whispercpp 1.3.5 → 1.3.7

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 (1017) hide show
  1. checksums.yaml +4 -4
  2. data/.document +3 -0
  3. data/.rdoc_options +2 -0
  4. data/LICENSE +1 -1
  5. data/README.md +133 -3
  6. data/Rakefile +18 -3
  7. data/ext/dependencies.rb +10 -4
  8. data/ext/dependencies_for_windows.rb +17 -0
  9. data/ext/extconf.rb +20 -7
  10. data/ext/options.rb +54 -14
  11. data/ext/options_for_windows.rb +51 -0
  12. data/ext/ruby_whisper.c +56 -46
  13. data/ext/ruby_whisper.h +165 -2
  14. data/ext/ruby_whisper_context.c +297 -126
  15. data/ext/ruby_whisper_context_params.c +163 -0
  16. data/ext/ruby_whisper_log_queue.c +180 -0
  17. data/ext/ruby_whisper_log_settable.h +47 -0
  18. data/ext/ruby_whisper_model.c +0 -1
  19. data/ext/ruby_whisper_parakeet.c +49 -0
  20. data/ext/ruby_whisper_parakeet_context.c +304 -0
  21. data/ext/ruby_whisper_parakeet_context_params.c +117 -0
  22. data/ext/ruby_whisper_parakeet_model.c +84 -0
  23. data/ext/ruby_whisper_parakeet_params.c +548 -0
  24. data/ext/ruby_whisper_parakeet_segment.c +157 -0
  25. data/ext/ruby_whisper_parakeet_token.c +188 -0
  26. data/ext/ruby_whisper_parakeet_transcribe.cpp +58 -0
  27. data/ext/ruby_whisper_params.c +256 -66
  28. data/ext/ruby_whisper_segment.c +6 -7
  29. data/ext/ruby_whisper_token.c +29 -9
  30. data/ext/ruby_whisper_transcribe.cpp +46 -16
  31. data/ext/ruby_whisper_vad_context.c +48 -1
  32. data/ext/ruby_whisper_vad_context_detect.cpp +6 -5
  33. data/ext/ruby_whisper_vad_params.c +0 -1
  34. data/ext/ruby_whisper_vad_segment.c +0 -1
  35. data/ext/ruby_whisper_vad_segments.c +0 -1
  36. data/ext/sources/CMakeLists.txt +41 -3
  37. data/ext/sources/CMakePresets.json +95 -0
  38. data/ext/sources/cmake/parakeet-config.cmake.in +30 -0
  39. data/ext/sources/cmake/parakeet.pc.in +10 -0
  40. data/ext/sources/cmake/whisper-config.cmake.in +5 -40
  41. data/ext/sources/cmake/whisper.pc.in +1 -1
  42. data/ext/sources/examples/CMakeLists.txt +4 -2
  43. data/ext/sources/examples/bench/bench.cpp +24 -19
  44. data/ext/sources/examples/cli/cli.cpp +51 -9
  45. data/ext/sources/examples/common-ggml.cpp +4 -0
  46. data/ext/sources/examples/common-whisper.cpp +139 -67
  47. data/ext/sources/examples/common-whisper.h +11 -0
  48. data/ext/sources/examples/ffmpeg-transcode.cpp +211 -341
  49. data/ext/sources/examples/miniaudio.h +4507 -2131
  50. data/ext/sources/examples/parakeet-cli/CMakeLists.txt +8 -0
  51. data/ext/sources/examples/parakeet-cli/parakeet-cli.cpp +243 -0
  52. data/ext/sources/examples/parakeet-quantize/CMakeLists.txt +7 -0
  53. data/ext/sources/examples/parakeet-quantize/parakeet-quantize.cpp +230 -0
  54. data/ext/sources/examples/server/server.cpp +213 -163
  55. data/ext/sources/ggml/CMakeLists.txt +29 -15
  56. data/ext/sources/ggml/cmake/FindNCCL.cmake +36 -0
  57. data/ext/sources/ggml/cmake/ggml-config.cmake.in +12 -2
  58. data/ext/sources/ggml/include/ggml-alloc.h +1 -0
  59. data/ext/sources/ggml/include/ggml-backend.h +73 -11
  60. data/ext/sources/ggml/include/ggml-cann.h +1 -1
  61. data/ext/sources/ggml/include/ggml-cpu.h +5 -0
  62. data/ext/sources/ggml/include/ggml-cuda.h +3 -0
  63. data/ext/sources/ggml/include/ggml-openvino.h +37 -0
  64. data/ext/sources/ggml/include/ggml-opt.h +1 -1
  65. data/ext/sources/ggml/include/ggml-rpc.h +8 -3
  66. data/ext/sources/ggml/include/ggml-virtgpu.h +14 -0
  67. data/ext/sources/ggml/include/ggml.h +155 -16
  68. data/ext/sources/ggml/include/gguf.h +10 -2
  69. data/ext/sources/ggml/src/CMakeLists.txt +25 -5
  70. data/ext/sources/ggml/src/ggml-alloc.c +9 -10
  71. data/ext/sources/ggml/src/ggml-backend-dl.cpp +48 -0
  72. data/ext/sources/ggml/src/ggml-backend-dl.h +45 -0
  73. data/ext/sources/ggml/src/ggml-backend-impl.h +22 -2
  74. data/ext/sources/ggml/src/ggml-backend-meta.cpp +2263 -0
  75. data/ext/sources/ggml/src/ggml-backend-reg.cpp +40 -86
  76. data/ext/sources/ggml/src/ggml-backend.cpp +114 -10
  77. data/ext/sources/ggml/src/ggml-blas/CMakeLists.txt +1 -1
  78. data/ext/sources/ggml/src/ggml-blas/ggml-blas.cpp +10 -2
  79. data/ext/sources/ggml/src/ggml-cann/acl_tensor.cpp +1 -1
  80. data/ext/sources/ggml/src/ggml-cann/acl_tensor.h +1 -1
  81. data/ext/sources/ggml/src/ggml-cann/aclnn_ops.cpp +1016 -442
  82. data/ext/sources/ggml/src/ggml-cann/aclnn_ops.h +111 -85
  83. data/ext/sources/ggml/src/ggml-cann/common.h +23 -14
  84. data/ext/sources/ggml/src/ggml-cann/ggml-cann.cpp +255 -92
  85. data/ext/sources/ggml/src/ggml-common.h +22 -0
  86. data/ext/sources/ggml/src/ggml-cpu/CMakeLists.txt +68 -34
  87. data/ext/sources/ggml/src/ggml-cpu/amx/amx.cpp +44 -19
  88. data/ext/sources/ggml/src/ggml-cpu/amx/common.h +34 -10
  89. data/ext/sources/ggml/src/ggml-cpu/amx/mmq.cpp +101 -101
  90. data/ext/sources/ggml/src/ggml-cpu/arch/arm/quants.c +194 -1
  91. data/ext/sources/ggml/src/ggml-cpu/arch/arm/repack.cpp +2874 -613
  92. data/ext/sources/ggml/src/ggml-cpu/arch/loongarch/quants.c +151 -1
  93. data/ext/sources/ggml/src/ggml-cpu/arch/powerpc/quants.c +0 -1
  94. data/ext/sources/ggml/src/ggml-cpu/arch/riscv/quants.c +5480 -840
  95. data/ext/sources/ggml/src/ggml-cpu/arch/riscv/repack.cpp +1361 -0
  96. data/ext/sources/ggml/src/ggml-cpu/arch/s390/quants.c +8 -11
  97. data/ext/sources/ggml/src/ggml-cpu/arch/wasm/quants.c +72 -1
  98. data/ext/sources/ggml/src/ggml-cpu/arch/x86/quants.c +186 -36
  99. data/ext/sources/ggml/src/ggml-cpu/arch/x86/repack.cpp +119 -19
  100. data/ext/sources/ggml/src/ggml-cpu/arch-fallback.h +112 -26
  101. data/ext/sources/ggml/src/ggml-cpu/binary-ops.cpp +2 -6
  102. data/ext/sources/ggml/src/ggml-cpu/cmake/FindSMTIME.cmake +32 -0
  103. data/ext/sources/ggml/src/ggml-cpu/common.h +8 -0
  104. data/ext/sources/ggml/src/ggml-cpu/ggml-cpu-impl.h +13 -0
  105. data/ext/sources/ggml/src/ggml-cpu/ggml-cpu.c +153 -16
  106. data/ext/sources/ggml/src/ggml-cpu/ggml-cpu.cpp +17 -0
  107. data/ext/sources/ggml/src/ggml-cpu/kleidiai/kernels.cpp +21 -20
  108. data/ext/sources/ggml/src/ggml-cpu/kleidiai/kleidiai.cpp +976 -251
  109. data/ext/sources/ggml/src/ggml-cpu/llamafile/sgemm.cpp +671 -266
  110. data/ext/sources/ggml/src/ggml-cpu/ops.cpp +1277 -263
  111. data/ext/sources/ggml/src/ggml-cpu/ops.h +4 -0
  112. data/ext/sources/ggml/src/ggml-cpu/quants.c +95 -0
  113. data/ext/sources/ggml/src/ggml-cpu/quants.h +6 -0
  114. data/ext/sources/ggml/src/ggml-cpu/repack.cpp +2893 -679
  115. data/ext/sources/ggml/src/ggml-cpu/repack.h +119 -8
  116. data/ext/sources/ggml/src/ggml-cpu/simd-gemm.h +226 -0
  117. data/ext/sources/ggml/src/ggml-cpu/simd-mappings.h +114 -19
  118. data/ext/sources/ggml/src/ggml-cpu/spacemit/ime.cpp +1402 -687
  119. data/ext/sources/ggml/src/ggml-cpu/spacemit/ime.h +8 -0
  120. data/ext/sources/ggml/src/ggml-cpu/spacemit/ime1_kernels.cpp +597 -2766
  121. data/ext/sources/ggml/src/ggml-cpu/spacemit/ime2_kernels.cpp +5768 -0
  122. data/ext/sources/ggml/src/ggml-cpu/spacemit/ime_env.cpp +320 -0
  123. data/ext/sources/ggml/src/ggml-cpu/spacemit/ime_env.h +55 -0
  124. data/ext/sources/ggml/src/ggml-cpu/spacemit/ime_kernels.h +182 -19
  125. data/ext/sources/ggml/src/ggml-cpu/spacemit/repack.cpp +1795 -0
  126. data/ext/sources/ggml/src/ggml-cpu/spacemit/repack.h +14 -0
  127. data/ext/sources/ggml/src/ggml-cpu/spacemit/rvv_kernels.cpp +3178 -0
  128. data/ext/sources/ggml/src/ggml-cpu/spacemit/rvv_kernels.h +95 -0
  129. data/ext/sources/ggml/src/ggml-cpu/spacemit/spine_barrier.h +34 -0
  130. data/ext/sources/ggml/src/ggml-cpu/spacemit/spine_mem_pool.cpp +760 -0
  131. data/ext/sources/ggml/src/ggml-cpu/spacemit/spine_mem_pool.h +32 -0
  132. data/ext/sources/ggml/src/ggml-cpu/spacemit/spine_tcm.h +409 -0
  133. data/ext/sources/ggml/src/ggml-cpu/unary-ops.cpp +1 -1
  134. data/ext/sources/ggml/src/ggml-cpu/vec.cpp +54 -53
  135. data/ext/sources/ggml/src/ggml-cpu/vec.h +225 -240
  136. data/ext/sources/ggml/src/ggml-cuda/CMakeLists.txt +18 -8
  137. data/ext/sources/ggml/src/ggml-cuda/allreduce.cu +971 -0
  138. data/ext/sources/ggml/src/ggml-cuda/allreduce.cuh +29 -0
  139. data/ext/sources/ggml/src/ggml-cuda/argsort.cu +73 -28
  140. data/ext/sources/ggml/src/ggml-cuda/binbcast.cu +69 -41
  141. data/ext/sources/ggml/src/ggml-cuda/binbcast.cuh +1 -0
  142. data/ext/sources/ggml/src/ggml-cuda/common.cuh +359 -29
  143. data/ext/sources/ggml/src/ggml-cuda/concat.cu +120 -114
  144. data/ext/sources/ggml/src/ggml-cuda/conv2d-transpose.cu +45 -21
  145. data/ext/sources/ggml/src/ggml-cuda/conv2d-transpose.cuh +1 -0
  146. data/ext/sources/ggml/src/ggml-cuda/convert.cu +94 -27
  147. data/ext/sources/ggml/src/ggml-cuda/convert.cuh +10 -0
  148. data/ext/sources/ggml/src/ggml-cuda/cpy.cu +20 -9
  149. data/ext/sources/ggml/src/ggml-cuda/dequantize.cuh +22 -0
  150. data/ext/sources/ggml/src/ggml-cuda/fattn-common.cuh +333 -85
  151. data/ext/sources/ggml/src/ggml-cuda/fattn-mma-f16.cuh +632 -190
  152. data/ext/sources/ggml/src/ggml-cuda/fattn-tile.cu +12 -0
  153. data/ext/sources/ggml/src/ggml-cuda/fattn-tile.cuh +162 -49
  154. data/ext/sources/ggml/src/ggml-cuda/fattn-vec.cuh +43 -18
  155. data/ext/sources/ggml/src/ggml-cuda/fattn-wmma-f16.cu +44 -14
  156. data/ext/sources/ggml/src/ggml-cuda/fattn-wmma-f16.cuh +1 -1
  157. data/ext/sources/ggml/src/ggml-cuda/fattn.cu +241 -23
  158. data/ext/sources/ggml/src/ggml-cuda/fattn.cuh +2 -0
  159. data/ext/sources/ggml/src/ggml-cuda/fwht.cu +101 -0
  160. data/ext/sources/ggml/src/ggml-cuda/fwht.cuh +4 -0
  161. data/ext/sources/ggml/src/ggml-cuda/gated_delta_net.cu +312 -0
  162. data/ext/sources/ggml/src/ggml-cuda/gated_delta_net.cuh +4 -0
  163. data/ext/sources/ggml/src/ggml-cuda/getrows.cu +34 -12
  164. data/ext/sources/ggml/src/ggml-cuda/ggml-cuda.cu +1454 -599
  165. data/ext/sources/ggml/src/ggml-cuda/im2col.cu +32 -29
  166. data/ext/sources/ggml/src/ggml-cuda/mean.cu +13 -10
  167. data/ext/sources/ggml/src/ggml-cuda/mma.cuh +397 -183
  168. data/ext/sources/ggml/src/ggml-cuda/mmf.cu +30 -10
  169. data/ext/sources/ggml/src/ggml-cuda/mmf.cuh +161 -88
  170. data/ext/sources/ggml/src/ggml-cuda/mmq.cu +18 -12
  171. data/ext/sources/ggml/src/ggml-cuda/mmq.cuh +522 -431
  172. data/ext/sources/ggml/src/ggml-cuda/mmvf.cu +139 -72
  173. data/ext/sources/ggml/src/ggml-cuda/mmvf.cuh +2 -0
  174. data/ext/sources/ggml/src/ggml-cuda/mmvq.cu +608 -88
  175. data/ext/sources/ggml/src/ggml-cuda/mmvq.cuh +6 -0
  176. data/ext/sources/ggml/src/ggml-cuda/norm.cu +47 -79
  177. data/ext/sources/ggml/src/ggml-cuda/out-prod.cu +23 -7
  178. data/ext/sources/ggml/src/ggml-cuda/pad.cu +13 -10
  179. data/ext/sources/ggml/src/ggml-cuda/quantize.cu +134 -27
  180. data/ext/sources/ggml/src/ggml-cuda/quantize.cuh +1 -1
  181. data/ext/sources/ggml/src/ggml-cuda/reduce_rows.cuh +7 -17
  182. data/ext/sources/ggml/src/ggml-cuda/rope.cu +244 -137
  183. data/ext/sources/ggml/src/ggml-cuda/scale.cu +4 -1
  184. data/ext/sources/ggml/src/ggml-cuda/set-rows.cu +14 -6
  185. data/ext/sources/ggml/src/ggml-cuda/snake.cu +72 -0
  186. data/ext/sources/ggml/src/ggml-cuda/snake.cuh +8 -0
  187. data/ext/sources/ggml/src/ggml-cuda/softcap.cu +4 -1
  188. data/ext/sources/ggml/src/ggml-cuda/softmax.cu +8 -83
  189. data/ext/sources/ggml/src/ggml-cuda/solve_tri.cu +1 -1
  190. data/ext/sources/ggml/src/ggml-cuda/ssm-conv.cu +96 -40
  191. data/ext/sources/ggml/src/ggml-cuda/ssm-conv.cuh +1 -1
  192. data/ext/sources/ggml/src/ggml-cuda/ssm-scan.cu +40 -18
  193. data/ext/sources/ggml/src/ggml-cuda/sumrows.cu +8 -4
  194. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_1-ncols2_16.cu +1 -0
  195. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_1-ncols2_32.cu +6 -0
  196. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_1-ncols2_8.cu +2 -0
  197. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_4.cu +2 -0
  198. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_16.cu +1 -0
  199. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_32.cu +6 -0
  200. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_4.cu +2 -0
  201. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_8.cu +2 -0
  202. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_16.cu +1 -0
  203. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_4.cu +2 -0
  204. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_4-ncols2_8.cu +2 -0
  205. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_4.cu +2 -0
  206. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_8.cu +2 -0
  207. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq192-dv128.cu +5 -0
  208. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq320-dv256.cu +5 -0
  209. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq512-dv512.cu +5 -0
  210. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-bf16.cu +7 -0
  211. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-f16.cu +7 -0
  212. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-q4_0.cu +7 -0
  213. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-q4_1.cu +7 -0
  214. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-q5_0.cu +7 -0
  215. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-q5_1.cu +7 -0
  216. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-q8_0.cu +7 -0
  217. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-bf16.cu +7 -0
  218. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-bf16.cu +7 -0
  219. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-bf16.cu +7 -0
  220. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-bf16.cu +7 -0
  221. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-bf16.cu +7 -0
  222. data/ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-bf16.cu +7 -0
  223. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-nvfp4.cu +5 -0
  224. data/ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q1_0.cu +5 -0
  225. data/ext/sources/ggml/src/ggml-cuda/top-k.cu +5 -5
  226. data/ext/sources/ggml/src/ggml-cuda/topk-moe.cu +202 -135
  227. data/ext/sources/ggml/src/ggml-cuda/topk-moe.cuh +20 -14
  228. data/ext/sources/ggml/src/ggml-cuda/unary.cu +86 -2
  229. data/ext/sources/ggml/src/ggml-cuda/unary.cuh +4 -0
  230. data/ext/sources/ggml/src/ggml-cuda/vecdotq.cuh +111 -17
  231. data/ext/sources/ggml/src/ggml-cuda/vendors/cuda.h +7 -2
  232. data/ext/sources/ggml/src/ggml-cuda/vendors/hip.h +30 -2
  233. data/ext/sources/ggml/src/ggml-cuda/vendors/musa.h +3 -0
  234. data/ext/sources/ggml/src/ggml-hexagon/CMakeLists.txt +84 -46
  235. data/ext/sources/ggml/src/ggml-hexagon/ggml-hexagon.cpp +1612 -753
  236. data/ext/sources/ggml/src/ggml-hexagon/htp/CMakeLists.txt +51 -11
  237. data/ext/sources/ggml/src/ggml-hexagon/htp/act-ops.c +361 -261
  238. data/ext/sources/ggml/src/ggml-hexagon/htp/argsort-ops.c +294 -0
  239. data/ext/sources/ggml/src/ggml-hexagon/htp/binary-ops.c +753 -241
  240. data/ext/sources/ggml/src/ggml-hexagon/htp/cmake-toolchain.cmake +5 -5
  241. data/ext/sources/ggml/src/ggml-hexagon/htp/concat-ops.c +277 -0
  242. data/ext/sources/ggml/src/ggml-hexagon/htp/cpy-ops.c +295 -0
  243. data/ext/sources/ggml/src/ggml-hexagon/htp/cumsum-ops.c +270 -0
  244. data/ext/sources/ggml/src/ggml-hexagon/htp/diag-ops.c +216 -0
  245. data/ext/sources/ggml/src/ggml-hexagon/htp/fill-ops.c +123 -0
  246. data/ext/sources/ggml/src/ggml-hexagon/htp/flash-attn-ops.c +471 -296
  247. data/ext/sources/ggml/src/ggml-hexagon/htp/gated-delta-net-ops.c +1148 -0
  248. data/ext/sources/ggml/src/ggml-hexagon/htp/get-rows-ops.c +159 -53
  249. data/ext/sources/ggml/src/ggml-hexagon/htp/{htp-dma.c → hex-dma.c} +3 -3
  250. data/ext/sources/ggml/src/ggml-hexagon/htp/hex-dma.h +372 -0
  251. data/ext/sources/ggml/src/ggml-hexagon/htp/hex-dump.h +86 -0
  252. data/ext/sources/ggml/src/ggml-hexagon/htp/hex-fastdiv.h +37 -0
  253. data/ext/sources/ggml/src/ggml-hexagon/htp/hex-utils.h +137 -0
  254. data/ext/sources/ggml/src/ggml-hexagon/htp/hmx-flash-attn-ops.c +1878 -0
  255. data/ext/sources/ggml/src/ggml-hexagon/htp/hmx-matmul-ops.c +2066 -0
  256. data/ext/sources/ggml/src/ggml-hexagon/htp/hmx-ops.c +6 -0
  257. data/ext/sources/ggml/src/ggml-hexagon/htp/hmx-ops.h +88 -0
  258. data/ext/sources/ggml/src/ggml-hexagon/htp/hmx-profile.h +34 -0
  259. data/ext/sources/ggml/src/ggml-hexagon/htp/hmx-queue.c +158 -0
  260. data/ext/sources/ggml/src/ggml-hexagon/htp/hmx-queue.h +134 -0
  261. data/ext/sources/ggml/src/ggml-hexagon/htp/hmx-utils.h +200 -0
  262. data/ext/sources/ggml/src/ggml-hexagon/htp/htp-ctx.h +97 -14
  263. data/ext/sources/ggml/src/ggml-hexagon/htp/htp-ops.h +163 -67
  264. data/ext/sources/ggml/src/ggml-hexagon/htp/htp_iface.idl +9 -3
  265. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-arith.h +443 -0
  266. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-base.h +308 -0
  267. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-copy.h +262 -0
  268. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-div.h +291 -0
  269. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-dump.h +129 -0
  270. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-exp.h +216 -0
  271. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-flash-attn.h +47 -0
  272. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-floor.h +100 -0
  273. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-inverse.h +210 -0
  274. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-log.h +65 -0
  275. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-pow.h +42 -0
  276. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-reduce.h +296 -0
  277. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-repl.h +74 -0
  278. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-scale.h +133 -0
  279. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-sigmoid.h +142 -0
  280. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-sin-cos.h +90 -0
  281. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-sqrt.h +126 -0
  282. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-types.h +36 -0
  283. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-utils.h +18 -1348
  284. data/ext/sources/ggml/src/ggml-hexagon/htp/main.c +547 -635
  285. data/ext/sources/ggml/src/ggml-hexagon/htp/matmul-ops.c +3556 -1101
  286. data/ext/sources/ggml/src/ggml-hexagon/htp/pad-ops.c +547 -0
  287. data/ext/sources/ggml/src/ggml-hexagon/htp/repeat-ops.c +148 -0
  288. data/ext/sources/ggml/src/ggml-hexagon/htp/rope-ops.c +475 -269
  289. data/ext/sources/ggml/src/ggml-hexagon/htp/set-rows-ops.c +94 -72
  290. data/ext/sources/ggml/src/ggml-hexagon/htp/softmax-ops.c +222 -217
  291. data/ext/sources/ggml/src/ggml-hexagon/htp/solve-tri-ops.c +267 -0
  292. data/ext/sources/ggml/src/ggml-hexagon/htp/ssm-conv.c +432 -0
  293. data/ext/sources/ggml/src/ggml-hexagon/htp/sum-rows-ops.c +128 -0
  294. data/ext/sources/ggml/src/ggml-hexagon/htp/unary-ops.c +886 -117
  295. data/ext/sources/ggml/src/ggml-hexagon/htp/vtcm-utils.h +16 -0
  296. data/ext/sources/ggml/src/ggml-hexagon/htp/worker-pool.c +1 -5
  297. data/ext/sources/ggml/src/ggml-hexagon/htp-drv.cpp +418 -0
  298. data/ext/sources/ggml/src/ggml-hexagon/htp-drv.h +121 -0
  299. data/ext/sources/ggml/src/ggml-hexagon/htp-opnode.h +272 -0
  300. data/ext/sources/ggml/src/ggml-hexagon/libdl.h +79 -0
  301. data/ext/sources/ggml/src/ggml-hexagon/libggml-htp.inf +40 -0
  302. data/ext/sources/ggml/src/ggml-hip/CMakeLists.txt +28 -9
  303. data/ext/sources/ggml/src/ggml-impl.h +68 -1
  304. data/ext/sources/ggml/src/ggml-metal/CMakeLists.txt +10 -10
  305. data/ext/sources/ggml/src/ggml-metal/ggml-metal-common.cpp +13 -2
  306. data/ext/sources/ggml/src/ggml-metal/ggml-metal-context.h +8 -0
  307. data/ext/sources/ggml/src/ggml-metal/ggml-metal-context.m +147 -17
  308. data/ext/sources/ggml/src/ggml-metal/ggml-metal-device.cpp +409 -83
  309. data/ext/sources/ggml/src/ggml-metal/ggml-metal-device.h +54 -5
  310. data/ext/sources/ggml/src/ggml-metal/ggml-metal-device.m +254 -52
  311. data/ext/sources/ggml/src/ggml-metal/ggml-metal-impl.h +254 -23
  312. data/ext/sources/ggml/src/ggml-metal/ggml-metal-ops.cpp +756 -285
  313. data/ext/sources/ggml/src/ggml-metal/ggml-metal-ops.h +7 -4
  314. data/ext/sources/ggml/src/ggml-metal/ggml-metal.cpp +359 -133
  315. data/ext/sources/ggml/src/ggml-metal/ggml-metal.metal +1867 -1123
  316. data/ext/sources/ggml/src/ggml-musa/CMakeLists.txt +5 -6
  317. data/ext/sources/ggml/src/ggml-opencl/CMakeLists.txt +71 -4
  318. data/ext/sources/ggml/src/ggml-opencl/ggml-opencl.cpp +14127 -5314
  319. data/ext/sources/ggml/src/ggml-opencl/kernels/concat.cl +97 -88
  320. data/ext/sources/ggml/src/ggml-opencl/kernels/cpy.cl +104 -0
  321. data/ext/sources/ggml/src/ggml-opencl/kernels/cumsum.cl +139 -0
  322. data/ext/sources/ggml/src/ggml-opencl/kernels/cvt.cl +1978 -67
  323. data/ext/sources/ggml/src/ggml-opencl/kernels/diag.cl +27 -0
  324. data/ext/sources/ggml/src/ggml-opencl/kernels/exp.cl +125 -0
  325. data/ext/sources/ggml/src/ggml-opencl/kernels/expm1.cl +87 -56
  326. data/ext/sources/ggml/src/ggml-opencl/kernels/gated_delta_net.cl +249 -0
  327. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_mxfp4_f32_ns.cl +306 -0
  328. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q4_0_f32_ns.cl +256 -0
  329. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q4_1_f32_ns.cl +258 -0
  330. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q4_k_f32_ns.cl +283 -0
  331. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q5_0_f32_ns.cl +260 -0
  332. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q5_1_f32_ns.cl +262 -0
  333. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q5_k_f32_ns.cl +288 -0
  334. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q6_k_f32_ns.cl +267 -0
  335. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_iq4_nl_f32.cl +150 -0
  336. data/ext/sources/ggml/src/ggml-opencl/kernels/{mul_mat_Ab_Bi_8x4.cl → gemm_noshuffle_q4_0_f32.cl} +1 -1
  337. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q4_1_f32.cl +132 -0
  338. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q4_k_f32.cl +172 -0
  339. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q5_0_f32.cl +131 -0
  340. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q5_1_f32.cl +134 -0
  341. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q5_k_f32.cl +176 -0
  342. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q6_k_f32.cl +140 -0
  343. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q8_0_f32.cl +129 -0
  344. data/ext/sources/ggml/src/ggml-opencl/kernels/gemm_xmem_f16_f32_os8.cl +233 -0
  345. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_mxfp4_f32_ns.cl +165 -0
  346. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q4_0_f32_ns.cl +120 -0
  347. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q4_1_f32_ns.cl +123 -0
  348. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q4_k_f32_ns.cl +155 -0
  349. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q5_0_f32_ns.cl +123 -0
  350. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q5_1_f32_ns.cl +125 -0
  351. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q5_k_f32_ns.cl +160 -0
  352. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q6_k_f32_ns.cl +141 -0
  353. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_iq4_nl_f32.cl +302 -0
  354. data/ext/sources/ggml/src/ggml-opencl/kernels/{gemv_noshuffle_general.cl → gemv_noshuffle_q4_0_f32.cl} +5 -5
  355. data/ext/sources/ggml/src/ggml-opencl/kernels/{gemv_noshuffle.cl → gemv_noshuffle_q4_0_f32_spec.cl} +5 -5
  356. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q4_1_f32.cl +283 -0
  357. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q4_k_f32.cl +318 -0
  358. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q5_0_f32.cl +291 -0
  359. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q5_1_f32.cl +294 -0
  360. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q5_k_f32.cl +326 -0
  361. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q6_k_f32.cl +293 -0
  362. data/ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q8_0_f32.cl +195 -0
  363. data/ext/sources/ggml/src/ggml-opencl/kernels/get_rows.cl +15 -9
  364. data/ext/sources/ggml/src/ggml-opencl/kernels/l2_norm.cl +71 -0
  365. data/ext/sources/ggml/src/ggml-opencl/kernels/mean.cl +114 -13
  366. data/ext/sources/ggml/src/ggml-opencl/kernels/moe_reorder_b.cl +30 -0
  367. data/ext/sources/ggml/src/ggml-opencl/kernels/moe_sort_by_expert.cl +82 -0
  368. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_iq4_nl_f32_l4_lm.cl +171 -0
  369. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q4_0_f32_l4_lm.cl +163 -0
  370. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q4_1_f32_l4_lm.cl +165 -0
  371. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q4_k_f32_l4_lm.cl +179 -0
  372. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q5_0_f32_l4_lm.cl +173 -0
  373. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q5_1_f32_l4_lm.cl +175 -0
  374. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q5_k_f32_l4_lm.cl +192 -0
  375. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q6_k_f32_l4_lm.cl +158 -0
  376. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_iq4_nl_f32.cl +164 -0
  377. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_iq4_nl_f32_flat.cl +202 -0
  378. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_1_f32.cl +219 -0
  379. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_1_f32_flat.cl +229 -0
  380. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_k_f32.cl +180 -0
  381. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_k_f32_flat.cl +196 -0
  382. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_0_f32.cl +241 -0
  383. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_0_f32_flat.cl +243 -0
  384. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_1_f32.cl +243 -0
  385. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_1_f32_flat.cl +247 -0
  386. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_k_f32.cl +187 -0
  387. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_k_f32_flat.cl +203 -0
  388. data/ext/sources/ggml/src/ggml-opencl/kernels/{mul_mv_q6_k.cl → mul_mv_q6_k_f32.cl} +4 -0
  389. data/ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q6_k_f32_flat.cl +178 -0
  390. data/ext/sources/ggml/src/ggml-opencl/kernels/neg.cl +125 -0
  391. data/ext/sources/ggml/src/ggml-opencl/kernels/repeat.cl +31 -32
  392. data/ext/sources/ggml/src/ggml-opencl/kernels/scale.cl +14 -4
  393. data/ext/sources/ggml/src/ggml-opencl/kernels/softplus.cl +88 -60
  394. data/ext/sources/ggml/src/ggml-opencl/kernels/solve_tri.cl +51 -0
  395. data/ext/sources/ggml/src/ggml-opencl/kernels/sum_rows.cl +114 -13
  396. data/ext/sources/ggml/src/ggml-opencl/kernels/tanh.cl +94 -48
  397. data/ext/sources/ggml/src/ggml-opencl/kernels/transpose.cl +26 -0
  398. data/ext/sources/ggml/src/ggml-opencl/kernels/tri.cl +32 -0
  399. data/ext/sources/ggml/src/ggml-openvino/.clang-format +154 -0
  400. data/ext/sources/ggml/src/ggml-openvino/CMakeLists.txt +22 -0
  401. data/ext/sources/ggml/src/ggml-openvino/ggml-decoder.cpp +985 -0
  402. data/ext/sources/ggml/src/ggml-openvino/ggml-decoder.h +294 -0
  403. data/ext/sources/ggml/src/ggml-openvino/ggml-openvino-extra.cpp +380 -0
  404. data/ext/sources/ggml/src/ggml-openvino/ggml-openvino-extra.h +182 -0
  405. data/ext/sources/ggml/src/ggml-openvino/ggml-openvino.cpp +1132 -0
  406. data/ext/sources/ggml/src/ggml-openvino/ggml-quants.cpp +956 -0
  407. data/ext/sources/ggml/src/ggml-openvino/ggml-quants.h +153 -0
  408. data/ext/sources/ggml/src/ggml-openvino/openvino/decoder.h +74 -0
  409. data/ext/sources/ggml/src/ggml-openvino/openvino/frontend.cpp +27 -0
  410. data/ext/sources/ggml/src/ggml-openvino/openvino/frontend.h +23 -0
  411. data/ext/sources/ggml/src/ggml-openvino/openvino/input_model.cpp +17 -0
  412. data/ext/sources/ggml/src/ggml-openvino/openvino/input_model.h +29 -0
  413. data/ext/sources/ggml/src/ggml-openvino/openvino/node_context.h +112 -0
  414. data/ext/sources/ggml/src/ggml-openvino/openvino/op/cont.cpp +48 -0
  415. data/ext/sources/ggml/src/ggml-openvino/openvino/op/cpy.cpp +21 -0
  416. data/ext/sources/ggml/src/ggml-openvino/openvino/op/flash_attn_ext.cpp +90 -0
  417. data/ext/sources/ggml/src/ggml-openvino/openvino/op/get_rows.cpp +69 -0
  418. data/ext/sources/ggml/src/ggml-openvino/openvino/op/glu_geglu.cpp +61 -0
  419. data/ext/sources/ggml/src/ggml-openvino/openvino/op/glu_swiglu.cpp +62 -0
  420. data/ext/sources/ggml/src/ggml-openvino/openvino/op/mulmat.cpp +90 -0
  421. data/ext/sources/ggml/src/ggml-openvino/openvino/op/permute.cpp +102 -0
  422. data/ext/sources/ggml/src/ggml-openvino/openvino/op/reshape.cpp +83 -0
  423. data/ext/sources/ggml/src/ggml-openvino/openvino/op/rms_norm.cpp +46 -0
  424. data/ext/sources/ggml/src/ggml-openvino/openvino/op/rope.cpp +149 -0
  425. data/ext/sources/ggml/src/ggml-openvino/openvino/op/scale.cpp +41 -0
  426. data/ext/sources/ggml/src/ggml-openvino/openvino/op/set_rows.cpp +76 -0
  427. data/ext/sources/ggml/src/ggml-openvino/openvino/op/softmax.cpp +89 -0
  428. data/ext/sources/ggml/src/ggml-openvino/openvino/op/transpose.cpp +23 -0
  429. data/ext/sources/ggml/src/ggml-openvino/openvino/op/unary_gelu.cpp +25 -0
  430. data/ext/sources/ggml/src/ggml-openvino/openvino/op/unary_silu.cpp +27 -0
  431. data/ext/sources/ggml/src/ggml-openvino/openvino/op/view.cpp +53 -0
  432. data/ext/sources/ggml/src/ggml-openvino/openvino/op_table.cpp +47 -0
  433. data/ext/sources/ggml/src/ggml-openvino/openvino/op_table.h +40 -0
  434. data/ext/sources/ggml/src/ggml-openvino/openvino/pass/fuse_to_sdpa.cpp +60 -0
  435. data/ext/sources/ggml/src/ggml-openvino/openvino/pass/fuse_to_sdpa.h +17 -0
  436. data/ext/sources/ggml/src/ggml-openvino/openvino/pass/mark_decompression_convert_constant_folding.h +29 -0
  437. data/ext/sources/ggml/src/ggml-openvino/openvino/pass/squeeze_matmul.cpp +58 -0
  438. data/ext/sources/ggml/src/ggml-openvino/openvino/pass/squeeze_matmul.h +17 -0
  439. data/ext/sources/ggml/src/ggml-openvino/openvino/rt_info/weightless_caching_attributes.hpp +41 -0
  440. data/ext/sources/ggml/src/ggml-openvino/openvino/translate_session.cpp +317 -0
  441. data/ext/sources/ggml/src/ggml-openvino/openvino/translate_session.h +28 -0
  442. data/ext/sources/ggml/src/ggml-openvino/openvino/utils.cpp +257 -0
  443. data/ext/sources/ggml/src/ggml-openvino/openvino/utils.h +86 -0
  444. data/ext/sources/ggml/src/ggml-openvino/utils.cpp +880 -0
  445. data/ext/sources/ggml/src/ggml-openvino/utils.h +143 -0
  446. data/ext/sources/ggml/src/ggml-opt.cpp +1 -0
  447. data/ext/sources/ggml/src/ggml-quants.c +385 -119
  448. data/ext/sources/ggml/src/ggml-quants.h +6 -0
  449. data/ext/sources/ggml/src/ggml-rpc/CMakeLists.txt +24 -0
  450. data/ext/sources/ggml/src/ggml-rpc/ggml-rpc.cpp +167 -311
  451. data/ext/sources/ggml/src/ggml-rpc/transport.cpp +683 -0
  452. data/ext/sources/ggml/src/ggml-rpc/transport.h +34 -0
  453. data/ext/sources/ggml/src/ggml-sycl/CMakeLists.txt +64 -91
  454. data/ext/sources/ggml/src/ggml-sycl/add-id.cpp +5 -1
  455. data/ext/sources/ggml/src/ggml-sycl/backend.hpp +4 -1
  456. data/ext/sources/ggml/src/ggml-sycl/binbcast.cpp +21 -20
  457. data/ext/sources/ggml/src/ggml-sycl/common.cpp +74 -2
  458. data/ext/sources/ggml/src/ggml-sycl/common.hpp +356 -11
  459. data/ext/sources/ggml/src/ggml-sycl/convert.cpp +184 -14
  460. data/ext/sources/ggml/src/ggml-sycl/convert.hpp +31 -1
  461. data/ext/sources/ggml/src/ggml-sycl/count-equal.cpp +1 -1
  462. data/ext/sources/ggml/src/ggml-sycl/cumsum.cpp +148 -0
  463. data/ext/sources/ggml/src/ggml-sycl/cumsum.hpp +5 -0
  464. data/ext/sources/ggml/src/ggml-sycl/dequantize.hpp +663 -0
  465. data/ext/sources/ggml/src/ggml-sycl/diag.cpp +67 -0
  466. data/ext/sources/ggml/src/ggml-sycl/diag.hpp +5 -0
  467. data/ext/sources/ggml/src/ggml-sycl/dmmv.cpp +586 -6
  468. data/ext/sources/ggml/src/ggml-sycl/dpct/helper.hpp +791 -47
  469. data/ext/sources/ggml/src/ggml-sycl/element_wise.cpp +77 -156
  470. data/ext/sources/ggml/src/ggml-sycl/element_wise.hpp +2 -2
  471. data/ext/sources/ggml/src/ggml-sycl/fattn-buffers.cpp +56 -0
  472. data/ext/sources/ggml/src/ggml-sycl/fattn-buffers.hpp +63 -0
  473. data/ext/sources/ggml/src/ggml-sycl/fattn-common.hpp +1181 -0
  474. data/ext/sources/ggml/src/ggml-sycl/fattn-tile.cpp +59 -0
  475. data/ext/sources/ggml/src/ggml-sycl/fattn-tile.hpp +1246 -0
  476. data/ext/sources/ggml/src/ggml-sycl/fattn-vec.hpp +674 -0
  477. data/ext/sources/ggml/src/ggml-sycl/fattn.cpp +227 -0
  478. data/ext/sources/ggml/src/ggml-sycl/fattn.hpp +22 -0
  479. data/ext/sources/ggml/src/ggml-sycl/fill.cpp +55 -0
  480. data/ext/sources/ggml/src/ggml-sycl/fill.hpp +5 -0
  481. data/ext/sources/ggml/src/ggml-sycl/gated_delta_net.cpp +347 -0
  482. data/ext/sources/ggml/src/ggml-sycl/gated_delta_net.hpp +9 -0
  483. data/ext/sources/ggml/src/ggml-sycl/gemm.hpp +3 -0
  484. data/ext/sources/ggml/src/ggml-sycl/getrows.cpp +79 -3
  485. data/ext/sources/ggml/src/ggml-sycl/ggml-sycl.cpp +1134 -236
  486. data/ext/sources/ggml/src/ggml-sycl/im2col.cpp +353 -89
  487. data/ext/sources/ggml/src/ggml-sycl/im2col.hpp +5 -3
  488. data/ext/sources/ggml/src/ggml-sycl/mmvq.cpp +1344 -26
  489. data/ext/sources/ggml/src/ggml-sycl/mmvq.hpp +16 -0
  490. data/ext/sources/ggml/src/ggml-sycl/norm.cpp +65 -66
  491. data/ext/sources/ggml/src/ggml-sycl/outprod.cpp +3 -3
  492. data/ext/sources/ggml/src/ggml-sycl/pad.cpp +27 -27
  493. data/ext/sources/ggml/src/ggml-sycl/presets.hpp +3 -0
  494. data/ext/sources/ggml/src/ggml-sycl/quants.hpp +72 -1
  495. data/ext/sources/ggml/src/ggml-sycl/rope.cpp +450 -287
  496. data/ext/sources/ggml/src/ggml-sycl/rope.hpp +6 -0
  497. data/ext/sources/ggml/src/ggml-sycl/set_rows.cpp +7 -1
  498. data/ext/sources/ggml/src/ggml-sycl/softmax.cpp +6 -6
  499. data/ext/sources/ggml/src/ggml-sycl/solve_tri.cpp +172 -0
  500. data/ext/sources/ggml/src/ggml-sycl/solve_tri.hpp +8 -0
  501. data/ext/sources/ggml/src/ggml-sycl/ssm_conv.cpp +6 -1
  502. data/ext/sources/ggml/src/ggml-sycl/ssm_scan.cpp +156 -0
  503. data/ext/sources/ggml/src/ggml-sycl/ssm_scan.hpp +5 -0
  504. data/ext/sources/ggml/src/ggml-sycl/sycl_hw.cpp +62 -10
  505. data/ext/sources/ggml/src/ggml-sycl/sycl_hw.hpp +18 -6
  506. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq112-dv112.cpp +5 -0
  507. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq128-dv128.cpp +5 -0
  508. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq256-dv256.cpp +5 -0
  509. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq40-dv40.cpp +5 -0
  510. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq512-dv512.cpp +6 -0
  511. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq576-dv512.cpp +5 -0
  512. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq64-dv64.cpp +5 -0
  513. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq72-dv72.cpp +5 -0
  514. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq80-dv80.cpp +5 -0
  515. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq96-dv96.cpp +5 -0
  516. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-f16.cpp +8 -0
  517. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q4_0.cpp +8 -0
  518. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q4_1.cpp +8 -0
  519. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q5_0.cpp +8 -0
  520. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q5_1.cpp +8 -0
  521. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q8_0.cpp +8 -0
  522. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-f16.cpp +8 -0
  523. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q4_0.cpp +8 -0
  524. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q4_1.cpp +8 -0
  525. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q5_0.cpp +8 -0
  526. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q5_1.cpp +8 -0
  527. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q8_0.cpp +8 -0
  528. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-f16.cpp +8 -0
  529. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q4_0.cpp +8 -0
  530. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q4_1.cpp +8 -0
  531. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q5_0.cpp +8 -0
  532. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q5_1.cpp +8 -0
  533. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q8_0.cpp +8 -0
  534. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-f16.cpp +8 -0
  535. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q4_0.cpp +8 -0
  536. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q4_1.cpp +8 -0
  537. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q5_0.cpp +8 -0
  538. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q5_1.cpp +8 -0
  539. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q8_0.cpp +8 -0
  540. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-f16.cpp +8 -0
  541. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q4_0.cpp +8 -0
  542. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q4_1.cpp +8 -0
  543. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q5_0.cpp +8 -0
  544. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q5_1.cpp +8 -0
  545. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q8_0.cpp +8 -0
  546. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-f16.cpp +8 -0
  547. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q4_0.cpp +8 -0
  548. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q4_1.cpp +8 -0
  549. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q5_0.cpp +8 -0
  550. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q5_1.cpp +8 -0
  551. data/ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q8_0.cpp +8 -0
  552. data/ext/sources/ggml/src/ggml-sycl/type.hpp +112 -0
  553. data/ext/sources/ggml/src/ggml-sycl/upscale.cpp +410 -0
  554. data/ext/sources/ggml/src/ggml-sycl/upscale.hpp +9 -0
  555. data/ext/sources/ggml/src/ggml-sycl/vecdotq.hpp +228 -53
  556. data/ext/sources/ggml/src/ggml-sycl/wkv.cpp +1 -1
  557. data/ext/sources/ggml/src/ggml-virtgpu/CMakeLists.txt +70 -0
  558. data/ext/sources/ggml/src/ggml-virtgpu/apir_cs_ggml-rpc-front.cpp +87 -0
  559. data/ext/sources/ggml/src/ggml-virtgpu/backend/CMakeLists.txt +21 -0
  560. data/ext/sources/ggml/src/ggml-virtgpu/backend/apir_cs_ggml-rpc-back.cpp +115 -0
  561. data/ext/sources/ggml/src/ggml-virtgpu/backend/backend-convert.h +13 -0
  562. data/ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched-backend.cpp +102 -0
  563. data/ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched-buffer-type.cpp +105 -0
  564. data/ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched-buffer.cpp +179 -0
  565. data/ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched-device.cpp +148 -0
  566. data/ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched.cpp +51 -0
  567. data/ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched.gen.h +73 -0
  568. data/ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched.h +27 -0
  569. data/ext/sources/ggml/src/ggml-virtgpu/backend/backend-virgl-apir.h +32 -0
  570. data/ext/sources/ggml/src/ggml-virtgpu/backend/backend.cpp +144 -0
  571. data/ext/sources/ggml/src/ggml-virtgpu/backend/shared/api_remoting.h +95 -0
  572. data/ext/sources/ggml/src/ggml-virtgpu/backend/shared/apir_backend.gen.h +94 -0
  573. data/ext/sources/ggml/src/ggml-virtgpu/backend/shared/apir_backend.h +50 -0
  574. data/ext/sources/ggml/src/ggml-virtgpu/backend/shared/apir_cs.h +378 -0
  575. data/ext/sources/ggml/src/ggml-virtgpu/backend/shared/apir_cs_ggml.h +232 -0
  576. data/ext/sources/ggml/src/ggml-virtgpu/backend/shared/apir_cs_rpc.h +58 -0
  577. data/ext/sources/ggml/src/ggml-virtgpu/ggml-backend-buffer-type.cpp +81 -0
  578. data/ext/sources/ggml/src/ggml-virtgpu/ggml-backend-buffer.cpp +123 -0
  579. data/ext/sources/ggml/src/ggml-virtgpu/ggml-backend-device.cpp +160 -0
  580. data/ext/sources/ggml/src/ggml-virtgpu/ggml-backend-reg.cpp +213 -0
  581. data/ext/sources/ggml/src/ggml-virtgpu/ggml-backend.cpp +71 -0
  582. data/ext/sources/ggml/src/ggml-virtgpu/ggml-remoting.h +71 -0
  583. data/ext/sources/ggml/src/ggml-virtgpu/ggmlremoting_functions.yaml +166 -0
  584. data/ext/sources/ggml/src/ggml-virtgpu/include/apir_hw.h +9 -0
  585. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu-apir.h +15 -0
  586. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward-backend.cpp +58 -0
  587. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward-buffer-type.cpp +110 -0
  588. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward-buffer.cpp +173 -0
  589. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward-device.cpp +192 -0
  590. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward-impl.h +36 -0
  591. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward.gen.h +53 -0
  592. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu-shm.cpp +99 -0
  593. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu-shm.h +23 -0
  594. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu-utils.cpp +179 -0
  595. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu-utils.h +86 -0
  596. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu.cpp +545 -0
  597. data/ext/sources/ggml/src/ggml-virtgpu/virtgpu.h +115 -0
  598. data/ext/sources/ggml/src/ggml-vulkan/CMakeLists.txt +12 -1
  599. data/ext/sources/ggml/src/ggml-vulkan/ggml-vulkan.cpp +3250 -940
  600. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt +4 -0
  601. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/acc.comp +16 -8
  602. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/contig_copy.comp +6 -2
  603. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_mm.comp +146 -13
  604. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/copy.comp +3 -1
  605. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/copy_from_quant.comp +1 -1
  606. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/copy_to_quant.comp +25 -1
  607. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.glsl +88 -0
  608. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.glsl +643 -1
  609. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_nvfp4.comp +32 -0
  610. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q1_0.comp +29 -0
  611. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/diag.comp +0 -1
  612. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dot_product_funcs.glsl +27 -0
  613. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/elu.comp +27 -0
  614. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/exp.comp +0 -1
  615. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/coopmat2_decode_vector.comp +7 -0
  616. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp +533 -180
  617. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_base.glsl +113 -68
  618. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp +412 -222
  619. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp +222 -83
  620. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_dequant.glsl +131 -0
  621. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_mask_opt.comp +162 -0
  622. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_mmq_funcs.glsl +203 -0
  623. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_split_k_reduce.comp +9 -8
  624. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/fwht.comp +115 -0
  625. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/gated_delta_net.comp +189 -0
  626. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.glsl +0 -1
  627. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/glu_head.glsl +10 -1
  628. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/glu_main.glsl +16 -6
  629. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/im2col.comp +76 -54
  630. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/im2col_3d.comp +0 -1
  631. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/l2_norm.comp +12 -9
  632. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/log.comp +0 -1
  633. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp +122 -27
  634. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.glsl +20 -17
  635. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iface.glsl +6 -6
  636. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp +1 -1
  637. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp +1 -1
  638. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp +1 -1
  639. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vecq.comp +1 -0
  640. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vecq_funcs.glsl +88 -55
  641. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp +22 -20
  642. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp +51 -14
  643. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_funcs.glsl +159 -125
  644. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_id_funcs.glsl +3 -1
  645. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq.comp +5 -3
  646. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.glsl +8 -8
  647. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_shmem_types.glsl +24 -9
  648. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/multi_add.comp +0 -1
  649. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp +2 -3
  650. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_funcs.glsl +39 -63
  651. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.glsl +0 -1
  652. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_multi.comp +7 -4
  653. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_neox.comp +7 -4
  654. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_norm.comp +7 -4
  655. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_params.glsl +13 -7
  656. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_vision.comp +7 -4
  657. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sgn.comp +21 -0
  658. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/snake.comp +49 -0
  659. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/ssm_conv.comp +27 -11
  660. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/tri.comp +0 -1
  661. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/types.glsl +79 -2
  662. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp +193 -149
  663. data/ext/sources/ggml/src/ggml-webgpu/CMakeLists.txt +5 -2
  664. data/ext/sources/ggml/src/ggml-webgpu/ggml-webgpu-shader-lib.hpp +3221 -97
  665. data/ext/sources/ggml/src/ggml-webgpu/ggml-webgpu.cpp +3493 -1997
  666. data/ext/sources/ggml/src/ggml-webgpu/pre_wgsl.hpp +37 -7
  667. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/add_id.wgsl +64 -0
  668. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/argmax.wgsl +72 -0
  669. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/argsort.wgsl +106 -0
  670. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/argsort_merge.wgsl +134 -0
  671. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/binary.wgsl +142 -0
  672. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/common_decls.tmpl +115 -141
  673. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/concat.wgsl +93 -0
  674. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/conv2d.wgsl +165 -0
  675. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/{cpy.tmpl.wgsl → cpy.wgsl} +25 -44
  676. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/cumsum.wgsl +66 -0
  677. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn.wgsl +198 -230
  678. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_quant_staging.tmpl +124 -0
  679. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_tile.wgsl +397 -0
  680. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_vec_blk.wgsl +101 -0
  681. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_vec_reduce.wgsl +84 -0
  682. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_vec_split.wgsl +619 -0
  683. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/gated_delta_net.wgsl +149 -0
  684. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/{get_rows.tmpl.wgsl → get_rows.wgsl} +234 -335
  685. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/glu.wgsl +155 -0
  686. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/im2col.wgsl +101 -0
  687. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_decls.tmpl +871 -42
  688. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_id.wgsl +195 -0
  689. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_id_gather.wgsl +52 -0
  690. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_id_vec.wgsl +154 -0
  691. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_reg_tile.wgsl +149 -0
  692. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/{mul_mat_subgroup_matrix.tmpl.wgsl → mul_mat_subgroup_matrix.wgsl} +36 -138
  693. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_vec.wgsl +151 -0
  694. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_vec_acc.tmpl +1432 -0
  695. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_vec_q_acc.tmpl +303 -0
  696. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/pad.wgsl +86 -0
  697. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/quant_inner_loops.tmpl +21 -0
  698. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/quantize_q8.wgsl +173 -0
  699. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/repeat.wgsl +67 -0
  700. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/rms_norm_mul.wgsl +152 -0
  701. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/{rope.tmpl.wgsl → rope.wgsl} +71 -142
  702. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/row_norm.wgsl +153 -0
  703. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/{scale.tmpl.wgsl → scale.wgsl} +15 -40
  704. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/set.wgsl +109 -0
  705. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/set_rows.wgsl +39 -12
  706. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/set_rows_quant.wgsl +224 -0
  707. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/{soft_max.tmpl.wgsl → soft_max.wgsl} +106 -206
  708. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/solve_tri.wgsl +121 -0
  709. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/ssm_conv.wgsl +65 -0
  710. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/ssm_scan.wgsl +193 -0
  711. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/sum_rows.wgsl +55 -0
  712. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/unary.wgsl +213 -0
  713. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/upscale.wgsl +240 -0
  714. data/ext/sources/ggml/src/ggml-zdnn/ggml-zdnn.cpp +24 -15
  715. data/ext/sources/ggml/src/ggml-zendnn/CMakeLists.txt +31 -32
  716. data/ext/sources/ggml/src/ggml-zendnn/ggml-zendnn.cpp +253 -16
  717. data/ext/sources/ggml/src/ggml.c +268 -52
  718. data/ext/sources/ggml/src/gguf.cpp +377 -47
  719. data/ext/sources/include/parakeet.h +342 -0
  720. data/ext/sources/include/whisper.h +10 -0
  721. data/ext/sources/media/matmul.png +0 -0
  722. data/ext/sources/src/CMakeLists.txt +23 -0
  723. data/ext/sources/src/parakeet-arch.h +188 -0
  724. data/ext/sources/src/parakeet.cpp +3838 -0
  725. data/ext/sources/src/whisper.cpp +62 -40
  726. data/extsources.rb +26 -10
  727. data/lib/whisper/log_settable.rb +36 -0
  728. data/lib/whisper/model/uri.rb +13 -1
  729. data/lib/whisper/output.rb +74 -0
  730. data/sig/whisper.rbs +445 -55
  731. data/test/helper.rb +2 -0
  732. data/test/jfk_reader/jfk_reader.c +50 -7
  733. data/test/test_callback.rb +1 -0
  734. data/test/test_context_params.rb +82 -0
  735. data/test/test_package.rb +6 -5
  736. data/test/test_parakeet.rb +28 -0
  737. data/test/test_parakeet_callback.rb +107 -0
  738. data/test/test_parakeet_context.rb +116 -0
  739. data/test/test_parakeet_context_params.rb +24 -0
  740. data/test/test_parakeet_model.rb +21 -0
  741. data/test/test_parakeet_params.rb +78 -0
  742. data/test/test_parakeet_segment.rb +42 -0
  743. data/test/test_parakeet_token.rb +73 -0
  744. data/test/test_params.rb +2 -0
  745. data/test/test_token.rb +11 -0
  746. data/test/test_vad_context.rb +58 -8
  747. data/test/test_vad_segment.rb +1 -1
  748. data/test/test_whisper.rb +44 -6
  749. data/whispercpp.gemspec +2 -2
  750. metadata +426 -280
  751. data/ext/sources/bindings/javascript/CMakeLists.txt +0 -41
  752. data/ext/sources/bindings/javascript/emscripten.cpp +0 -93
  753. data/ext/sources/bindings/javascript/libwhisper.worker.js +0 -1
  754. data/ext/sources/bindings/javascript/package.json +0 -26
  755. data/ext/sources/bindings/javascript/whisper.js +0 -19
  756. data/ext/sources/examples/addon.node/CMakeLists.txt +0 -31
  757. data/ext/sources/examples/addon.node/__test__/whisper.spec.js +0 -133
  758. data/ext/sources/examples/addon.node/addon.cpp +0 -557
  759. data/ext/sources/examples/addon.node/index.js +0 -59
  760. data/ext/sources/examples/addon.node/package.json +0 -16
  761. data/ext/sources/examples/addon.node/vad-example.js +0 -132
  762. data/ext/sources/examples/bench.wasm/CMakeLists.txt +0 -49
  763. data/ext/sources/examples/bench.wasm/emscripten.cpp +0 -87
  764. data/ext/sources/examples/bench.wasm/index-tmpl.html +0 -285
  765. data/ext/sources/examples/coi-serviceworker.js +0 -146
  766. data/ext/sources/examples/command/CMakeLists.txt +0 -10
  767. data/ext/sources/examples/command/command.cpp +0 -802
  768. data/ext/sources/examples/command/commands.txt +0 -9
  769. data/ext/sources/examples/command.wasm/CMakeLists.txt +0 -50
  770. data/ext/sources/examples/command.wasm/emscripten.cpp +0 -327
  771. data/ext/sources/examples/command.wasm/index-tmpl.html +0 -415
  772. data/ext/sources/examples/generate-karaoke.sh +0 -57
  773. data/ext/sources/examples/helpers.js +0 -191
  774. data/ext/sources/examples/livestream.sh +0 -112
  775. data/ext/sources/examples/lsp/CMakeLists.txt +0 -10
  776. data/ext/sources/examples/lsp/lsp.cpp +0 -471
  777. data/ext/sources/examples/lsp/whisper.vim +0 -362
  778. data/ext/sources/examples/python/test_whisper_processor.py +0 -7
  779. data/ext/sources/examples/python/whisper_processor.py +0 -54
  780. data/ext/sources/examples/server/bench.js +0 -29
  781. data/ext/sources/examples/server.py +0 -120
  782. data/ext/sources/examples/stream/CMakeLists.txt +0 -10
  783. data/ext/sources/examples/stream/stream.cpp +0 -437
  784. data/ext/sources/examples/stream.wasm/CMakeLists.txt +0 -49
  785. data/ext/sources/examples/stream.wasm/emscripten.cpp +0 -216
  786. data/ext/sources/examples/stream.wasm/index-tmpl.html +0 -491
  787. data/ext/sources/examples/sycl/CMakeLists.txt +0 -9
  788. data/ext/sources/examples/sycl/build.sh +0 -22
  789. data/ext/sources/examples/sycl/ls-sycl-device.cpp +0 -11
  790. data/ext/sources/examples/sycl/run-whisper.sh +0 -17
  791. data/ext/sources/examples/talk-llama/CMakeLists.txt +0 -47
  792. data/ext/sources/examples/talk-llama/eleven-labs.py +0 -80
  793. data/ext/sources/examples/talk-llama/llama-adapter.cpp +0 -494
  794. data/ext/sources/examples/talk-llama/llama-adapter.h +0 -88
  795. data/ext/sources/examples/talk-llama/llama-arch.cpp +0 -2559
  796. data/ext/sources/examples/talk-llama/llama-arch.h +0 -586
  797. data/ext/sources/examples/talk-llama/llama-batch.cpp +0 -917
  798. data/ext/sources/examples/talk-llama/llama-batch.h +0 -173
  799. data/ext/sources/examples/talk-llama/llama-chat.cpp +0 -876
  800. data/ext/sources/examples/talk-llama/llama-chat.h +0 -70
  801. data/ext/sources/examples/talk-llama/llama-context.cpp +0 -3645
  802. data/ext/sources/examples/talk-llama/llama-context.h +0 -360
  803. data/ext/sources/examples/talk-llama/llama-cparams.cpp +0 -5
  804. data/ext/sources/examples/talk-llama/llama-cparams.h +0 -42
  805. data/ext/sources/examples/talk-llama/llama-grammar.cpp +0 -1464
  806. data/ext/sources/examples/talk-llama/llama-grammar.h +0 -194
  807. data/ext/sources/examples/talk-llama/llama-graph.cpp +0 -2282
  808. data/ext/sources/examples/talk-llama/llama-graph.h +0 -910
  809. data/ext/sources/examples/talk-llama/llama-hparams.cpp +0 -241
  810. data/ext/sources/examples/talk-llama/llama-hparams.h +0 -284
  811. data/ext/sources/examples/talk-llama/llama-impl.cpp +0 -171
  812. data/ext/sources/examples/talk-llama/llama-impl.h +0 -63
  813. data/ext/sources/examples/talk-llama/llama-io.cpp +0 -15
  814. data/ext/sources/examples/talk-llama/llama-io.h +0 -35
  815. data/ext/sources/examples/talk-llama/llama-kv-cache-iswa.cpp +0 -328
  816. data/ext/sources/examples/talk-llama/llama-kv-cache-iswa.h +0 -137
  817. data/ext/sources/examples/talk-llama/llama-kv-cache.cpp +0 -2100
  818. data/ext/sources/examples/talk-llama/llama-kv-cache.h +0 -390
  819. data/ext/sources/examples/talk-llama/llama-kv-cells.h +0 -533
  820. data/ext/sources/examples/talk-llama/llama-memory-hybrid.cpp +0 -268
  821. data/ext/sources/examples/talk-llama/llama-memory-hybrid.h +0 -139
  822. data/ext/sources/examples/talk-llama/llama-memory-recurrent.cpp +0 -1167
  823. data/ext/sources/examples/talk-llama/llama-memory-recurrent.h +0 -182
  824. data/ext/sources/examples/talk-llama/llama-memory.cpp +0 -59
  825. data/ext/sources/examples/talk-llama/llama-memory.h +0 -122
  826. data/ext/sources/examples/talk-llama/llama-mmap.cpp +0 -735
  827. data/ext/sources/examples/talk-llama/llama-mmap.h +0 -73
  828. data/ext/sources/examples/talk-llama/llama-model-loader.cpp +0 -1247
  829. data/ext/sources/examples/talk-llama/llama-model-loader.h +0 -176
  830. data/ext/sources/examples/talk-llama/llama-model-saver.cpp +0 -285
  831. data/ext/sources/examples/talk-llama/llama-model-saver.h +0 -37
  832. data/ext/sources/examples/talk-llama/llama-model.cpp +0 -8338
  833. data/ext/sources/examples/talk-llama/llama-model.h +0 -544
  834. data/ext/sources/examples/talk-llama/llama-quant.cpp +0 -1072
  835. data/ext/sources/examples/talk-llama/llama-quant.h +0 -1
  836. data/ext/sources/examples/talk-llama/llama-sampling.cpp +0 -3771
  837. data/ext/sources/examples/talk-llama/llama-sampling.h +0 -44
  838. data/ext/sources/examples/talk-llama/llama-vocab.cpp +0 -3900
  839. data/ext/sources/examples/talk-llama/llama-vocab.h +0 -182
  840. data/ext/sources/examples/talk-llama/llama.cpp +0 -1140
  841. data/ext/sources/examples/talk-llama/llama.h +0 -1540
  842. data/ext/sources/examples/talk-llama/models/afmoe.cpp +0 -191
  843. data/ext/sources/examples/talk-llama/models/apertus.cpp +0 -125
  844. data/ext/sources/examples/talk-llama/models/arcee.cpp +0 -135
  845. data/ext/sources/examples/talk-llama/models/arctic.cpp +0 -138
  846. data/ext/sources/examples/talk-llama/models/arwkv7.cpp +0 -86
  847. data/ext/sources/examples/talk-llama/models/baichuan.cpp +0 -122
  848. data/ext/sources/examples/talk-llama/models/bailingmoe.cpp +0 -144
  849. data/ext/sources/examples/talk-llama/models/bailingmoe2.cpp +0 -135
  850. data/ext/sources/examples/talk-llama/models/bert.cpp +0 -178
  851. data/ext/sources/examples/talk-llama/models/bitnet.cpp +0 -160
  852. data/ext/sources/examples/talk-llama/models/bloom.cpp +0 -101
  853. data/ext/sources/examples/talk-llama/models/chameleon.cpp +0 -178
  854. data/ext/sources/examples/talk-llama/models/chatglm.cpp +0 -132
  855. data/ext/sources/examples/talk-llama/models/codeshell.cpp +0 -111
  856. data/ext/sources/examples/talk-llama/models/cogvlm.cpp +0 -102
  857. data/ext/sources/examples/talk-llama/models/cohere2-iswa.cpp +0 -134
  858. data/ext/sources/examples/talk-llama/models/command-r.cpp +0 -122
  859. data/ext/sources/examples/talk-llama/models/dbrx.cpp +0 -123
  860. data/ext/sources/examples/talk-llama/models/deci.cpp +0 -135
  861. data/ext/sources/examples/talk-llama/models/deepseek.cpp +0 -144
  862. data/ext/sources/examples/talk-llama/models/deepseek2.cpp +0 -259
  863. data/ext/sources/examples/talk-llama/models/dots1.cpp +0 -134
  864. data/ext/sources/examples/talk-llama/models/dream.cpp +0 -105
  865. data/ext/sources/examples/talk-llama/models/ernie4-5-moe.cpp +0 -150
  866. data/ext/sources/examples/talk-llama/models/ernie4-5.cpp +0 -110
  867. data/ext/sources/examples/talk-llama/models/exaone.cpp +0 -114
  868. data/ext/sources/examples/talk-llama/models/exaone4.cpp +0 -123
  869. data/ext/sources/examples/talk-llama/models/falcon-h1.cpp +0 -113
  870. data/ext/sources/examples/talk-llama/models/falcon.cpp +0 -120
  871. data/ext/sources/examples/talk-llama/models/gemma-embedding.cpp +0 -116
  872. data/ext/sources/examples/talk-llama/models/gemma.cpp +0 -112
  873. data/ext/sources/examples/talk-llama/models/gemma2-iswa.cpp +0 -128
  874. data/ext/sources/examples/talk-llama/models/gemma3.cpp +0 -155
  875. data/ext/sources/examples/talk-llama/models/gemma3n-iswa.cpp +0 -384
  876. data/ext/sources/examples/talk-llama/models/glm4-moe.cpp +0 -170
  877. data/ext/sources/examples/talk-llama/models/glm4.cpp +0 -150
  878. data/ext/sources/examples/talk-llama/models/gpt2.cpp +0 -105
  879. data/ext/sources/examples/talk-llama/models/gptneox.cpp +0 -144
  880. data/ext/sources/examples/talk-llama/models/granite-hybrid.cpp +0 -196
  881. data/ext/sources/examples/talk-llama/models/granite.cpp +0 -211
  882. data/ext/sources/examples/talk-llama/models/graph-context-mamba.cpp +0 -283
  883. data/ext/sources/examples/talk-llama/models/grok.cpp +0 -159
  884. data/ext/sources/examples/talk-llama/models/grovemoe.cpp +0 -141
  885. data/ext/sources/examples/talk-llama/models/hunyuan-dense.cpp +0 -132
  886. data/ext/sources/examples/talk-llama/models/hunyuan-moe.cpp +0 -154
  887. data/ext/sources/examples/talk-llama/models/internlm2.cpp +0 -120
  888. data/ext/sources/examples/talk-llama/models/jais.cpp +0 -86
  889. data/ext/sources/examples/talk-llama/models/jamba.cpp +0 -106
  890. data/ext/sources/examples/talk-llama/models/lfm2.cpp +0 -175
  891. data/ext/sources/examples/talk-llama/models/llada-moe.cpp +0 -122
  892. data/ext/sources/examples/talk-llama/models/llada.cpp +0 -99
  893. data/ext/sources/examples/talk-llama/models/llama-iswa.cpp +0 -178
  894. data/ext/sources/examples/talk-llama/models/llama.cpp +0 -168
  895. data/ext/sources/examples/talk-llama/models/maincoder.cpp +0 -117
  896. data/ext/sources/examples/talk-llama/models/mamba.cpp +0 -55
  897. data/ext/sources/examples/talk-llama/models/mimo2-iswa.cpp +0 -123
  898. data/ext/sources/examples/talk-llama/models/minicpm3.cpp +0 -199
  899. data/ext/sources/examples/talk-llama/models/minimax-m2.cpp +0 -124
  900. data/ext/sources/examples/talk-llama/models/mistral3.cpp +0 -160
  901. data/ext/sources/examples/talk-llama/models/models.h +0 -569
  902. data/ext/sources/examples/talk-llama/models/modern-bert.cpp +0 -116
  903. data/ext/sources/examples/talk-llama/models/mpt.cpp +0 -126
  904. data/ext/sources/examples/talk-llama/models/nemotron-h.cpp +0 -150
  905. data/ext/sources/examples/talk-llama/models/nemotron.cpp +0 -122
  906. data/ext/sources/examples/talk-llama/models/neo-bert.cpp +0 -104
  907. data/ext/sources/examples/talk-llama/models/olmo.cpp +0 -121
  908. data/ext/sources/examples/talk-llama/models/olmo2.cpp +0 -150
  909. data/ext/sources/examples/talk-llama/models/olmoe.cpp +0 -124
  910. data/ext/sources/examples/talk-llama/models/openai-moe-iswa.cpp +0 -127
  911. data/ext/sources/examples/talk-llama/models/openelm.cpp +0 -124
  912. data/ext/sources/examples/talk-llama/models/orion.cpp +0 -123
  913. data/ext/sources/examples/talk-llama/models/pangu-embedded.cpp +0 -121
  914. data/ext/sources/examples/talk-llama/models/phi2.cpp +0 -121
  915. data/ext/sources/examples/talk-llama/models/phi3.cpp +0 -152
  916. data/ext/sources/examples/talk-llama/models/plamo.cpp +0 -110
  917. data/ext/sources/examples/talk-llama/models/plamo2.cpp +0 -316
  918. data/ext/sources/examples/talk-llama/models/plamo3.cpp +0 -128
  919. data/ext/sources/examples/talk-llama/models/plm.cpp +0 -168
  920. data/ext/sources/examples/talk-llama/models/qwen.cpp +0 -108
  921. data/ext/sources/examples/talk-llama/models/qwen2.cpp +0 -126
  922. data/ext/sources/examples/talk-llama/models/qwen2moe.cpp +0 -151
  923. data/ext/sources/examples/talk-llama/models/qwen2vl.cpp +0 -117
  924. data/ext/sources/examples/talk-llama/models/qwen3.cpp +0 -117
  925. data/ext/sources/examples/talk-llama/models/qwen3moe.cpp +0 -124
  926. data/ext/sources/examples/talk-llama/models/qwen3next.cpp +0 -873
  927. data/ext/sources/examples/talk-llama/models/qwen3vl-moe.cpp +0 -149
  928. data/ext/sources/examples/talk-llama/models/qwen3vl.cpp +0 -141
  929. data/ext/sources/examples/talk-llama/models/refact.cpp +0 -94
  930. data/ext/sources/examples/talk-llama/models/rnd1.cpp +0 -126
  931. data/ext/sources/examples/talk-llama/models/rwkv6-base.cpp +0 -162
  932. data/ext/sources/examples/talk-llama/models/rwkv6.cpp +0 -94
  933. data/ext/sources/examples/talk-llama/models/rwkv6qwen2.cpp +0 -86
  934. data/ext/sources/examples/talk-llama/models/rwkv7-base.cpp +0 -135
  935. data/ext/sources/examples/talk-llama/models/rwkv7.cpp +0 -90
  936. data/ext/sources/examples/talk-llama/models/seed-oss.cpp +0 -124
  937. data/ext/sources/examples/talk-llama/models/smallthinker.cpp +0 -126
  938. data/ext/sources/examples/talk-llama/models/smollm3.cpp +0 -128
  939. data/ext/sources/examples/talk-llama/models/stablelm.cpp +0 -146
  940. data/ext/sources/examples/talk-llama/models/starcoder.cpp +0 -100
  941. data/ext/sources/examples/talk-llama/models/starcoder2.cpp +0 -121
  942. data/ext/sources/examples/talk-llama/models/t5-dec.cpp +0 -166
  943. data/ext/sources/examples/talk-llama/models/t5-enc.cpp +0 -96
  944. data/ext/sources/examples/talk-llama/models/wavtokenizer-dec.cpp +0 -149
  945. data/ext/sources/examples/talk-llama/models/xverse.cpp +0 -108
  946. data/ext/sources/examples/talk-llama/prompts/talk-alpaca.txt +0 -23
  947. data/ext/sources/examples/talk-llama/speak +0 -40
  948. data/ext/sources/examples/talk-llama/speak.bat +0 -1
  949. data/ext/sources/examples/talk-llama/speak.ps1 +0 -14
  950. data/ext/sources/examples/talk-llama/talk-llama.cpp +0 -813
  951. data/ext/sources/examples/talk-llama/unicode-data.cpp +0 -7034
  952. data/ext/sources/examples/talk-llama/unicode-data.h +0 -20
  953. data/ext/sources/examples/talk-llama/unicode.cpp +0 -1147
  954. data/ext/sources/examples/talk-llama/unicode.h +0 -111
  955. data/ext/sources/examples/wchess/CMakeLists.txt +0 -10
  956. data/ext/sources/examples/wchess/libwchess/CMakeLists.txt +0 -19
  957. data/ext/sources/examples/wchess/libwchess/Chessboard.cpp +0 -803
  958. data/ext/sources/examples/wchess/libwchess/Chessboard.h +0 -33
  959. data/ext/sources/examples/wchess/libwchess/WChess.cpp +0 -193
  960. data/ext/sources/examples/wchess/libwchess/WChess.h +0 -63
  961. data/ext/sources/examples/wchess/libwchess/test-chessboard.cpp +0 -117
  962. data/ext/sources/examples/wchess/wchess.cmd/CMakeLists.txt +0 -8
  963. data/ext/sources/examples/wchess/wchess.cmd/wchess.cmd.cpp +0 -253
  964. data/ext/sources/examples/whisper.wasm/CMakeLists.txt +0 -50
  965. data/ext/sources/examples/whisper.wasm/emscripten.cpp +0 -118
  966. data/ext/sources/examples/whisper.wasm/index-tmpl.html +0 -659
  967. data/ext/sources/ggml/cmake/BuildTypes.cmake +0 -54
  968. data/ext/sources/ggml/src/ggml-cpu/llamafile/sgemm-ppc.h +0 -333
  969. data/ext/sources/ggml/src/ggml-cuda/template-instances/generate_cu_files.py +0 -99
  970. data/ext/sources/ggml/src/ggml-hexagon/htp/htp-dma.h +0 -157
  971. data/ext/sources/ggml/src/ggml-hexagon/htp/htp-msg.h +0 -165
  972. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-exp.c +0 -94
  973. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-inverse.c +0 -72
  974. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-sigmoid.c +0 -49
  975. data/ext/sources/ggml/src/ggml-hexagon/htp/hvx-utils.c +0 -1020
  976. data/ext/sources/ggml/src/ggml-hexagon/htp/ops-utils.h +0 -149
  977. data/ext/sources/ggml/src/ggml-hexagon/htp-utils.c +0 -454
  978. data/ext/sources/ggml/src/ggml-hexagon/htp-utils.h +0 -221
  979. data/ext/sources/ggml/src/ggml-hexagon/op-desc.h +0 -153
  980. data/ext/sources/ggml/src/ggml-opencl/kernels/embed_kernel.py +0 -26
  981. data/ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rte.glsl +0 -5
  982. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/bin_op.tmpl.wgsl +0 -188
  983. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/binary_head.tmpl +0 -45
  984. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/embed_wgsl.py +0 -147
  985. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/glu.tmpl.wgsl +0 -323
  986. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat.tmpl.wgsl +0 -907
  987. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_reg_tile.tmpl.wgsl +0 -247
  988. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_vec.tmpl.wgsl +0 -267
  989. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/rms_norm.wgsl +0 -123
  990. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/set_rows.tmpl.wgsl +0 -112
  991. data/ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/unary_op.wgsl +0 -483
  992. data/ext/sources/tests/CMakeLists.txt +0 -112
  993. data/ext/sources/tests/earnings21/eval.mk +0 -58
  994. data/ext/sources/tests/earnings21/eval.py +0 -68
  995. data/ext/sources/tests/earnings21/normalizers/__init__.py +0 -2
  996. data/ext/sources/tests/earnings21/normalizers/basic.py +0 -80
  997. data/ext/sources/tests/earnings21/normalizers/english.json +0 -1741
  998. data/ext/sources/tests/earnings21/normalizers/english.py +0 -550
  999. data/ext/sources/tests/earnings21/requirements.txt +0 -6
  1000. data/ext/sources/tests/en-0-ref.txt +0 -1
  1001. data/ext/sources/tests/en-1-ref.txt +0 -1
  1002. data/ext/sources/tests/en-2-ref.txt +0 -1
  1003. data/ext/sources/tests/es-0-ref.txt +0 -1
  1004. data/ext/sources/tests/librispeech/eval.mk +0 -39
  1005. data/ext/sources/tests/librispeech/eval.py +0 -47
  1006. data/ext/sources/tests/librispeech/normalizers/__init__.py +0 -2
  1007. data/ext/sources/tests/librispeech/normalizers/basic.py +0 -80
  1008. data/ext/sources/tests/librispeech/normalizers/english.json +0 -1741
  1009. data/ext/sources/tests/librispeech/normalizers/english.py +0 -550
  1010. data/ext/sources/tests/librispeech/requirements.txt +0 -6
  1011. data/ext/sources/tests/run-tests.sh +0 -130
  1012. data/ext/sources/tests/test-c.c +0 -3
  1013. data/ext/sources/tests/test-vad-full.cpp +0 -56
  1014. data/ext/sources/tests/test-vad.cpp +0 -83
  1015. data/ext/sources/tests/test-whisper.js +0 -58
  1016. data/lib/whisper/context.rb +0 -15
  1017. data/lib/whisper/segment.rb +0 -58
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: 1.3.5
4
+ version: 1.3.7
5
5
  platform: ruby
6
6
  authors:
7
7
  - Georgi Gerganov
@@ -20,19 +20,34 @@ extra_rdoc_files:
20
20
  - LICENSE
21
21
  - README.md
22
22
  files:
23
+ - ".document"
23
24
  - ".gitignore"
25
+ - ".rdoc_options"
24
26
  - LICENSE
25
27
  - README.md
26
28
  - Rakefile
27
29
  - ext/.gitignore
28
30
  - ext/dependencies.rb
31
+ - ext/dependencies_for_windows.rb
29
32
  - ext/extconf.rb
30
33
  - ext/options.rb
34
+ - ext/options_for_windows.rb
31
35
  - ext/ruby_whisper.c
32
36
  - ext/ruby_whisper.h
33
37
  - ext/ruby_whisper_context.c
38
+ - ext/ruby_whisper_context_params.c
34
39
  - ext/ruby_whisper_error.c
40
+ - ext/ruby_whisper_log_queue.c
41
+ - ext/ruby_whisper_log_settable.h
35
42
  - ext/ruby_whisper_model.c
43
+ - ext/ruby_whisper_parakeet.c
44
+ - ext/ruby_whisper_parakeet_context.c
45
+ - ext/ruby_whisper_parakeet_context_params.c
46
+ - ext/ruby_whisper_parakeet_model.c
47
+ - ext/ruby_whisper_parakeet_params.c
48
+ - ext/ruby_whisper_parakeet_segment.c
49
+ - ext/ruby_whisper_parakeet_token.c
50
+ - ext/ruby_whisper_parakeet_transcribe.cpp
36
51
  - ext/ruby_whisper_params.c
37
52
  - ext/ruby_whisper_segment.c
38
53
  - ext/ruby_whisper_token.c
@@ -44,43 +59,25 @@ files:
44
59
  - ext/ruby_whisper_vad_segments.c
45
60
  - ext/sources/CMakeGraphVizOptions.cmake
46
61
  - ext/sources/CMakeLists.txt
47
- - ext/sources/bindings/javascript/CMakeLists.txt
48
- - ext/sources/bindings/javascript/emscripten.cpp
49
- - ext/sources/bindings/javascript/libwhisper.worker.js
62
+ - ext/sources/CMakePresets.json
50
63
  - ext/sources/bindings/javascript/package-tmpl.json
51
- - ext/sources/bindings/javascript/package.json
52
- - ext/sources/bindings/javascript/whisper.js
53
64
  - ext/sources/cmake/DefaultTargetOptions.cmake
54
65
  - ext/sources/cmake/FindFFmpeg.cmake
55
66
  - ext/sources/cmake/arm64-apple-clang.cmake
56
67
  - ext/sources/cmake/arm64-windows-llvm.cmake
57
68
  - ext/sources/cmake/build-info.cmake
58
69
  - ext/sources/cmake/git-vars.cmake
70
+ - ext/sources/cmake/parakeet-config.cmake.in
71
+ - ext/sources/cmake/parakeet.pc.in
59
72
  - ext/sources/cmake/riscv64-spacemit-linux-gnu-gcc.cmake
60
73
  - ext/sources/cmake/whisper-config.cmake.in
61
74
  - ext/sources/cmake/whisper.pc.in
62
75
  - ext/sources/cmake/x64-windows-llvm.cmake
63
76
  - ext/sources/examples/CMakeLists.txt
64
- - ext/sources/examples/addon.node/CMakeLists.txt
65
- - ext/sources/examples/addon.node/__test__/whisper.spec.js
66
- - ext/sources/examples/addon.node/addon.cpp
67
- - ext/sources/examples/addon.node/index.js
68
- - ext/sources/examples/addon.node/package.json
69
- - ext/sources/examples/addon.node/vad-example.js
70
- - ext/sources/examples/bench.wasm/CMakeLists.txt
71
- - ext/sources/examples/bench.wasm/emscripten.cpp
72
- - ext/sources/examples/bench.wasm/index-tmpl.html
73
77
  - ext/sources/examples/bench/CMakeLists.txt
74
78
  - ext/sources/examples/bench/bench.cpp
75
79
  - ext/sources/examples/cli/CMakeLists.txt
76
80
  - ext/sources/examples/cli/cli.cpp
77
- - ext/sources/examples/coi-serviceworker.js
78
- - ext/sources/examples/command.wasm/CMakeLists.txt
79
- - ext/sources/examples/command.wasm/emscripten.cpp
80
- - ext/sources/examples/command.wasm/index-tmpl.html
81
- - ext/sources/examples/command/CMakeLists.txt
82
- - ext/sources/examples/command/command.cpp
83
- - ext/sources/examples/command/commands.txt
84
81
  - ext/sources/examples/common-ggml.cpp
85
82
  - ext/sources/examples/common-ggml.h
86
83
  - ext/sources/examples/common-sdl.cpp
@@ -92,215 +89,24 @@ files:
92
89
  - ext/sources/examples/deprecation-warning/CMakeLists.txt
93
90
  - ext/sources/examples/deprecation-warning/deprecation-warning.cpp
94
91
  - ext/sources/examples/ffmpeg-transcode.cpp
95
- - ext/sources/examples/generate-karaoke.sh
96
92
  - ext/sources/examples/grammar-parser.cpp
97
93
  - ext/sources/examples/grammar-parser.h
98
- - ext/sources/examples/helpers.js
99
94
  - ext/sources/examples/json.hpp
100
- - ext/sources/examples/livestream.sh
101
- - ext/sources/examples/lsp/CMakeLists.txt
102
- - ext/sources/examples/lsp/lsp.cpp
103
- - ext/sources/examples/lsp/whisper.vim
104
95
  - ext/sources/examples/miniaudio.h
105
- - ext/sources/examples/python/test_whisper_processor.py
106
- - ext/sources/examples/python/whisper_processor.py
96
+ - ext/sources/examples/parakeet-cli/CMakeLists.txt
97
+ - ext/sources/examples/parakeet-cli/parakeet-cli.cpp
98
+ - ext/sources/examples/parakeet-quantize/CMakeLists.txt
99
+ - ext/sources/examples/parakeet-quantize/parakeet-quantize.cpp
107
100
  - ext/sources/examples/quantize/CMakeLists.txt
108
101
  - ext/sources/examples/quantize/quantize.cpp
109
- - ext/sources/examples/server.py
110
102
  - ext/sources/examples/server/CMakeLists.txt
111
- - ext/sources/examples/server/bench.js
112
103
  - ext/sources/examples/server/httplib.h
113
104
  - ext/sources/examples/server/server.cpp
114
105
  - ext/sources/examples/stb_vorbis.c
115
- - ext/sources/examples/stream.wasm/CMakeLists.txt
116
- - ext/sources/examples/stream.wasm/emscripten.cpp
117
- - ext/sources/examples/stream.wasm/index-tmpl.html
118
- - ext/sources/examples/stream/CMakeLists.txt
119
- - ext/sources/examples/stream/stream.cpp
120
- - ext/sources/examples/sycl/CMakeLists.txt
121
- - ext/sources/examples/sycl/build.sh
122
- - ext/sources/examples/sycl/ls-sycl-device.cpp
123
- - ext/sources/examples/sycl/run-whisper.sh
124
- - ext/sources/examples/talk-llama/CMakeLists.txt
125
- - ext/sources/examples/talk-llama/eleven-labs.py
126
- - ext/sources/examples/talk-llama/llama-adapter.cpp
127
- - ext/sources/examples/talk-llama/llama-adapter.h
128
- - ext/sources/examples/talk-llama/llama-arch.cpp
129
- - ext/sources/examples/talk-llama/llama-arch.h
130
- - ext/sources/examples/talk-llama/llama-batch.cpp
131
- - ext/sources/examples/talk-llama/llama-batch.h
132
- - ext/sources/examples/talk-llama/llama-chat.cpp
133
- - ext/sources/examples/talk-llama/llama-chat.h
134
- - ext/sources/examples/talk-llama/llama-context.cpp
135
- - ext/sources/examples/talk-llama/llama-context.h
136
- - ext/sources/examples/talk-llama/llama-cparams.cpp
137
- - ext/sources/examples/talk-llama/llama-cparams.h
138
- - ext/sources/examples/talk-llama/llama-grammar.cpp
139
- - ext/sources/examples/talk-llama/llama-grammar.h
140
- - ext/sources/examples/talk-llama/llama-graph.cpp
141
- - ext/sources/examples/talk-llama/llama-graph.h
142
- - ext/sources/examples/talk-llama/llama-hparams.cpp
143
- - ext/sources/examples/talk-llama/llama-hparams.h
144
- - ext/sources/examples/talk-llama/llama-impl.cpp
145
- - ext/sources/examples/talk-llama/llama-impl.h
146
- - ext/sources/examples/talk-llama/llama-io.cpp
147
- - ext/sources/examples/talk-llama/llama-io.h
148
- - ext/sources/examples/talk-llama/llama-kv-cache-iswa.cpp
149
- - ext/sources/examples/talk-llama/llama-kv-cache-iswa.h
150
- - ext/sources/examples/talk-llama/llama-kv-cache.cpp
151
- - ext/sources/examples/talk-llama/llama-kv-cache.h
152
- - ext/sources/examples/talk-llama/llama-kv-cells.h
153
- - ext/sources/examples/talk-llama/llama-memory-hybrid.cpp
154
- - ext/sources/examples/talk-llama/llama-memory-hybrid.h
155
- - ext/sources/examples/talk-llama/llama-memory-recurrent.cpp
156
- - ext/sources/examples/talk-llama/llama-memory-recurrent.h
157
- - ext/sources/examples/talk-llama/llama-memory.cpp
158
- - ext/sources/examples/talk-llama/llama-memory.h
159
- - ext/sources/examples/talk-llama/llama-mmap.cpp
160
- - ext/sources/examples/talk-llama/llama-mmap.h
161
- - ext/sources/examples/talk-llama/llama-model-loader.cpp
162
- - ext/sources/examples/talk-llama/llama-model-loader.h
163
- - ext/sources/examples/talk-llama/llama-model-saver.cpp
164
- - ext/sources/examples/talk-llama/llama-model-saver.h
165
- - ext/sources/examples/talk-llama/llama-model.cpp
166
- - ext/sources/examples/talk-llama/llama-model.h
167
- - ext/sources/examples/talk-llama/llama-quant.cpp
168
- - ext/sources/examples/talk-llama/llama-quant.h
169
- - ext/sources/examples/talk-llama/llama-sampling.cpp
170
- - ext/sources/examples/talk-llama/llama-sampling.h
171
- - ext/sources/examples/talk-llama/llama-vocab.cpp
172
- - ext/sources/examples/talk-llama/llama-vocab.h
173
- - ext/sources/examples/talk-llama/llama.cpp
174
- - ext/sources/examples/talk-llama/llama.h
175
- - ext/sources/examples/talk-llama/models/afmoe.cpp
176
- - ext/sources/examples/talk-llama/models/apertus.cpp
177
- - ext/sources/examples/talk-llama/models/arcee.cpp
178
- - ext/sources/examples/talk-llama/models/arctic.cpp
179
- - ext/sources/examples/talk-llama/models/arwkv7.cpp
180
- - ext/sources/examples/talk-llama/models/baichuan.cpp
181
- - ext/sources/examples/talk-llama/models/bailingmoe.cpp
182
- - ext/sources/examples/talk-llama/models/bailingmoe2.cpp
183
- - ext/sources/examples/talk-llama/models/bert.cpp
184
- - ext/sources/examples/talk-llama/models/bitnet.cpp
185
- - ext/sources/examples/talk-llama/models/bloom.cpp
186
- - ext/sources/examples/talk-llama/models/chameleon.cpp
187
- - ext/sources/examples/talk-llama/models/chatglm.cpp
188
- - ext/sources/examples/talk-llama/models/codeshell.cpp
189
- - ext/sources/examples/talk-llama/models/cogvlm.cpp
190
- - ext/sources/examples/talk-llama/models/cohere2-iswa.cpp
191
- - ext/sources/examples/talk-llama/models/command-r.cpp
192
- - ext/sources/examples/talk-llama/models/dbrx.cpp
193
- - ext/sources/examples/talk-llama/models/deci.cpp
194
- - ext/sources/examples/talk-llama/models/deepseek.cpp
195
- - ext/sources/examples/talk-llama/models/deepseek2.cpp
196
- - ext/sources/examples/talk-llama/models/dots1.cpp
197
- - ext/sources/examples/talk-llama/models/dream.cpp
198
- - ext/sources/examples/talk-llama/models/ernie4-5-moe.cpp
199
- - ext/sources/examples/talk-llama/models/ernie4-5.cpp
200
- - ext/sources/examples/talk-llama/models/exaone.cpp
201
- - ext/sources/examples/talk-llama/models/exaone4.cpp
202
- - ext/sources/examples/talk-llama/models/falcon-h1.cpp
203
- - ext/sources/examples/talk-llama/models/falcon.cpp
204
- - ext/sources/examples/talk-llama/models/gemma-embedding.cpp
205
- - ext/sources/examples/talk-llama/models/gemma.cpp
206
- - ext/sources/examples/talk-llama/models/gemma2-iswa.cpp
207
- - ext/sources/examples/talk-llama/models/gemma3.cpp
208
- - ext/sources/examples/talk-llama/models/gemma3n-iswa.cpp
209
- - ext/sources/examples/talk-llama/models/glm4-moe.cpp
210
- - ext/sources/examples/talk-llama/models/glm4.cpp
211
- - ext/sources/examples/talk-llama/models/gpt2.cpp
212
- - ext/sources/examples/talk-llama/models/gptneox.cpp
213
- - ext/sources/examples/talk-llama/models/granite-hybrid.cpp
214
- - ext/sources/examples/talk-llama/models/granite.cpp
215
- - ext/sources/examples/talk-llama/models/graph-context-mamba.cpp
216
- - ext/sources/examples/talk-llama/models/grok.cpp
217
- - ext/sources/examples/talk-llama/models/grovemoe.cpp
218
- - ext/sources/examples/talk-llama/models/hunyuan-dense.cpp
219
- - ext/sources/examples/talk-llama/models/hunyuan-moe.cpp
220
- - ext/sources/examples/talk-llama/models/internlm2.cpp
221
- - ext/sources/examples/talk-llama/models/jais.cpp
222
- - ext/sources/examples/talk-llama/models/jamba.cpp
223
- - ext/sources/examples/talk-llama/models/lfm2.cpp
224
- - ext/sources/examples/talk-llama/models/llada-moe.cpp
225
- - ext/sources/examples/talk-llama/models/llada.cpp
226
- - ext/sources/examples/talk-llama/models/llama-iswa.cpp
227
- - ext/sources/examples/talk-llama/models/llama.cpp
228
- - ext/sources/examples/talk-llama/models/maincoder.cpp
229
- - ext/sources/examples/talk-llama/models/mamba.cpp
230
- - ext/sources/examples/talk-llama/models/mimo2-iswa.cpp
231
- - ext/sources/examples/talk-llama/models/minicpm3.cpp
232
- - ext/sources/examples/talk-llama/models/minimax-m2.cpp
233
- - ext/sources/examples/talk-llama/models/mistral3.cpp
234
- - ext/sources/examples/talk-llama/models/models.h
235
- - ext/sources/examples/talk-llama/models/modern-bert.cpp
236
- - ext/sources/examples/talk-llama/models/mpt.cpp
237
- - ext/sources/examples/talk-llama/models/nemotron-h.cpp
238
- - ext/sources/examples/talk-llama/models/nemotron.cpp
239
- - ext/sources/examples/talk-llama/models/neo-bert.cpp
240
- - ext/sources/examples/talk-llama/models/olmo.cpp
241
- - ext/sources/examples/talk-llama/models/olmo2.cpp
242
- - ext/sources/examples/talk-llama/models/olmoe.cpp
243
- - ext/sources/examples/talk-llama/models/openai-moe-iswa.cpp
244
- - ext/sources/examples/talk-llama/models/openelm.cpp
245
- - ext/sources/examples/talk-llama/models/orion.cpp
246
- - ext/sources/examples/talk-llama/models/pangu-embedded.cpp
247
- - ext/sources/examples/talk-llama/models/phi2.cpp
248
- - ext/sources/examples/talk-llama/models/phi3.cpp
249
- - ext/sources/examples/talk-llama/models/plamo.cpp
250
- - ext/sources/examples/talk-llama/models/plamo2.cpp
251
- - ext/sources/examples/talk-llama/models/plamo3.cpp
252
- - ext/sources/examples/talk-llama/models/plm.cpp
253
- - ext/sources/examples/talk-llama/models/qwen.cpp
254
- - ext/sources/examples/talk-llama/models/qwen2.cpp
255
- - ext/sources/examples/talk-llama/models/qwen2moe.cpp
256
- - ext/sources/examples/talk-llama/models/qwen2vl.cpp
257
- - ext/sources/examples/talk-llama/models/qwen3.cpp
258
- - ext/sources/examples/talk-llama/models/qwen3moe.cpp
259
- - ext/sources/examples/talk-llama/models/qwen3next.cpp
260
- - ext/sources/examples/talk-llama/models/qwen3vl-moe.cpp
261
- - ext/sources/examples/talk-llama/models/qwen3vl.cpp
262
- - ext/sources/examples/talk-llama/models/refact.cpp
263
- - ext/sources/examples/talk-llama/models/rnd1.cpp
264
- - ext/sources/examples/talk-llama/models/rwkv6-base.cpp
265
- - ext/sources/examples/talk-llama/models/rwkv6.cpp
266
- - ext/sources/examples/talk-llama/models/rwkv6qwen2.cpp
267
- - ext/sources/examples/talk-llama/models/rwkv7-base.cpp
268
- - ext/sources/examples/talk-llama/models/rwkv7.cpp
269
- - ext/sources/examples/talk-llama/models/seed-oss.cpp
270
- - ext/sources/examples/talk-llama/models/smallthinker.cpp
271
- - ext/sources/examples/talk-llama/models/smollm3.cpp
272
- - ext/sources/examples/talk-llama/models/stablelm.cpp
273
- - ext/sources/examples/talk-llama/models/starcoder.cpp
274
- - ext/sources/examples/talk-llama/models/starcoder2.cpp
275
- - ext/sources/examples/talk-llama/models/t5-dec.cpp
276
- - ext/sources/examples/talk-llama/models/t5-enc.cpp
277
- - ext/sources/examples/talk-llama/models/wavtokenizer-dec.cpp
278
- - ext/sources/examples/talk-llama/models/xverse.cpp
279
- - ext/sources/examples/talk-llama/prompts/talk-alpaca.txt
280
- - ext/sources/examples/talk-llama/speak
281
- - ext/sources/examples/talk-llama/speak.bat
282
- - ext/sources/examples/talk-llama/speak.ps1
283
- - ext/sources/examples/talk-llama/talk-llama.cpp
284
- - ext/sources/examples/talk-llama/unicode-data.cpp
285
- - ext/sources/examples/talk-llama/unicode-data.h
286
- - ext/sources/examples/talk-llama/unicode.cpp
287
- - ext/sources/examples/talk-llama/unicode.h
288
106
  - ext/sources/examples/vad-speech-segments/CMakeLists.txt
289
107
  - ext/sources/examples/vad-speech-segments/speech.cpp
290
- - ext/sources/examples/wchess/CMakeLists.txt
291
- - ext/sources/examples/wchess/libwchess/CMakeLists.txt
292
- - ext/sources/examples/wchess/libwchess/Chessboard.cpp
293
- - ext/sources/examples/wchess/libwchess/Chessboard.h
294
- - ext/sources/examples/wchess/libwchess/WChess.cpp
295
- - ext/sources/examples/wchess/libwchess/WChess.h
296
- - ext/sources/examples/wchess/libwchess/test-chessboard.cpp
297
- - ext/sources/examples/wchess/wchess.cmd/CMakeLists.txt
298
- - ext/sources/examples/wchess/wchess.cmd/wchess.cmd.cpp
299
- - ext/sources/examples/whisper.wasm/CMakeLists.txt
300
- - ext/sources/examples/whisper.wasm/emscripten.cpp
301
- - ext/sources/examples/whisper.wasm/index-tmpl.html
302
108
  - ext/sources/ggml/CMakeLists.txt
303
- - ext/sources/ggml/cmake/BuildTypes.cmake
109
+ - ext/sources/ggml/cmake/FindNCCL.cmake
304
110
  - ext/sources/ggml/cmake/GitVars.cmake
305
111
  - ext/sources/ggml/cmake/common.cmake
306
112
  - ext/sources/ggml/cmake/ggml-config.cmake.in
@@ -314,9 +120,11 @@ files:
314
120
  - ext/sources/ggml/include/ggml-hexagon.h
315
121
  - ext/sources/ggml/include/ggml-metal.h
316
122
  - ext/sources/ggml/include/ggml-opencl.h
123
+ - ext/sources/ggml/include/ggml-openvino.h
317
124
  - ext/sources/ggml/include/ggml-opt.h
318
125
  - ext/sources/ggml/include/ggml-rpc.h
319
126
  - ext/sources/ggml/include/ggml-sycl.h
127
+ - ext/sources/ggml/include/ggml-virtgpu.h
320
128
  - ext/sources/ggml/include/ggml-vulkan.h
321
129
  - ext/sources/ggml/include/ggml-webgpu.h
322
130
  - ext/sources/ggml/include/ggml-zdnn.h
@@ -325,7 +133,10 @@ files:
325
133
  - ext/sources/ggml/include/gguf.h
326
134
  - ext/sources/ggml/src/CMakeLists.txt
327
135
  - ext/sources/ggml/src/ggml-alloc.c
136
+ - ext/sources/ggml/src/ggml-backend-dl.cpp
137
+ - ext/sources/ggml/src/ggml-backend-dl.h
328
138
  - ext/sources/ggml/src/ggml-backend-impl.h
139
+ - ext/sources/ggml/src/ggml-backend-meta.cpp
329
140
  - ext/sources/ggml/src/ggml-backend-reg.cpp
330
141
  - ext/sources/ggml/src/ggml-backend.cpp
331
142
  - ext/sources/ggml/src/ggml-blas/CMakeLists.txt
@@ -363,6 +174,7 @@ files:
363
174
  - ext/sources/ggml/src/ggml-cpu/binary-ops.cpp
364
175
  - ext/sources/ggml/src/ggml-cpu/binary-ops.h
365
176
  - ext/sources/ggml/src/ggml-cpu/cmake/FindSIMD.cmake
177
+ - ext/sources/ggml/src/ggml-cpu/cmake/FindSMTIME.cmake
366
178
  - ext/sources/ggml/src/ggml-cpu/common.h
367
179
  - ext/sources/ggml/src/ggml-cpu/ggml-cpu-impl.h
368
180
  - ext/sources/ggml/src/ggml-cpu/ggml-cpu.c
@@ -373,7 +185,6 @@ files:
373
185
  - ext/sources/ggml/src/ggml-cpu/kleidiai/kernels.h
374
186
  - ext/sources/ggml/src/ggml-cpu/kleidiai/kleidiai.cpp
375
187
  - ext/sources/ggml/src/ggml-cpu/kleidiai/kleidiai.h
376
- - ext/sources/ggml/src/ggml-cpu/llamafile/sgemm-ppc.h
377
188
  - ext/sources/ggml/src/ggml-cpu/llamafile/sgemm.cpp
378
189
  - ext/sources/ggml/src/ggml-cpu/llamafile/sgemm.h
379
190
  - ext/sources/ggml/src/ggml-cpu/ops.cpp
@@ -382,11 +193,23 @@ files:
382
193
  - ext/sources/ggml/src/ggml-cpu/quants.h
383
194
  - ext/sources/ggml/src/ggml-cpu/repack.cpp
384
195
  - ext/sources/ggml/src/ggml-cpu/repack.h
196
+ - ext/sources/ggml/src/ggml-cpu/simd-gemm.h
385
197
  - ext/sources/ggml/src/ggml-cpu/simd-mappings.h
386
198
  - ext/sources/ggml/src/ggml-cpu/spacemit/ime.cpp
387
199
  - ext/sources/ggml/src/ggml-cpu/spacemit/ime.h
388
200
  - ext/sources/ggml/src/ggml-cpu/spacemit/ime1_kernels.cpp
201
+ - ext/sources/ggml/src/ggml-cpu/spacemit/ime2_kernels.cpp
202
+ - ext/sources/ggml/src/ggml-cpu/spacemit/ime_env.cpp
203
+ - ext/sources/ggml/src/ggml-cpu/spacemit/ime_env.h
389
204
  - ext/sources/ggml/src/ggml-cpu/spacemit/ime_kernels.h
205
+ - ext/sources/ggml/src/ggml-cpu/spacemit/repack.cpp
206
+ - ext/sources/ggml/src/ggml-cpu/spacemit/repack.h
207
+ - ext/sources/ggml/src/ggml-cpu/spacemit/rvv_kernels.cpp
208
+ - ext/sources/ggml/src/ggml-cpu/spacemit/rvv_kernels.h
209
+ - ext/sources/ggml/src/ggml-cpu/spacemit/spine_barrier.h
210
+ - ext/sources/ggml/src/ggml-cpu/spacemit/spine_mem_pool.cpp
211
+ - ext/sources/ggml/src/ggml-cpu/spacemit/spine_mem_pool.h
212
+ - ext/sources/ggml/src/ggml-cpu/spacemit/spine_tcm.h
390
213
  - ext/sources/ggml/src/ggml-cpu/traits.cpp
391
214
  - ext/sources/ggml/src/ggml-cpu/traits.h
392
215
  - ext/sources/ggml/src/ggml-cpu/unary-ops.cpp
@@ -398,6 +221,8 @@ files:
398
221
  - ext/sources/ggml/src/ggml-cuda/acc.cuh
399
222
  - ext/sources/ggml/src/ggml-cuda/add-id.cu
400
223
  - ext/sources/ggml/src/ggml-cuda/add-id.cuh
224
+ - ext/sources/ggml/src/ggml-cuda/allreduce.cu
225
+ - ext/sources/ggml/src/ggml-cuda/allreduce.cuh
401
226
  - ext/sources/ggml/src/ggml-cuda/arange.cu
402
227
  - ext/sources/ggml/src/ggml-cuda/arange.cuh
403
228
  - ext/sources/ggml/src/ggml-cuda/argmax.cu
@@ -447,6 +272,10 @@ files:
447
272
  - ext/sources/ggml/src/ggml-cuda/fattn.cuh
448
273
  - ext/sources/ggml/src/ggml-cuda/fill.cu
449
274
  - ext/sources/ggml/src/ggml-cuda/fill.cuh
275
+ - ext/sources/ggml/src/ggml-cuda/fwht.cu
276
+ - ext/sources/ggml/src/ggml-cuda/fwht.cuh
277
+ - ext/sources/ggml/src/ggml-cuda/gated_delta_net.cu
278
+ - ext/sources/ggml/src/ggml-cuda/gated_delta_net.cuh
450
279
  - ext/sources/ggml/src/ggml-cuda/getrows.cu
451
280
  - ext/sources/ggml/src/ggml-cuda/getrows.cuh
452
281
  - ext/sources/ggml/src/ggml-cuda/ggml-cuda.cu
@@ -494,6 +323,8 @@ files:
494
323
  - ext/sources/ggml/src/ggml-cuda/set-rows.cuh
495
324
  - ext/sources/ggml/src/ggml-cuda/set.cu
496
325
  - ext/sources/ggml/src/ggml-cuda/set.cuh
326
+ - ext/sources/ggml/src/ggml-cuda/snake.cu
327
+ - ext/sources/ggml/src/ggml-cuda/snake.cuh
497
328
  - ext/sources/ggml/src/ggml-cuda/softcap.cu
498
329
  - ext/sources/ggml/src/ggml-cuda/softcap.cuh
499
330
  - ext/sources/ggml/src/ggml-cuda/softmax.cu
@@ -509,11 +340,13 @@ files:
509
340
  - ext/sources/ggml/src/ggml-cuda/sumrows.cu
510
341
  - ext/sources/ggml/src/ggml-cuda/sumrows.cuh
511
342
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_1-ncols2_16.cu
343
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_1-ncols2_32.cu
512
344
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_1-ncols2_8.cu
513
345
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_1.cu
514
346
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_2.cu
515
347
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_16-ncols2_4.cu
516
348
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_16.cu
349
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_32.cu
517
350
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_4.cu
518
351
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_2-ncols2_8.cu
519
352
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_32-ncols2_1.cu
@@ -529,50 +362,65 @@ files:
529
362
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-mma-f16-instance-ncols1_8-ncols2_8.cu
530
363
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq112-dv112.cu
531
364
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq128-dv128.cu
365
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq192-dv128.cu
532
366
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq256-dv256.cu
367
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq320-dv256.cu
533
368
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq40-dv40.cu
369
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq512-dv512.cu
534
370
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq576-dv512.cu
535
371
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq64-dv64.cu
536
372
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq72-dv72.cu
537
373
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq80-dv80.cu
538
374
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq96-dv96.cu
375
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-bf16.cu
376
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-f16.cu
377
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-q4_0.cu
378
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-q4_1.cu
379
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-q5_0.cu
380
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-q5_1.cu
381
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-bf16-q8_0.cu
382
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-bf16.cu
539
383
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-f16.cu
540
384
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-q4_0.cu
541
385
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-q4_1.cu
542
386
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-q5_0.cu
543
387
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-q5_1.cu
544
388
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-q8_0.cu
389
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-bf16.cu
545
390
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-f16.cu
546
391
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-q4_0.cu
547
392
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-q4_1.cu
548
393
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-q5_0.cu
549
394
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-q5_1.cu
550
395
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-q8_0.cu
396
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-bf16.cu
551
397
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-f16.cu
552
398
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-q4_0.cu
553
399
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-q4_1.cu
554
400
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-q5_0.cu
555
401
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-q5_1.cu
556
402
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-q8_0.cu
403
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-bf16.cu
557
404
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-f16.cu
558
405
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-q4_0.cu
559
406
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-q4_1.cu
560
407
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-q5_0.cu
561
408
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-q5_1.cu
562
409
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-q8_0.cu
410
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-bf16.cu
563
411
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-f16.cu
564
412
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-q4_0.cu
565
413
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-q4_1.cu
566
414
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-q5_0.cu
567
415
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-q5_1.cu
568
416
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-q8_0.cu
417
+ - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-bf16.cu
569
418
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-f16.cu
570
419
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-q4_0.cu
571
420
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-q4_1.cu
572
421
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-q5_0.cu
573
422
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-q5_1.cu
574
423
  - ext/sources/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-q8_0.cu
575
- - ext/sources/ggml/src/ggml-cuda/template-instances/generate_cu_files.py
576
424
  - ext/sources/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_1.cu
577
425
  - ext/sources/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_10.cu
578
426
  - ext/sources/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_11.cu
@@ -598,6 +446,8 @@ files:
598
446
  - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq4_nl.cu
599
447
  - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-iq4_xs.cu
600
448
  - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-mxfp4.cu
449
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-nvfp4.cu
450
+ - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q1_0.cu
601
451
  - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q2_k.cu
602
452
  - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q3_k.cu
603
453
  - ext/sources/ggml/src/ggml-cuda/template-instances/mmq-instance-q4_0.cu
@@ -628,35 +478,73 @@ files:
628
478
  - ext/sources/ggml/src/ggml-cuda/wkv.cuh
629
479
  - ext/sources/ggml/src/ggml-hexagon/CMakeLists.txt
630
480
  - ext/sources/ggml/src/ggml-hexagon/ggml-hexagon.cpp
631
- - ext/sources/ggml/src/ggml-hexagon/htp-utils.c
632
- - ext/sources/ggml/src/ggml-hexagon/htp-utils.h
481
+ - ext/sources/ggml/src/ggml-hexagon/htp-drv.cpp
482
+ - ext/sources/ggml/src/ggml-hexagon/htp-drv.h
483
+ - ext/sources/ggml/src/ggml-hexagon/htp-opnode.h
633
484
  - ext/sources/ggml/src/ggml-hexagon/htp/CMakeLists.txt
634
485
  - ext/sources/ggml/src/ggml-hexagon/htp/act-ops.c
486
+ - ext/sources/ggml/src/ggml-hexagon/htp/argsort-ops.c
635
487
  - ext/sources/ggml/src/ggml-hexagon/htp/binary-ops.c
636
488
  - ext/sources/ggml/src/ggml-hexagon/htp/cmake-toolchain.cmake
489
+ - ext/sources/ggml/src/ggml-hexagon/htp/concat-ops.c
490
+ - ext/sources/ggml/src/ggml-hexagon/htp/cpy-ops.c
491
+ - ext/sources/ggml/src/ggml-hexagon/htp/cumsum-ops.c
492
+ - ext/sources/ggml/src/ggml-hexagon/htp/diag-ops.c
493
+ - ext/sources/ggml/src/ggml-hexagon/htp/fill-ops.c
637
494
  - ext/sources/ggml/src/ggml-hexagon/htp/flash-attn-ops.c
495
+ - ext/sources/ggml/src/ggml-hexagon/htp/gated-delta-net-ops.c
638
496
  - ext/sources/ggml/src/ggml-hexagon/htp/get-rows-ops.c
497
+ - ext/sources/ggml/src/ggml-hexagon/htp/hex-dma.c
498
+ - ext/sources/ggml/src/ggml-hexagon/htp/hex-dma.h
499
+ - ext/sources/ggml/src/ggml-hexagon/htp/hex-dump.h
500
+ - ext/sources/ggml/src/ggml-hexagon/htp/hex-fastdiv.h
501
+ - ext/sources/ggml/src/ggml-hexagon/htp/hex-utils.h
502
+ - ext/sources/ggml/src/ggml-hexagon/htp/hmx-flash-attn-ops.c
503
+ - ext/sources/ggml/src/ggml-hexagon/htp/hmx-matmul-ops.c
504
+ - ext/sources/ggml/src/ggml-hexagon/htp/hmx-ops.c
505
+ - ext/sources/ggml/src/ggml-hexagon/htp/hmx-ops.h
506
+ - ext/sources/ggml/src/ggml-hexagon/htp/hmx-profile.h
507
+ - ext/sources/ggml/src/ggml-hexagon/htp/hmx-queue.c
508
+ - ext/sources/ggml/src/ggml-hexagon/htp/hmx-queue.h
509
+ - ext/sources/ggml/src/ggml-hexagon/htp/hmx-utils.h
639
510
  - ext/sources/ggml/src/ggml-hexagon/htp/htp-ctx.h
640
- - ext/sources/ggml/src/ggml-hexagon/htp/htp-dma.c
641
- - ext/sources/ggml/src/ggml-hexagon/htp/htp-dma.h
642
- - ext/sources/ggml/src/ggml-hexagon/htp/htp-msg.h
643
511
  - ext/sources/ggml/src/ggml-hexagon/htp/htp-ops.h
644
512
  - ext/sources/ggml/src/ggml-hexagon/htp/htp_iface.idl
645
- - ext/sources/ggml/src/ggml-hexagon/htp/hvx-exp.c
646
- - ext/sources/ggml/src/ggml-hexagon/htp/hvx-inverse.c
647
- - ext/sources/ggml/src/ggml-hexagon/htp/hvx-sigmoid.c
648
- - ext/sources/ggml/src/ggml-hexagon/htp/hvx-utils.c
513
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-arith.h
514
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-base.h
515
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-copy.h
516
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-div.h
517
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-dump.h
518
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-exp.h
519
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-flash-attn.h
520
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-floor.h
521
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-inverse.h
522
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-log.h
523
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-pow.h
524
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-reduce.h
525
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-repl.h
526
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-scale.h
527
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-sigmoid.h
528
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-sin-cos.h
529
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-sqrt.h
530
+ - ext/sources/ggml/src/ggml-hexagon/htp/hvx-types.h
649
531
  - ext/sources/ggml/src/ggml-hexagon/htp/hvx-utils.h
650
532
  - ext/sources/ggml/src/ggml-hexagon/htp/main.c
651
533
  - ext/sources/ggml/src/ggml-hexagon/htp/matmul-ops.c
652
- - ext/sources/ggml/src/ggml-hexagon/htp/ops-utils.h
534
+ - ext/sources/ggml/src/ggml-hexagon/htp/pad-ops.c
535
+ - ext/sources/ggml/src/ggml-hexagon/htp/repeat-ops.c
653
536
  - ext/sources/ggml/src/ggml-hexagon/htp/rope-ops.c
654
537
  - ext/sources/ggml/src/ggml-hexagon/htp/set-rows-ops.c
655
538
  - ext/sources/ggml/src/ggml-hexagon/htp/softmax-ops.c
539
+ - ext/sources/ggml/src/ggml-hexagon/htp/solve-tri-ops.c
540
+ - ext/sources/ggml/src/ggml-hexagon/htp/ssm-conv.c
541
+ - ext/sources/ggml/src/ggml-hexagon/htp/sum-rows-ops.c
656
542
  - ext/sources/ggml/src/ggml-hexagon/htp/unary-ops.c
543
+ - ext/sources/ggml/src/ggml-hexagon/htp/vtcm-utils.h
657
544
  - ext/sources/ggml/src/ggml-hexagon/htp/worker-pool.c
658
545
  - ext/sources/ggml/src/ggml-hexagon/htp/worker-pool.h
659
- - ext/sources/ggml/src/ggml-hexagon/op-desc.h
546
+ - ext/sources/ggml/src/ggml-hexagon/libdl.h
547
+ - ext/sources/ggml/src/ggml-hexagon/libggml-htp.inf
660
548
  - ext/sources/ggml/src/ggml-hip/CMakeLists.txt
661
549
  - ext/sources/ggml/src/ggml-impl.h
662
550
  - ext/sources/ggml/src/ggml-metal/CMakeLists.txt
@@ -685,32 +573,79 @@ files:
685
573
  - ext/sources/ggml/src/ggml-opencl/kernels/conv2d.cl
686
574
  - ext/sources/ggml/src/ggml-opencl/kernels/conv2d_f16_f32.cl
687
575
  - ext/sources/ggml/src/ggml-opencl/kernels/cpy.cl
576
+ - ext/sources/ggml/src/ggml-opencl/kernels/cumsum.cl
688
577
  - ext/sources/ggml/src/ggml-opencl/kernels/cvt.cl
578
+ - ext/sources/ggml/src/ggml-opencl/kernels/diag.cl
689
579
  - ext/sources/ggml/src/ggml-opencl/kernels/diag_mask_inf.cl
690
580
  - ext/sources/ggml/src/ggml-opencl/kernels/div.cl
691
- - ext/sources/ggml/src/ggml-opencl/kernels/embed_kernel.py
581
+ - ext/sources/ggml/src/ggml-opencl/kernels/exp.cl
692
582
  - ext/sources/ggml/src/ggml-opencl/kernels/expm1.cl
693
583
  - ext/sources/ggml/src/ggml-opencl/kernels/fill.cl
694
584
  - ext/sources/ggml/src/ggml-opencl/kernels/flash_attn_f16.cl
695
585
  - ext/sources/ggml/src/ggml-opencl/kernels/flash_attn_f32.cl
696
586
  - ext/sources/ggml/src/ggml-opencl/kernels/flash_attn_f32_f16.cl
587
+ - ext/sources/ggml/src/ggml-opencl/kernels/gated_delta_net.cl
697
588
  - ext/sources/ggml/src/ggml-opencl/kernels/gelu.cl
698
589
  - ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_mxfp4_f32.cl
590
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_mxfp4_f32_ns.cl
591
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q4_0_f32_ns.cl
592
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q4_1_f32_ns.cl
593
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q4_k_f32_ns.cl
594
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q5_0_f32_ns.cl
595
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q5_1_f32_ns.cl
596
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q5_k_f32_ns.cl
597
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_moe_q6_k_f32_ns.cl
598
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_iq4_nl_f32.cl
599
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q4_0_f32.cl
600
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q4_1_f32.cl
601
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q4_k_f32.cl
602
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q5_0_f32.cl
603
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q5_1_f32.cl
604
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q5_k_f32.cl
605
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q6_k_f32.cl
606
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_noshuffle_q8_0_f32.cl
607
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemm_xmem_f16_f32_os8.cl
699
608
  - ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_mxfp4_f32.cl
700
- - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle.cl
701
- - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_general.cl
609
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_mxfp4_f32_ns.cl
610
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q4_0_f32_ns.cl
611
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q4_1_f32_ns.cl
612
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q4_k_f32_ns.cl
613
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q5_0_f32_ns.cl
614
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q5_1_f32_ns.cl
615
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q5_k_f32_ns.cl
616
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_moe_q6_k_f32_ns.cl
617
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_iq4_nl_f32.cl
618
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q4_0_f32.cl
619
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q4_0_f32_spec.cl
620
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q4_1_f32.cl
621
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q4_k_f32.cl
622
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q5_0_f32.cl
623
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q5_1_f32.cl
624
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q5_k_f32.cl
625
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q6_k_f32.cl
626
+ - ext/sources/ggml/src/ggml-opencl/kernels/gemv_noshuffle_q8_0_f32.cl
702
627
  - ext/sources/ggml/src/ggml-opencl/kernels/get_rows.cl
703
628
  - ext/sources/ggml/src/ggml-opencl/kernels/glu.cl
704
629
  - ext/sources/ggml/src/ggml-opencl/kernels/group_norm.cl
705
630
  - ext/sources/ggml/src/ggml-opencl/kernels/im2col_f16.cl
706
631
  - ext/sources/ggml/src/ggml-opencl/kernels/im2col_f32.cl
632
+ - ext/sources/ggml/src/ggml-opencl/kernels/l2_norm.cl
707
633
  - ext/sources/ggml/src/ggml-opencl/kernels/mean.cl
634
+ - ext/sources/ggml/src/ggml-opencl/kernels/moe_reorder_b.cl
635
+ - ext/sources/ggml/src/ggml-opencl/kernels/moe_sort_by_expert.cl
708
636
  - ext/sources/ggml/src/ggml-opencl/kernels/mul.cl
709
- - ext/sources/ggml/src/ggml-opencl/kernels/mul_mat_Ab_Bi_8x4.cl
710
637
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mat_f16_f32.cl
711
638
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_f16_f32_kq_kqv.cl
712
639
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_f16_f32_l4_lm.cl
713
640
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_f32_f32_l4_lm.cl
641
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_iq4_nl_f32_l4_lm.cl
642
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q4_0_f32_l4_lm.cl
643
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q4_1_f32_l4_lm.cl
644
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q4_k_f32_l4_lm.cl
645
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q5_0_f32_l4_lm.cl
646
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q5_1_f32_l4_lm.cl
647
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q5_k_f32_l4_lm.cl
648
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q6_k_f32_l4_lm.cl
714
649
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mm_q8_0_f32_l4_lm.cl
715
650
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f16_f16.cl
716
651
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_f16_f32.cl
@@ -722,6 +657,8 @@ files:
722
657
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_id_q4_0_f32_8x_flat.cl
723
658
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_id_q8_0_f32.cl
724
659
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_id_q8_0_f32_flat.cl
660
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_iq4_nl_f32.cl
661
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_iq4_nl_f32_flat.cl
725
662
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_mxfp4_f32.cl
726
663
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_mxfp4_f32_flat.cl
727
664
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32.cl
@@ -729,9 +666,21 @@ files:
729
666
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32_1d_8x_flat.cl
730
667
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32_8x_flat.cl
731
668
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_0_f32_v.cl
732
- - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q6_k.cl
669
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_1_f32.cl
670
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_1_f32_flat.cl
671
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_k_f32.cl
672
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q4_k_f32_flat.cl
673
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_0_f32.cl
674
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_0_f32_flat.cl
675
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_1_f32.cl
676
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_1_f32_flat.cl
677
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_k_f32.cl
678
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q5_k_f32_flat.cl
679
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q6_k_f32.cl
680
+ - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q6_k_f32_flat.cl
733
681
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q8_0_f32.cl
734
682
  - ext/sources/ggml/src/ggml-opencl/kernels/mul_mv_q8_0_f32_flat.cl
683
+ - ext/sources/ggml/src/ggml-opencl/kernels/neg.cl
735
684
  - ext/sources/ggml/src/ggml-opencl/kernels/norm.cl
736
685
  - ext/sources/ggml/src/ggml-opencl/kernels/pad.cl
737
686
  - ext/sources/ggml/src/ggml-opencl/kernels/relu.cl
@@ -747,6 +696,7 @@ files:
747
696
  - ext/sources/ggml/src/ggml-opencl/kernels/softmax_f16.cl
748
697
  - ext/sources/ggml/src/ggml-opencl/kernels/softmax_f32.cl
749
698
  - ext/sources/ggml/src/ggml-opencl/kernels/softplus.cl
699
+ - ext/sources/ggml/src/ggml-opencl/kernels/solve_tri.cl
750
700
  - ext/sources/ggml/src/ggml-opencl/kernels/sqr.cl
751
701
  - ext/sources/ggml/src/ggml-opencl/kernels/sqrt.cl
752
702
  - ext/sources/ggml/src/ggml-opencl/kernels/ssm_conv.cl
@@ -754,13 +704,63 @@ files:
754
704
  - ext/sources/ggml/src/ggml-opencl/kernels/sum_rows.cl
755
705
  - ext/sources/ggml/src/ggml-opencl/kernels/tanh.cl
756
706
  - ext/sources/ggml/src/ggml-opencl/kernels/transpose.cl
707
+ - ext/sources/ggml/src/ggml-opencl/kernels/tri.cl
757
708
  - ext/sources/ggml/src/ggml-opencl/kernels/tsembd.cl
758
709
  - ext/sources/ggml/src/ggml-opencl/kernels/upscale.cl
710
+ - ext/sources/ggml/src/ggml-openvino/.clang-format
711
+ - ext/sources/ggml/src/ggml-openvino/CMakeLists.txt
712
+ - ext/sources/ggml/src/ggml-openvino/ggml-decoder.cpp
713
+ - ext/sources/ggml/src/ggml-openvino/ggml-decoder.h
714
+ - ext/sources/ggml/src/ggml-openvino/ggml-openvino-extra.cpp
715
+ - ext/sources/ggml/src/ggml-openvino/ggml-openvino-extra.h
716
+ - ext/sources/ggml/src/ggml-openvino/ggml-openvino.cpp
717
+ - ext/sources/ggml/src/ggml-openvino/ggml-quants.cpp
718
+ - ext/sources/ggml/src/ggml-openvino/ggml-quants.h
719
+ - ext/sources/ggml/src/ggml-openvino/openvino/decoder.h
720
+ - ext/sources/ggml/src/ggml-openvino/openvino/frontend.cpp
721
+ - ext/sources/ggml/src/ggml-openvino/openvino/frontend.h
722
+ - ext/sources/ggml/src/ggml-openvino/openvino/input_model.cpp
723
+ - ext/sources/ggml/src/ggml-openvino/openvino/input_model.h
724
+ - ext/sources/ggml/src/ggml-openvino/openvino/node_context.h
725
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/cont.cpp
726
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/cpy.cpp
727
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/flash_attn_ext.cpp
728
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/get_rows.cpp
729
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/glu_geglu.cpp
730
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/glu_swiglu.cpp
731
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/mulmat.cpp
732
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/permute.cpp
733
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/reshape.cpp
734
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/rms_norm.cpp
735
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/rope.cpp
736
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/scale.cpp
737
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/set_rows.cpp
738
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/softmax.cpp
739
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/transpose.cpp
740
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/unary_gelu.cpp
741
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/unary_silu.cpp
742
+ - ext/sources/ggml/src/ggml-openvino/openvino/op/view.cpp
743
+ - ext/sources/ggml/src/ggml-openvino/openvino/op_table.cpp
744
+ - ext/sources/ggml/src/ggml-openvino/openvino/op_table.h
745
+ - ext/sources/ggml/src/ggml-openvino/openvino/pass/fuse_to_sdpa.cpp
746
+ - ext/sources/ggml/src/ggml-openvino/openvino/pass/fuse_to_sdpa.h
747
+ - ext/sources/ggml/src/ggml-openvino/openvino/pass/mark_decompression_convert_constant_folding.h
748
+ - ext/sources/ggml/src/ggml-openvino/openvino/pass/squeeze_matmul.cpp
749
+ - ext/sources/ggml/src/ggml-openvino/openvino/pass/squeeze_matmul.h
750
+ - ext/sources/ggml/src/ggml-openvino/openvino/rt_info/weightless_caching_attributes.hpp
751
+ - ext/sources/ggml/src/ggml-openvino/openvino/translate_session.cpp
752
+ - ext/sources/ggml/src/ggml-openvino/openvino/translate_session.h
753
+ - ext/sources/ggml/src/ggml-openvino/openvino/utils.cpp
754
+ - ext/sources/ggml/src/ggml-openvino/openvino/utils.h
755
+ - ext/sources/ggml/src/ggml-openvino/utils.cpp
756
+ - ext/sources/ggml/src/ggml-openvino/utils.h
759
757
  - ext/sources/ggml/src/ggml-opt.cpp
760
758
  - ext/sources/ggml/src/ggml-quants.c
761
759
  - ext/sources/ggml/src/ggml-quants.h
762
760
  - ext/sources/ggml/src/ggml-rpc/CMakeLists.txt
763
761
  - ext/sources/ggml/src/ggml-rpc/ggml-rpc.cpp
762
+ - ext/sources/ggml/src/ggml-rpc/transport.cpp
763
+ - ext/sources/ggml/src/ggml-rpc/transport.h
764
764
  - ext/sources/ggml/src/ggml-sycl/CMakeLists.txt
765
765
  - ext/sources/ggml/src/ggml-sycl/add-id.cpp
766
766
  - ext/sources/ggml/src/ggml-sycl/add-id.hpp
@@ -779,12 +779,28 @@ files:
779
779
  - ext/sources/ggml/src/ggml-sycl/count-equal.hpp
780
780
  - ext/sources/ggml/src/ggml-sycl/cpy.cpp
781
781
  - ext/sources/ggml/src/ggml-sycl/cpy.hpp
782
+ - ext/sources/ggml/src/ggml-sycl/cumsum.cpp
783
+ - ext/sources/ggml/src/ggml-sycl/cumsum.hpp
782
784
  - ext/sources/ggml/src/ggml-sycl/dequantize.hpp
785
+ - ext/sources/ggml/src/ggml-sycl/diag.cpp
786
+ - ext/sources/ggml/src/ggml-sycl/diag.hpp
783
787
  - ext/sources/ggml/src/ggml-sycl/dmmv.cpp
784
788
  - ext/sources/ggml/src/ggml-sycl/dmmv.hpp
785
789
  - ext/sources/ggml/src/ggml-sycl/dpct/helper.hpp
786
790
  - ext/sources/ggml/src/ggml-sycl/element_wise.cpp
787
791
  - ext/sources/ggml/src/ggml-sycl/element_wise.hpp
792
+ - ext/sources/ggml/src/ggml-sycl/fattn-buffers.cpp
793
+ - ext/sources/ggml/src/ggml-sycl/fattn-buffers.hpp
794
+ - ext/sources/ggml/src/ggml-sycl/fattn-common.hpp
795
+ - ext/sources/ggml/src/ggml-sycl/fattn-tile.cpp
796
+ - ext/sources/ggml/src/ggml-sycl/fattn-tile.hpp
797
+ - ext/sources/ggml/src/ggml-sycl/fattn-vec.hpp
798
+ - ext/sources/ggml/src/ggml-sycl/fattn.cpp
799
+ - ext/sources/ggml/src/ggml-sycl/fattn.hpp
800
+ - ext/sources/ggml/src/ggml-sycl/fill.cpp
801
+ - ext/sources/ggml/src/ggml-sycl/fill.hpp
802
+ - ext/sources/ggml/src/ggml-sycl/gated_delta_net.cpp
803
+ - ext/sources/ggml/src/ggml-sycl/gated_delta_net.hpp
788
804
  - ext/sources/ggml/src/ggml-sycl/gemm.hpp
789
805
  - ext/sources/ggml/src/ggml-sycl/getrows.cpp
790
806
  - ext/sources/ggml/src/ggml-sycl/getrows.hpp
@@ -820,17 +836,111 @@ files:
820
836
  - ext/sources/ggml/src/ggml-sycl/set_rows.hpp
821
837
  - ext/sources/ggml/src/ggml-sycl/softmax.cpp
822
838
  - ext/sources/ggml/src/ggml-sycl/softmax.hpp
839
+ - ext/sources/ggml/src/ggml-sycl/solve_tri.cpp
840
+ - ext/sources/ggml/src/ggml-sycl/solve_tri.hpp
823
841
  - ext/sources/ggml/src/ggml-sycl/ssm_conv.cpp
824
842
  - ext/sources/ggml/src/ggml-sycl/ssm_conv.hpp
843
+ - ext/sources/ggml/src/ggml-sycl/ssm_scan.cpp
844
+ - ext/sources/ggml/src/ggml-sycl/ssm_scan.hpp
825
845
  - ext/sources/ggml/src/ggml-sycl/sycl_hw.cpp
826
846
  - ext/sources/ggml/src/ggml-sycl/sycl_hw.hpp
847
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq112-dv112.cpp
848
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq128-dv128.cpp
849
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq256-dv256.cpp
850
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq40-dv40.cpp
851
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq512-dv512.cpp
852
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq576-dv512.cpp
853
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq64-dv64.cpp
854
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq72-dv72.cpp
855
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq80-dv80.cpp
856
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-tile-instance-dkq96-dv96.cpp
857
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-f16.cpp
858
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q4_0.cpp
859
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q4_1.cpp
860
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q5_0.cpp
861
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q5_1.cpp
862
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-f16-q8_0.cpp
863
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-f16.cpp
864
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q4_0.cpp
865
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q4_1.cpp
866
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q5_0.cpp
867
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q5_1.cpp
868
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_0-q8_0.cpp
869
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-f16.cpp
870
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q4_0.cpp
871
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q4_1.cpp
872
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q5_0.cpp
873
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q5_1.cpp
874
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q4_1-q8_0.cpp
875
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-f16.cpp
876
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q4_0.cpp
877
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q4_1.cpp
878
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q5_0.cpp
879
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q5_1.cpp
880
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_0-q8_0.cpp
881
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-f16.cpp
882
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q4_0.cpp
883
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q4_1.cpp
884
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q5_0.cpp
885
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q5_1.cpp
886
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q5_1-q8_0.cpp
887
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-f16.cpp
888
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q4_0.cpp
889
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q4_1.cpp
890
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q5_0.cpp
891
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q5_1.cpp
892
+ - ext/sources/ggml/src/ggml-sycl/template-instances/fattn-vec-instance-q8_0-q8_0.cpp
827
893
  - ext/sources/ggml/src/ggml-sycl/tsembd.cpp
828
894
  - ext/sources/ggml/src/ggml-sycl/tsembd.hpp
895
+ - ext/sources/ggml/src/ggml-sycl/type.hpp
896
+ - ext/sources/ggml/src/ggml-sycl/upscale.cpp
897
+ - ext/sources/ggml/src/ggml-sycl/upscale.hpp
829
898
  - ext/sources/ggml/src/ggml-sycl/vecdotq.hpp
830
899
  - ext/sources/ggml/src/ggml-sycl/wkv.cpp
831
900
  - ext/sources/ggml/src/ggml-sycl/wkv.hpp
832
901
  - ext/sources/ggml/src/ggml-threading.cpp
833
902
  - ext/sources/ggml/src/ggml-threading.h
903
+ - ext/sources/ggml/src/ggml-virtgpu/CMakeLists.txt
904
+ - ext/sources/ggml/src/ggml-virtgpu/apir_cs_ggml-rpc-front.cpp
905
+ - ext/sources/ggml/src/ggml-virtgpu/backend/CMakeLists.txt
906
+ - ext/sources/ggml/src/ggml-virtgpu/backend/apir_cs_ggml-rpc-back.cpp
907
+ - ext/sources/ggml/src/ggml-virtgpu/backend/backend-convert.h
908
+ - ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched-backend.cpp
909
+ - ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched-buffer-type.cpp
910
+ - ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched-buffer.cpp
911
+ - ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched-device.cpp
912
+ - ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched.cpp
913
+ - ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched.gen.h
914
+ - ext/sources/ggml/src/ggml-virtgpu/backend/backend-dispatched.h
915
+ - ext/sources/ggml/src/ggml-virtgpu/backend/backend-virgl-apir.h
916
+ - ext/sources/ggml/src/ggml-virtgpu/backend/backend.cpp
917
+ - ext/sources/ggml/src/ggml-virtgpu/backend/shared/api_remoting.h
918
+ - ext/sources/ggml/src/ggml-virtgpu/backend/shared/apir_backend.gen.h
919
+ - ext/sources/ggml/src/ggml-virtgpu/backend/shared/apir_backend.h
920
+ - ext/sources/ggml/src/ggml-virtgpu/backend/shared/apir_cs.h
921
+ - ext/sources/ggml/src/ggml-virtgpu/backend/shared/apir_cs_ggml.h
922
+ - ext/sources/ggml/src/ggml-virtgpu/backend/shared/apir_cs_rpc.h
923
+ - ext/sources/ggml/src/ggml-virtgpu/ggml-backend-buffer-type.cpp
924
+ - ext/sources/ggml/src/ggml-virtgpu/ggml-backend-buffer.cpp
925
+ - ext/sources/ggml/src/ggml-virtgpu/ggml-backend-device.cpp
926
+ - ext/sources/ggml/src/ggml-virtgpu/ggml-backend-reg.cpp
927
+ - ext/sources/ggml/src/ggml-virtgpu/ggml-backend.cpp
928
+ - ext/sources/ggml/src/ggml-virtgpu/ggml-remoting.h
929
+ - ext/sources/ggml/src/ggml-virtgpu/ggmlremoting_functions.yaml
930
+ - ext/sources/ggml/src/ggml-virtgpu/include/apir_hw.h
931
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu-apir.h
932
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward-backend.cpp
933
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward-buffer-type.cpp
934
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward-buffer.cpp
935
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward-device.cpp
936
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward-impl.h
937
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu-forward.gen.h
938
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu-shm.cpp
939
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu-shm.h
940
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu-utils.cpp
941
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu-utils.h
942
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu.cpp
943
+ - ext/sources/ggml/src/ggml-virtgpu/virtgpu.h
834
944
  - ext/sources/ggml/src/ggml-vulkan/CMakeLists.txt
835
945
  - ext/sources/ggml/src/ggml-vulkan/cmake/host-toolchain.cmake.in
836
946
  - ext/sources/ggml/src/ggml-vulkan/ggml-vulkan.cpp
@@ -875,6 +985,8 @@ files:
875
985
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp
876
986
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_xs.comp
877
987
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_mxfp4.comp
988
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_nvfp4.comp
989
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q1_0.comp
878
990
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q2_k.comp
879
991
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q3_k.comp
880
992
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_0.comp
@@ -888,18 +1000,26 @@ files:
888
1000
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/diag.comp
889
1001
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/diag_mask_inf.comp
890
1002
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/div.comp
1003
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/dot_product_funcs.glsl
1004
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/elu.comp
891
1005
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/exp.comp
892
1006
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/bfloat16.comp
893
1007
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/coopmat.comp
894
1008
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/coopmat2.comp
1009
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/coopmat2_decode_vector.comp
895
1010
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/integer_dot.comp
896
1011
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/fill.comp
897
1012
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp
898
1013
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_base.glsl
899
1014
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp
900
1015
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp
1016
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_dequant.glsl
1017
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_mask_opt.comp
1018
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_mmq_funcs.glsl
901
1019
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_split_k_reduce.comp
902
1020
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/floor.comp
1021
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/fwht.comp
1022
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/gated_delta_net.comp
903
1023
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/geglu.comp
904
1024
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/geglu_erf.comp
905
1025
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/geglu_quick.comp
@@ -973,12 +1093,13 @@ files:
973
1093
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_params.glsl
974
1094
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rope_vision.comp
975
1095
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/round.comp
976
- - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/rte.glsl
977
1096
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp
1097
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sgn.comp
978
1098
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sigmoid.comp
979
1099
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/silu.comp
980
1100
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/silu_back.comp
981
1101
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp
1102
+ - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/snake.comp
982
1103
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp
983
1104
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/soft_max_back.comp
984
1105
  - ext/sources/ggml/src/ggml-vulkan/vulkan-shaders/soft_max_large1.comp
@@ -1015,27 +1136,54 @@ files:
1015
1136
  - ext/sources/ggml/src/ggml-webgpu/ggml-webgpu-shader-lib.hpp
1016
1137
  - ext/sources/ggml/src/ggml-webgpu/ggml-webgpu.cpp
1017
1138
  - ext/sources/ggml/src/ggml-webgpu/pre_wgsl.hpp
1018
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/bin_op.tmpl.wgsl
1019
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/binary_head.tmpl
1139
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/add_id.wgsl
1140
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/argmax.wgsl
1141
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/argsort.wgsl
1142
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/argsort_merge.wgsl
1143
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/binary.wgsl
1020
1144
  - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/common_decls.tmpl
1021
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/cpy.tmpl.wgsl
1022
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/embed_wgsl.py
1145
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/concat.wgsl
1146
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/conv2d.wgsl
1147
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/cpy.wgsl
1148
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/cumsum.wgsl
1023
1149
  - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn.wgsl
1024
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/get_rows.tmpl.wgsl
1025
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/glu.tmpl.wgsl
1150
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_quant_staging.tmpl
1151
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_tile.wgsl
1152
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_vec_blk.wgsl
1153
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_vec_reduce.wgsl
1154
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/flash_attn_vec_split.wgsl
1155
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/gated_delta_net.wgsl
1156
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/get_rows.wgsl
1157
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/glu.wgsl
1158
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/im2col.wgsl
1026
1159
  - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/memset.wgsl
1027
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat.tmpl.wgsl
1028
1160
  - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_decls.tmpl
1029
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_reg_tile.tmpl.wgsl
1030
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_subgroup_matrix.tmpl.wgsl
1031
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_vec.tmpl.wgsl
1032
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/rms_norm.wgsl
1033
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/rope.tmpl.wgsl
1034
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/scale.tmpl.wgsl
1035
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/set_rows.tmpl.wgsl
1161
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_id.wgsl
1162
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_id_gather.wgsl
1163
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_id_vec.wgsl
1164
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_reg_tile.wgsl
1165
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_subgroup_matrix.wgsl
1166
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_vec.wgsl
1167
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_vec_acc.tmpl
1168
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_vec_q_acc.tmpl
1169
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/pad.wgsl
1170
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/quant_inner_loops.tmpl
1171
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/quantize_q8.wgsl
1172
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/repeat.wgsl
1173
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/rms_norm_mul.wgsl
1174
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/rope.wgsl
1175
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/row_norm.wgsl
1176
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/scale.wgsl
1177
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/set.wgsl
1036
1178
  - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/set_rows.wgsl
1037
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/soft_max.tmpl.wgsl
1038
- - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/unary_op.wgsl
1179
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/set_rows_quant.wgsl
1180
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/soft_max.wgsl
1181
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/solve_tri.wgsl
1182
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/ssm_conv.wgsl
1183
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/ssm_scan.wgsl
1184
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/sum_rows.wgsl
1185
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/unary.wgsl
1186
+ - ext/sources/ggml/src/ggml-webgpu/wgsl-shaders/upscale.wgsl
1039
1187
  - ext/sources/ggml/src/ggml-zdnn/CMakeLists.txt
1040
1188
  - ext/sources/ggml/src/ggml-zdnn/common.hpp
1041
1189
  - ext/sources/ggml/src/ggml-zdnn/ggml-zdnn.cpp
@@ -1048,7 +1196,9 @@ files:
1048
1196
  - ext/sources/ggml/src/ggml.c
1049
1197
  - ext/sources/ggml/src/ggml.cpp
1050
1198
  - ext/sources/ggml/src/gguf.cpp
1199
+ - ext/sources/include/parakeet.h
1051
1200
  - ext/sources/include/whisper.h
1201
+ - ext/sources/media/matmul.png
1052
1202
  - ext/sources/src/CMakeLists.txt
1053
1203
  - ext/sources/src/coreml/whisper-compat.h
1054
1204
  - ext/sources/src/coreml/whisper-compat.m
@@ -1060,45 +1210,32 @@ files:
1060
1210
  - ext/sources/src/coreml/whisper-encoder.mm
1061
1211
  - ext/sources/src/openvino/whisper-openvino-encoder.cpp
1062
1212
  - ext/sources/src/openvino/whisper-openvino-encoder.h
1213
+ - ext/sources/src/parakeet-arch.h
1214
+ - ext/sources/src/parakeet.cpp
1063
1215
  - ext/sources/src/whisper-arch.h
1064
1216
  - ext/sources/src/whisper.cpp
1065
- - ext/sources/tests/CMakeLists.txt
1066
- - ext/sources/tests/earnings21/eval.mk
1067
- - ext/sources/tests/earnings21/eval.py
1068
- - ext/sources/tests/earnings21/normalizers/__init__.py
1069
- - ext/sources/tests/earnings21/normalizers/basic.py
1070
- - ext/sources/tests/earnings21/normalizers/english.json
1071
- - ext/sources/tests/earnings21/normalizers/english.py
1072
- - ext/sources/tests/earnings21/requirements.txt
1073
- - ext/sources/tests/en-0-ref.txt
1074
- - ext/sources/tests/en-1-ref.txt
1075
- - ext/sources/tests/en-2-ref.txt
1076
- - ext/sources/tests/es-0-ref.txt
1077
- - ext/sources/tests/librispeech/eval.mk
1078
- - ext/sources/tests/librispeech/eval.py
1079
- - ext/sources/tests/librispeech/normalizers/__init__.py
1080
- - ext/sources/tests/librispeech/normalizers/basic.py
1081
- - ext/sources/tests/librispeech/normalizers/english.json
1082
- - ext/sources/tests/librispeech/normalizers/english.py
1083
- - ext/sources/tests/librispeech/requirements.txt
1084
- - ext/sources/tests/run-tests.sh
1085
- - ext/sources/tests/test-c.c
1086
- - ext/sources/tests/test-vad-full.cpp
1087
- - ext/sources/tests/test-vad.cpp
1088
- - ext/sources/tests/test-whisper.js
1089
1217
  - extsources.rb
1090
- - lib/whisper/context.rb
1218
+ - lib/whisper/log_settable.rb
1091
1219
  - lib/whisper/model/uri.rb
1092
- - lib/whisper/segment.rb
1220
+ - lib/whisper/output.rb
1093
1221
  - sig/whisper.rbs
1094
1222
  - test/helper.rb
1095
1223
  - test/jfk_reader/.gitignore
1096
1224
  - test/jfk_reader/extconf.rb
1097
1225
  - test/jfk_reader/jfk_reader.c
1098
1226
  - test/test_callback.rb
1227
+ - test/test_context_params.rb
1099
1228
  - test/test_error.rb
1100
1229
  - test/test_model.rb
1101
1230
  - test/test_package.rb
1231
+ - test/test_parakeet.rb
1232
+ - test/test_parakeet_callback.rb
1233
+ - test/test_parakeet_context.rb
1234
+ - test/test_parakeet_context_params.rb
1235
+ - test/test_parakeet_model.rb
1236
+ - test/test_parakeet_params.rb
1237
+ - test/test_parakeet_segment.rb
1238
+ - test/test_parakeet_token.rb
1102
1239
  - test/test_params.rb
1103
1240
  - test/test_segment.rb
1104
1241
  - test/test_token.rb
@@ -1122,7 +1259,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
1122
1259
  requirements:
1123
1260
  - - ">="
1124
1261
  - !ruby/object:Gem::Version
1125
- version: 3.1.0
1262
+ version: 3.3.0
1126
1263
  required_rubygems_version: !ruby/object:Gem::Requirement
1127
1264
  requirements:
1128
1265
  - - ">="
@@ -1138,9 +1275,18 @@ test_files:
1138
1275
  - test/jfk_reader/extconf.rb
1139
1276
  - test/jfk_reader/jfk_reader.c
1140
1277
  - test/test_callback.rb
1278
+ - test/test_context_params.rb
1141
1279
  - test/test_error.rb
1142
1280
  - test/test_model.rb
1143
1281
  - test/test_package.rb
1282
+ - test/test_parakeet.rb
1283
+ - test/test_parakeet_callback.rb
1284
+ - test/test_parakeet_context.rb
1285
+ - test/test_parakeet_context_params.rb
1286
+ - test/test_parakeet_model.rb
1287
+ - test/test_parakeet_params.rb
1288
+ - test/test_parakeet_segment.rb
1289
+ - test/test_parakeet_token.rb
1144
1290
  - test/test_params.rb
1145
1291
  - test/test_segment.rb
1146
1292
  - test/test_token.rb