
+ - ext/sources/ggml/src/ggml-cuda/im2col.cu
323
+ - ext/sources/ggml/src/ggml-cuda/im2col.cuh
324
+ - ext/sources/ggml/src/ggml-cuda/mma.cuh
325
+ - ext/sources/ggml/src/ggml-cuda/mmq.cu
326
+ - ext/sources/ggml/src/ggml-cuda/mmq.cuh
327
+ - ext/sources/ggml/src/ggml-cuda/mmv.cu
328
+ - ext/sources/ggml/src/ggml-cuda/mmv.cuh
329
+ - ext/sources/ggml/src/ggml-cuda/mmvq.cu
330
+ - ext/sources/ggml/src/ggml-cuda/mmvq.cuh
331
+ - ext/sources/ggml/src/ggml-cuda/norm.cu
332
+ - ext/sources/ggml/src/ggml-cuda/norm.cuh
333
+ - ext/sources/ggml/src/ggml-cuda/opt-step-adamw.cu
334
+ - ext/sources/ggml/src/ggml-cuda/opt-step-adamw.cuh
335
+ - ext/sources/ggml/src/ggml-cuda/out-prod.cu
336
+ - ext/sources/ggml/src/ggml-cuda/out-prod.cuh
337
+ - ext/sources/ggml/src/ggml-cuda/pad.cu
338
+ - ext/sources/ggml/src/ggml-cuda/pad.cuh
339
+ - ext/sources/ggml/src/ggml-cuda/pool2d.cu
340
+ - ext/sources/ggml/src/ggml-cuda/pool2d.cuh
341
+ - ext/sources/ggml/src/ggml-cuda/quantize.cu
342
+ - ext/sources/ggml/src/ggml-cuda/quantize.cuh
343
+ - ext/sources/ggml/src/ggml-cuda/rope.cu
344
+ - ext/sources/ggml/src/ggml-cuda/rope.cuh
345
+ - ext/sources/ggml/src/ggml-cuda/scale.cu
346
+ - ext/sources/ggml/src/ggml-cuda/scale.cuh
347
+ - ext/sources/ggml/src/ggml-cuda/softmax.cu
348
+ - ext/sources/ggml/src/ggml-cuda/softmax.cuh
349
+ - ext/sources/ggml/src/ggml-cuda/ssm-conv.cu
350
+ - ext/sources/ggml/src/ggml-cuda/ssm-conv.cuh
351
+ - ext/sources/ggml/src/ggml-cuda/ssm-scan.cu
352
+ - ext/sources/ggml/src/ggml-cuda/ssm-scan.cuh
353
+ - ext/sources/ggml/src/ggml-cuda/sum.cu
354
+ - ext/sources/ggml/src/ggml-cuda/sum.cuh
355
+ - ext/sources/ggml/src/ggml-cuda/sumrows.cu
356
+ - ext/sources/ggml/src/ggml-cuda/sumrows.cuh
357
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_1-ncols2_16.cu
358
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_1-ncols2_8.cu
359
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_1.cu
360
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_2.cu
361
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_4.cu
362
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_16.cu
363
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_4.cu
364
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_8.cu
365
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_32-ncols2_1.cu
366
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_32-ncols2_2.cu
367
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_16.cu
368
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_2.cu
369
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_4.cu
370
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_8.cu
371
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_64-ncols2_1.cu
372
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_1.cu
373
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_2.cu
374
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_4.cu
375
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_8.cu
376
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-f16.cu
377
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q4_0.cu
378
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q4_1.cu
379
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q5_0.cu
380
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q5_1.cu
381
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q8_0.cu
382
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-f16.cu
383
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q4_0.cu
384
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q4_1.cu
385
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q5_0.cu
386
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q5_1.cu
387
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q8_0.cu
388
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-f16.cu
389
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q4_0.cu
390
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q4_1.cu
391
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q5_0.cu
392
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q5_1.cu
393
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q8_0.cu
394
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-f16.cu
395
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q4_0.cu
396
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q4_1.cu
397
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q5_0.cu
398
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q5_1.cu
399
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q8_0.cu
400
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-f16.cu
401
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q4_0.cu
402
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q4_1.cu
403
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q5_0.cu
404
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q5_1.cu
405
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q8_0.cu
406
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-f16.cu
407
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q4_0.cu
408
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q4_1.cu
409
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q5_0.cu
410
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q5_1.cu
411
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q8_0.cu
412
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs256-f16-f16.cu
413
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-f16.cu
414
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q4_0.cu
415
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q4_1.cu
416
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q5_0.cu
417
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q5_1.cu
418
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q8_0.cu
419
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-f16.cu
420
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q4_0.cu
421
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q4_1.cu
422
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q5_0.cu
423
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q5_1.cu
424
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q8_0.cu
425
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-f16.cu
426
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q4_0.cu
427
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q4_1.cu
428
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q5_0.cu
429
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q5_1.cu
430
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q8_0.cu
431
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-f16.cu
432
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q4_0.cu
433
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q4_1.cu
434
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q5_0.cu
435
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q5_1.cu
436
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q8_0.cu
437
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-f16.cu
438
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q4_0.cu
439
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q4_1.cu
440
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q5_0.cu
441
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q5_1.cu
442
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q8_0.cu
443
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-f16.cu
444
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q4_0.cu
445
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q4_1.cu
446
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q5_0.cu
447
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q5_1.cu
448
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q8_0.cu
449
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-f16.cu
450
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q4_0.cu
451
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q4_1.cu
452
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q5_0.cu
453
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q5_1.cu
454
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q8_0.cu
455
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs256-f16-f16.cu
456
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-f16.cu
457
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q4_0.cu
458
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q4_1.cu
459
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q5_0.cu
460
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q5_1.cu
461
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q8_0.cu
462
+ - ext/sources/ggml/src/ggml-cuda/template-instances/generate_cu_files.py
463
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq1_s.cu
464
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq2_s.cu
465
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq2_xs.cu
466
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq2_xxs.cu
467
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq3_s.cu
468
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq3_xxs.cu
469
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq4_nl.cu
470
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq4_xs.cu
471
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q2_k.cu
472
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q3_k.cu
473
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q4_0.cu
474
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q4_1.cu
475
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q4_k.cu
476
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q5_0.cu
477
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q5_1.cu
478
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q5_k.cu
479
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q6_k.cu
480
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q8_0.cu
481
+ - ext/sources/ggml/src/ggml-cuda/tsembd.cu
482
+ - ext/sources/ggml/src/ggml-cuda/tsembd.cuh
483
+ - ext/sources/ggml/src/ggml-cuda/unary.cu
484
+ - ext/sources/ggml/src/ggml-cuda/unary.cuh
485
+ - ext/sources/ggml/src/ggml-cuda/upscale.cu
486
+ - ext/sources/ggml/src/ggml-cuda/upscale.cuh
487
+ - ext/sources/ggml/src/ggml-cuda/vecdotq.cuh
488
+ - ext/sources/ggml/src/ggml-cuda/vendors/cuda.h
489
+ - ext/sources/ggml/src/ggml-cuda/vendors/hip.h
490
+ - ext/sources/ggml/src/ggml-cuda/vendors/musa.h
491
+ - ext/sources/ggml/src/ggml-cuda/wkv.cu
492
+ - ext/sources/ggml/src/ggml-cuda/wkv.cuh
493
+ - ext/sources/ggml/src/ggml-hip/CMakeLists.txt
494
+ - ext/sources/ggml/src/ggml-impl.h
495
+ - ext/sources/ggml/src/ggml-kompute/CMakeLists.txt
496
+ - ext/sources/ggml/src/ggml-kompute/ggml-kompute.cpp
497
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/common.comp
498
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_add.comp
499
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_addrow.comp
500
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f16_f16.comp
501
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f16_f32.comp
502
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f32_f16.comp
503
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_cpy_f32_f32.comp
504
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_diagmask.comp
505
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_gelu.comp
506
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows.comp
507
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows_f16.comp
508
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows_f32.comp
509
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q4_0.comp
510
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q4_1.comp
511
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_getrows_q6_k.comp
512
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul.comp
513
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_f16.comp
514
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_mat_f32.comp
515
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_0.comp
516
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_1.comp
517
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q4_k.comp
518
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q6_k.comp
519
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mat_q8_0.comp
520
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n.comp
521
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_mul_mv_q_n_pre.comp
522
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_norm.comp
523
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_relu.comp
524
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_rmsnorm.comp
525
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f16.comp
526
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_rope_neox_f32.comp
527
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f16.comp
528
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_rope_norm_f32.comp
529
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_scale.comp
530
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_scale_8.comp
531
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_silu.comp
532
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/op_softmax.comp
533
+ - ext/sources/ggml/src/ggml-kompute/kompute-shaders/rope_common.comp
534
+ - ext/sources/ggml/src/ggml-metal/CMakeLists.txt
535
+ - ext/sources/ggml/src/ggml-metal/ggml-metal-impl.h
536
+ - ext/sources/ggml/src/ggml-metal/ggml-metal.m
537
+ - ext/sources/ggml/src/ggml-metal/ggml-metal.metal
538
+ - ext/sources/ggml/src/ggml-musa/CMakeLists.txt
539
+ - ext/sources/ggml/src/ggml-musa/mudnn.cu
540
+ - ext/sources/ggml/src/ggml-musa/mudnn.cuh
541
+ - ext/sources/ggml/src/ggml-opencl/CMakeLists.txt
542
+ - ext/sources/ggml/src/ggml-opencl/ggml-opencl.cpp
543
+ - ext/sources/ggml/src/ggml-opencl/kernels/add.cl
544
+ - ext/sources/ggml/src/ggml-opencl/kernels/clamp.cl
545
+ - ext/sources/ggml/src/ggml-opencl/kernels/cpy.cl
546
+ - ext/sources/ggml/src/ggml-opencl/kernels/cvt.cl
547
+ - ext/sources/ggml/src/ggml-opencl/kernels/diag_mask_inf.cl
548
+ - ext/sources/ggml/src/ggml-opencl/kernels/embed_kernel.py
549
+ - ext/sources/ggml/src/ggml-opencl/kernels/gelu.cl
550
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle.cl
551
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_general.cl
552
+ - ext/sources/ggml/src/ggml-opencl/kernels/get_rows.cl
553
+ - ext/sources/ggml/src/ggml-opencl/kernels/im2col_f16.cl
554
+ - ext/sources/ggml/src/ggml-opencl/kernels/im2col_f32.cl
555
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul.cl
556
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mat_Ab_Bi_8x4.cl
557
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f16_f16.cl
558
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f16_f32.cl
559
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f16_f32_1row.cl
560
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f16_f32_l4.cl
561
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f32_f32.cl
562
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32.cl
563
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32_1d_16x_flat.cl
564
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32_1d_8x_flat.cl
565
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32_8x_flat.cl
566
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32_v.cl
567
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q6_k.cl
568
+ - ext/sources/ggml/src/ggml-opencl/kernels/norm.cl
569
+ - ext/sources/ggml/src/ggml-opencl/kernels/relu.cl
570
+ - ext/sources/ggml/src/ggml-opencl/kernels/rms_norm.cl
571
+ - ext/sources/ggml/src/ggml-opencl/kernels/rope.cl
572
+ - ext/sources/ggml/src/ggml-opencl/kernels/scale.cl
573
+ - ext/sources/ggml/src/ggml-opencl/kernels/silu.cl
574
+ - ext/sources/ggml/src/ggml-opencl/kernels/softmax_4_f16.cl
575
+ - ext/sources/ggml/src/ggml-opencl/kernels/softmax_4_f32.cl
576
+ - ext/sources/ggml/src/ggml-opencl/kernels/softmax_f16.cl
577
+ - ext/sources/ggml/src/ggml-opencl/kernels/softmax_f32.cl
578
+ - ext/sources/ggml/src/ggml-opencl/kernels/transpose.cl
579
+ - ext/sources/ggml/src/ggml-opt.cpp
580
+ - ext/sources/ggml/src/ggml-quants.c
581
+ - ext/sources/ggml/src/ggml-quants.h
582
+ - ext/sources/ggml/src/ggml-rpc/CMakeLists.txt
583
+ - ext/sources/ggml/src/ggml-rpc/ggml-rpc.cpp
584
+ - ext/sources/ggml/src/ggml-sycl/CMakeLists.txt
585
+ - ext/sources/ggml/src/ggml-sycl/backend.hpp
586
+ - ext/sources/ggml/src/ggml-sycl/binbcast.cpp
587
+ - ext/sources/ggml/src/ggml-sycl/binbcast.hpp
588
+ - ext/sources/ggml/src/ggml-sycl/common.cpp
589
+ - ext/sources/ggml/src/ggml-sycl/common.hpp
590
+ - ext/sources/ggml/src/ggml-sycl/concat.cpp
591
+ - ext/sources/ggml/src/ggml-sycl/concat.hpp
592
+ - ext/sources/ggml/src/ggml-sycl/conv.cpp
593
+ - ext/sources/ggml/src/ggml-sycl/conv.hpp
594
+ - ext/sources/ggml/src/ggml-sycl/convert.cpp
595
+ - ext/sources/ggml/src/ggml-sycl/convert.hpp
596
+ - ext/sources/ggml/src/ggml-sycl/cpy.cpp
597
+ - ext/sources/ggml/src/ggml-sycl/cpy.hpp
598
+ - ext/sources/ggml/src/ggml-sycl/dequantize.hpp
599
+ - ext/sources/ggml/src/ggml-sycl/dmmv.cpp
600
+ - ext/sources/ggml/src/ggml-sycl/dmmv.hpp
601
+ - ext/sources/ggml/src/ggml-sycl/dpct/helper.hpp
602
+ - ext/sources/ggml/src/ggml-sycl/element_wise.cpp
603
+ - ext/sources/ggml/src/ggml-sycl/element_wise.hpp
604
+ - ext/sources/ggml/src/ggml-sycl/gemm.hpp
605
+ - ext/sources/ggml/src/ggml-sycl/getrows.cpp
606
+ - ext/sources/ggml/src/ggml-sycl/getrows.hpp
607
+ - ext/sources/ggml/src/ggml-sycl/ggml-sycl.cpp
608
+ - ext/sources/ggml/src/ggml-sycl/gla.cpp
609
+ - ext/sources/ggml/src/ggml-sycl/gla.hpp
610
+ - ext/sources/ggml/src/ggml-sycl/im2col.cpp
611
+ - ext/sources/ggml/src/ggml-sycl/im2col.hpp
612
+ - ext/sources/ggml/src/ggml-sycl/mmq.cpp
613
+ - ext/sources/ggml/src/ggml-sycl/mmq.hpp
614
+ - ext/sources/ggml/src/ggml-sycl/mmvq.cpp
615
+ - ext/sources/ggml/src/ggml-sycl/mmvq.hpp
616
+ - ext/sources/ggml/src/ggml-sycl/norm.cpp
617
+ - ext/sources/ggml/src/ggml-sycl/norm.hpp
618
+ - ext/sources/ggml/src/ggml-sycl/outprod.cpp
619
+ - ext/sources/ggml/src/ggml-sycl/outprod.hpp
620
+ - ext/sources/ggml/src/ggml-sycl/presets.hpp
621
+ - ext/sources/ggml/src/ggml-sycl/quants.hpp
622
+ - ext/sources/ggml/src/ggml-sycl/rope.cpp
623
+ - ext/sources/ggml/src/ggml-sycl/rope.hpp
624
+ - ext/sources/ggml/src/ggml-sycl/softmax.cpp
625
+ - ext/sources/ggml/src/ggml-sycl/softmax.hpp
626
+ - ext/sources/ggml/src/ggml-sycl/sycl_hw.cpp
627
+ - ext/sources/ggml/src/ggml-sycl/sycl_hw.hpp
628
+ - ext/sources/ggml/src/ggml-sycl/tsembd.cpp
629
+ - ext/sources/ggml/src/ggml-sycl/tsembd.hpp
630
+ - ext/sources/ggml/src/ggml-sycl/vecdotq.hpp
631
+ - ext/sources/ggml/src/ggml-sycl/wkv.cpp
632
+ - ext/sources/ggml/src/ggml-sycl/wkv.hpp
633
+ - ext/sources/ggml/src/ggml-threading.cpp
634
+ - ext/sources/ggml/src/ggml-threading.h
635
+ - ext/sources/ggml/src/ggml-vulkan/CMakeLists.txt
636
+ - ext/sources/ggml/src/ggml-vulkan/cmake/host-toolchain.cmake.in
637
+ - ext/sources/ggml/src/ggml-vulkan/ggml-vulkan.cpp
638
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt
639
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/acc.comp
640
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/add.comp
641
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/argmax.comp
642
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/argsort.comp
643
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/clamp.comp
644
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/concat.comp
645
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/contig_copy.comp
646
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_dw.comp
647
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/copy.comp
648
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/copy_from_quant.comp
649
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/copy_to_quant.comp
650
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/cos.comp
651
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/count_equal.comp
652
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_f32.comp
653
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.comp
654
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.comp
655
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_head.comp
656
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_m.comp
657
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_s.comp
658
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_s.comp
659
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xs.comp
660
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xxs.comp
661
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_s.comp
662
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_xxs.comp
663
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp
664
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_xs.comp
665
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q2_k.comp
666
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q3_k.comp
667
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_0.comp
668
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_1.comp
669
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_k.comp
670
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_0.comp
671
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_1.comp
672
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_k.comp
673
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q6_k.comp
674
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q8_0.comp
675
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/diag_mask_inf.comp
676
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/div.comp
677
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp
678
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_base.comp
679
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp
680
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp
681
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_split_k_reduce.comp
682
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/gelu.comp
683
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/gelu_quick.comp
684
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.comp
685
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/generic_head.comp
686
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/generic_unary_head.comp
687
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/get_rows.comp
688
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/get_rows_quant.comp
689
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/group_norm.comp
690
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/im2col.comp
691
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/l2_norm.comp
692
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/leaky_relu.comp
693
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul.comp
694
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_split_k_reduce.comp
695
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp
696
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.comp
697
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_m.comp
698
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_s.comp
699
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_s.comp
700
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xs.comp
701
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xxs.comp
702
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_s.comp
703
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_xxs.comp
704
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_nc.comp
705
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_p021.comp
706
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
707
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
708
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
709
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
710
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
711
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp
712
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp
713
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq.comp
714
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.comp
715
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/norm.comp
716
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_adamw.comp
717
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/pad.comp
718
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/pool2d.comp
719
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/quantize_q8_1.comp
720
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/relu.comp
721
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/repeat.comp
722
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/repeat_back.comp
723
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp
724
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_back.comp
725
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.comp
726
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_multi.comp
727
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_neox.comp
728
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_norm.comp
729
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_vision.comp
730
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp
731
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sigmoid.comp
732
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/silu.comp
733
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/silu_back.comp
734
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp
735
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp
736
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/soft_max_back.comp
737
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/square.comp
738
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sub.comp
739
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sum_rows.comp
740
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/tanh.comp
741
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/test_bfloat16_support.comp
742
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat2_support.comp
743
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat_support.comp
744
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/test_integer_dot_support.comp
745
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/timestep_embedding.comp
746
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/types.comp
747
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/upscale.comp
748
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp
749
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/wkv6.comp
750
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/wkv7.comp
751
+ - ext/sources/ggml/src/ggml.c
752
+ - ext/sources/ggml/src/gguf.cpp
753
+ - ext/sources/include/whisper.h
754
+ - ext/sources/src/CMakeLists.txt
755
+ - ext/sources/src/coreml/whisper-decoder-impl.h
756
+ - ext/sources/src/coreml/whisper-decoder-impl.m
757
+ - ext/sources/src/coreml/whisper-encoder-impl.h
758
+ - ext/sources/src/coreml/whisper-encoder-impl.m
759
+ - ext/sources/src/coreml/whisper-encoder.h
760
+ - ext/sources/src/coreml/whisper-encoder.mm
761
+ - ext/sources/src/openvino/whisper-openvino-encoder.cpp
762
+ - ext/sources/src/openvino/whisper-openvino-encoder.h
763
+ - ext/sources/src/whisper-arch.h
764
+ - ext/sources/src/whisper.cpp
765
+ - ext/sources/tests/CMakeLists.txt
766
+ - ext/sources/tests/earnings21/eval.mk
767
+ - ext/sources/tests/earnings21/eval.py
768
+ - ext/sources/tests/earnings21/normalizers/__init__.py
769
+ - ext/sources/tests/earnings21/normalizers/basic.py
770
+ - ext/sources/tests/earnings21/normalizers/english.json
771
+ - ext/sources/tests/earnings21/normalizers/english.py
772
+ - ext/sources/tests/earnings21/requirements.txt
773
+ - ext/sources/tests/en-0-ref.txt
774
+ - ext/sources/tests/en-1-ref.txt
775
+ - ext/sources/tests/en-2-ref.txt
776
+ - ext/sources/tests/es-0-ref.txt
777
+ - ext/sources/tests/librispeech/eval.mk
778
+ - ext/sources/tests/librispeech/eval.py
779
+ - ext/sources/tests/librispeech/normalizers/__init__.py
780
+ - ext/sources/tests/librispeech/normalizers/basic.py
781
+ - ext/sources/tests/librispeech/normalizers/english.json
782
+ - ext/sources/tests/librispeech/normalizers/english.py
783
+ - ext/sources/tests/librispeech/requirements.txt
784
+ - ext/sources/tests/run-tests.sh
785
+ - ext/sources/tests/test-c.c
786
+ - ext/sources/tests/test-vad-full.cpp
787
+ - ext/sources/tests/test-vad.cpp
788
+ - ext/sources/tests/test-whisper.js
138
789
  - extsources.rb
139
- - lib/whisper.rb
140
790
  - lib/whisper/model/uri.rb
791
+ - sig/whisper.rbs
141
792
  - tests/helper.rb
142
793
  - tests/jfk_reader/.gitignore
143
794
  - tests/jfk_reader/extconf.rb
@@ -148,9 +799,11 @@ files:
148
799
  - tests/test_package.rb
149
800
  - tests/test_params.rb
150
801
  - tests/test_segment.rb
802
+ - tests/test_vad.rb
803
+ - tests/test_vad_params.rb
151
804
  - tests/test_whisper.rb
152
805
  - whispercpp.gemspec
153
- homepage: https://github.com/ggerganov/whisper.cpp
806
+ homepage: https://github.com/ggml-org/whisper.cpp
154
807
  licenses:
155
808
  - MIT
156
809
  metadata: {}
@@ -186,4 +839,6 @@ test_files:
186
839
  - tests/test_package.rb
187
840
  - tests/test_params.rb
188
841
  - tests/test_segment.rb
842
+ - tests/test_vad.rb
843
+ - tests/test_vad_params.rb
189
844
  - tests/test_whisper.rb