@novastera-oss/llamarn 0.4.0 → 0.4.3-beta4

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 (979) hide show
  1. package/RNLlamaCpp.podspec +4 -1
  2. package/android/CMakeLists.txt +13 -3
  3. package/android/src/main/cpp/include/llama.h +44 -21
  4. package/android/src/main/jniLibs/arm64-v8a/libggml-base.so +0 -0
  5. package/android/src/main/jniLibs/arm64-v8a/libggml-cpu.so +0 -0
  6. package/android/src/main/jniLibs/arm64-v8a/libggml.so +0 -0
  7. package/android/src/main/jniLibs/arm64-v8a/libllama.so +0 -0
  8. package/android/src/main/jniLibs/armeabi-v7a/libggml-base.so +0 -0
  9. package/android/src/main/jniLibs/armeabi-v7a/libggml-cpu.so +0 -0
  10. package/android/src/main/jniLibs/armeabi-v7a/libggml.so +0 -0
  11. package/android/src/main/jniLibs/armeabi-v7a/libllama.so +0 -0
  12. package/android/src/main/jniLibs/x86/libggml-base.so +0 -0
  13. package/android/src/main/jniLibs/x86/libggml-cpu.so +0 -0
  14. package/android/src/main/jniLibs/x86/libggml.so +0 -0
  15. package/android/src/main/jniLibs/x86/libllama.so +0 -0
  16. package/android/src/main/jniLibs/x86_64/libggml-base.so +0 -0
  17. package/android/src/main/jniLibs/x86_64/libggml-cpu.so +0 -0
  18. package/android/src/main/jniLibs/x86_64/libggml.so +0 -0
  19. package/android/src/main/jniLibs/x86_64/libllama.so +0 -0
  20. package/cpp/LlamaCppModel.cpp +2 -10
  21. package/cpp/SystemUtils.cpp +3 -7
  22. package/cpp/build-info.cpp +2 -2
  23. package/cpp/llama.cpp/CMakeLists.txt +12 -0
  24. package/cpp/llama.cpp/CODEOWNERS +116 -10
  25. package/cpp/llama.cpp/CONTRIBUTING.md +30 -3
  26. package/cpp/llama.cpp/README.md +13 -5
  27. package/cpp/llama.cpp/build-xcframework.sh +5 -0
  28. package/cpp/llama.cpp/cmake/riscv64-spacemit-linux-gnu-gcc.cmake +29 -0
  29. package/cpp/llama.cpp/common/CMakeLists.txt +12 -2
  30. package/cpp/llama.cpp/common/arg.cpp +303 -795
  31. package/cpp/llama.cpp/common/arg.h +2 -3
  32. package/cpp/llama.cpp/common/chat-parser-xml-toolcall.cpp +861 -0
  33. package/cpp/llama.cpp/common/chat-parser-xml-toolcall.h +45 -0
  34. package/cpp/llama.cpp/common/chat-parser.cpp +156 -15
  35. package/cpp/llama.cpp/common/chat-parser.h +13 -0
  36. package/cpp/llama.cpp/common/chat.cpp +1147 -88
  37. package/cpp/llama.cpp/common/chat.h +16 -3
  38. package/cpp/llama.cpp/common/common.cpp +70 -15
  39. package/cpp/llama.cpp/common/common.h +57 -19
  40. package/cpp/llama.cpp/common/download.cpp +1072 -0
  41. package/cpp/llama.cpp/common/download.h +55 -0
  42. package/cpp/llama.cpp/common/http.h +73 -0
  43. package/cpp/llama.cpp/common/json-partial.cpp +70 -2
  44. package/cpp/llama.cpp/common/json-schema-to-grammar.cpp +61 -22
  45. package/cpp/llama.cpp/common/json-schema-to-grammar.h +2 -0
  46. package/cpp/llama.cpp/common/log.cpp +59 -2
  47. package/cpp/llama.cpp/common/log.h +12 -4
  48. package/cpp/llama.cpp/common/sampling.cpp +84 -8
  49. package/cpp/llama.cpp/common/sampling.h +3 -1
  50. package/cpp/llama.cpp/common/speculative.cpp +1 -1
  51. package/cpp/llama.cpp/convert_hf_to_gguf.py +1608 -233
  52. package/cpp/llama.cpp/convert_hf_to_gguf_update.py +6 -1
  53. package/cpp/llama.cpp/convert_lora_to_gguf.py +37 -5
  54. package/cpp/llama.cpp/ggml/CMakeLists.txt +47 -28
  55. package/cpp/llama.cpp/ggml/include/ggml-backend.h +19 -1
  56. package/cpp/llama.cpp/ggml/include/ggml-cpu.h +1 -1
  57. package/cpp/llama.cpp/ggml/include/ggml-hexagon.h +19 -0
  58. package/cpp/llama.cpp/ggml/include/ggml-metal.h +1 -6
  59. package/cpp/llama.cpp/ggml/include/ggml-rpc.h +7 -9
  60. package/cpp/llama.cpp/ggml/include/ggml-zdnn.h +2 -1
  61. package/cpp/llama.cpp/ggml/include/ggml.h +199 -6
  62. package/cpp/llama.cpp/ggml/src/CMakeLists.txt +38 -0
  63. package/cpp/llama.cpp/ggml/src/ggml-alloc.c +299 -130
  64. package/cpp/llama.cpp/ggml/src/ggml-backend-impl.h +4 -4
  65. package/cpp/llama.cpp/ggml/src/ggml-backend-reg.cpp +21 -5
  66. package/cpp/llama.cpp/ggml/src/ggml-backend.cpp +99 -2
  67. package/cpp/llama.cpp/ggml/src/ggml-blas/CMakeLists.txt +1 -1
  68. package/cpp/llama.cpp/ggml/src/ggml-blas/ggml-blas.cpp +1 -0
  69. package/cpp/llama.cpp/ggml/src/ggml-cann/acl_tensor.cpp +57 -45
  70. package/cpp/llama.cpp/ggml/src/ggml-cann/acl_tensor.h +138 -47
  71. package/cpp/llama.cpp/ggml/src/ggml-cann/aclnn_ops.cpp +1584 -1773
  72. package/cpp/llama.cpp/ggml/src/ggml-cann/aclnn_ops.h +201 -317
  73. package/cpp/llama.cpp/ggml/src/ggml-cann/common.h +146 -187
  74. package/cpp/llama.cpp/ggml/src/ggml-cann/ggml-cann.cpp +771 -713
  75. package/cpp/llama.cpp/ggml/src/ggml-cpu/CMakeLists.txt +135 -77
  76. package/cpp/llama.cpp/ggml/src/ggml-cpu/amx/amx.cpp +5 -2
  77. package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/arm/quants.c +428 -26
  78. package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/loongarch/quants.c +16 -17
  79. package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/riscv/quants.c +318 -145
  80. package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/s390/cpu-feats.cpp +50 -0
  81. package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/s390/quants.c +155 -60
  82. package/cpp/llama.cpp/ggml/src/ggml-cpu/arch/x86/repack.cpp +8 -8
  83. package/cpp/llama.cpp/ggml/src/ggml-cpu/arch-fallback.h +0 -1
  84. package/cpp/llama.cpp/ggml/src/ggml-cpu/common.h +14 -0
  85. package/cpp/llama.cpp/ggml/src/ggml-cpu/ggml-cpu-impl.h +10 -9
  86. package/cpp/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.c +108 -64
  87. package/cpp/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.cpp +14 -4
  88. package/cpp/llama.cpp/ggml/src/ggml-cpu/kleidiai/kernels.cpp +530 -87
  89. package/cpp/llama.cpp/ggml/src/ggml-cpu/kleidiai/kernels.h +37 -45
  90. package/cpp/llama.cpp/ggml/src/ggml-cpu/kleidiai/kleidiai.cpp +349 -127
  91. package/cpp/llama.cpp/ggml/src/ggml-cpu/ops.cpp +947 -1218
  92. package/cpp/llama.cpp/ggml/src/ggml-cpu/ops.h +5 -4
  93. package/cpp/llama.cpp/ggml/src/ggml-cpu/repack.cpp +143 -29
  94. package/cpp/llama.cpp/ggml/src/ggml-cpu/simd-mappings.h +82 -76
  95. package/cpp/llama.cpp/ggml/src/ggml-cpu/spacemit/ime.cpp +1025 -0
  96. package/cpp/llama.cpp/ggml/src/ggml-cpu/spacemit/ime.h +13 -0
  97. package/cpp/llama.cpp/ggml/src/ggml-cpu/spacemit/ime1_kernels.cpp +3196 -0
  98. package/cpp/llama.cpp/ggml/src/ggml-cpu/spacemit/ime_kernels.h +26 -0
  99. package/cpp/llama.cpp/ggml/src/ggml-cpu/unary-ops.cpp +151 -0
  100. package/cpp/llama.cpp/ggml/src/ggml-cpu/unary-ops.h +7 -0
  101. package/cpp/llama.cpp/ggml/src/ggml-cpu/vec.cpp +233 -28
  102. package/cpp/llama.cpp/ggml/src/ggml-cpu/vec.h +326 -66
  103. package/cpp/llama.cpp/ggml/src/ggml-cuda/CMakeLists.txt +12 -3
  104. package/cpp/llama.cpp/ggml/src/ggml-cuda/argsort.cu +102 -6
  105. package/cpp/llama.cpp/ggml/src/ggml-cuda/binbcast.cu +110 -76
  106. package/cpp/llama.cpp/ggml/src/ggml-cuda/common.cuh +167 -38
  107. package/cpp/llama.cpp/ggml/src/ggml-cuda/conv2d.cu +6 -11
  108. package/cpp/llama.cpp/ggml/src/ggml-cuda/convert.cuh +12 -0
  109. package/cpp/llama.cpp/ggml/src/ggml-cuda/cpy-utils.cuh +1 -1
  110. package/cpp/llama.cpp/ggml/src/ggml-cuda/cpy.cu +245 -151
  111. package/cpp/llama.cpp/ggml/src/ggml-cuda/cpy.cuh +1 -5
  112. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-common.cuh +341 -289
  113. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-tile.cu +49 -0
  114. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-tile.cuh +1233 -0
  115. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-vec.cuh +586 -0
  116. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-wmma-f16.cu +6 -6
  117. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-wmma-f16.cuh +48 -0
  118. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn.cu +123 -220
  119. package/cpp/llama.cpp/ggml/src/ggml-cuda/getrows.cu +41 -39
  120. package/cpp/llama.cpp/ggml/src/ggml-cuda/ggml-cuda.cu +715 -45
  121. package/cpp/llama.cpp/ggml/src/ggml-cuda/im2col.cu +150 -0
  122. package/cpp/llama.cpp/ggml/src/ggml-cuda/im2col.cuh +1 -0
  123. package/cpp/llama.cpp/ggml/src/ggml-cuda/mma.cuh +321 -24
  124. package/cpp/llama.cpp/ggml/src/ggml-cuda/mmf.cu +93 -351
  125. package/cpp/llama.cpp/ggml/src/ggml-cuda/mmf.cuh +828 -1
  126. package/cpp/llama.cpp/ggml/src/ggml-cuda/mmid.cu +164 -0
  127. package/cpp/llama.cpp/ggml/src/ggml-cuda/mmid.cuh +5 -0
  128. package/cpp/llama.cpp/ggml/src/ggml-cuda/mmq.cu +3 -166
  129. package/cpp/llama.cpp/ggml/src/ggml-cuda/mmq.cuh +1 -1
  130. package/cpp/llama.cpp/ggml/src/ggml-cuda/mmvf.cu +371 -78
  131. package/cpp/llama.cpp/ggml/src/ggml-cuda/mmvf.cuh +3 -2
  132. package/cpp/llama.cpp/ggml/src/ggml-cuda/mmvq.cu +279 -147
  133. package/cpp/llama.cpp/ggml/src/ggml-cuda/mmvq.cuh +1 -1
  134. package/cpp/llama.cpp/ggml/src/ggml-cuda/norm.cu +97 -85
  135. package/cpp/llama.cpp/ggml/src/ggml-cuda/pad.cu +46 -23
  136. package/cpp/llama.cpp/ggml/src/ggml-cuda/pad_reflect_1d.cu +63 -54
  137. package/cpp/llama.cpp/ggml/src/ggml-cuda/quantize.cu +12 -10
  138. package/cpp/llama.cpp/ggml/src/ggml-cuda/rope.cu +192 -77
  139. package/cpp/llama.cpp/ggml/src/ggml-cuda/rope.cuh +2 -0
  140. package/cpp/llama.cpp/ggml/src/ggml-cuda/scale.cu +10 -9
  141. package/cpp/llama.cpp/ggml/src/ggml-cuda/set-rows.cu +137 -75
  142. package/cpp/llama.cpp/ggml/src/ggml-cuda/set.cu +39 -0
  143. package/cpp/llama.cpp/ggml/src/ggml-cuda/set.cuh +7 -0
  144. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq112-dv112.cu +5 -0
  145. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq128-dv128.cu +5 -0
  146. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq256-dv256.cu +5 -0
  147. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq40-dv40.cu +5 -0
  148. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq576-dv512.cu +5 -0
  149. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq64-dv64.cu +5 -0
  150. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq72-dv72.cu +5 -0
  151. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq80-dv80.cu +5 -0
  152. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-tile-instance-dkq96-dv96.cu +5 -0
  153. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-f16.cu +7 -0
  154. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-q4_0.cu +7 -0
  155. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-q4_1.cu +7 -0
  156. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-q5_0.cu +7 -0
  157. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-q5_1.cu +7 -0
  158. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-f16-q8_0.cu +7 -0
  159. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-f16.cu +7 -0
  160. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-q4_0.cu +7 -0
  161. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-q4_1.cu +7 -0
  162. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-q5_0.cu +7 -0
  163. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-q5_1.cu +7 -0
  164. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_0-q8_0.cu +7 -0
  165. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-f16.cu +7 -0
  166. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-q4_0.cu +7 -0
  167. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-q4_1.cu +7 -0
  168. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-q5_0.cu +7 -0
  169. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-q5_1.cu +7 -0
  170. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q4_1-q8_0.cu +7 -0
  171. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-f16.cu +7 -0
  172. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-q4_0.cu +7 -0
  173. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-q4_1.cu +7 -0
  174. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-q5_0.cu +7 -0
  175. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-q5_1.cu +7 -0
  176. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_0-q8_0.cu +7 -0
  177. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-f16.cu +7 -0
  178. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-q4_0.cu +7 -0
  179. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-q4_1.cu +7 -0
  180. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-q5_0.cu +7 -0
  181. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-q5_1.cu +7 -0
  182. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q5_1-q8_0.cu +7 -0
  183. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-f16.cu +7 -0
  184. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-q4_0.cu +7 -0
  185. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-q4_1.cu +7 -0
  186. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-q5_0.cu +7 -0
  187. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-q5_1.cu +7 -0
  188. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-instance-q8_0-q8_0.cu +7 -0
  189. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/generate_cu_files.py +40 -19
  190. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_1.cu +5 -0
  191. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_10.cu +5 -0
  192. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_11.cu +5 -0
  193. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_12.cu +5 -0
  194. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_13.cu +5 -0
  195. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_14.cu +5 -0
  196. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_15.cu +5 -0
  197. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_16.cu +5 -0
  198. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_2.cu +5 -0
  199. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_3.cu +5 -0
  200. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_4.cu +5 -0
  201. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_5.cu +5 -0
  202. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_6.cu +5 -0
  203. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_7.cu +5 -0
  204. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_8.cu +5 -0
  205. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/mmf-instance-ncols_9.cu +5 -0
  206. package/cpp/llama.cpp/ggml/src/ggml-cuda/topk-moe.cu +336 -0
  207. package/cpp/llama.cpp/ggml/src/ggml-cuda/topk-moe.cuh +16 -0
  208. package/cpp/llama.cpp/ggml/src/ggml-cuda/tsembd.cu +3 -3
  209. package/cpp/llama.cpp/ggml/src/ggml-cuda/unary.cu +105 -11
  210. package/cpp/llama.cpp/ggml/src/ggml-cuda/unary.cuh +36 -0
  211. package/cpp/llama.cpp/ggml/src/ggml-cuda/upscale.cu +87 -6
  212. package/cpp/llama.cpp/ggml/src/ggml-cuda/vendors/hip.h +28 -12
  213. package/cpp/llama.cpp/ggml/src/ggml-hexagon/CMakeLists.txt +68 -0
  214. package/cpp/llama.cpp/ggml/src/ggml-hexagon/ggml-hexagon.cpp +3807 -0
  215. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/CMakeLists.txt +40 -0
  216. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/act-ops.c +442 -0
  217. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/binary-ops.c +360 -0
  218. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/cmake-toolchain.cmake +157 -0
  219. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/htp-ctx.h +40 -0
  220. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/htp-dma.c +69 -0
  221. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/htp-dma.h +119 -0
  222. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/htp-msg.h +156 -0
  223. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/htp-ops.h +64 -0
  224. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/htp_iface.idl +16 -0
  225. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/hvx-exp.c +93 -0
  226. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/hvx-inverse.c +60 -0
  227. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/hvx-sigmoid.c +49 -0
  228. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/hvx-utils.c +960 -0
  229. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/hvx-utils.h +1032 -0
  230. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/main.c +829 -0
  231. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/matmul-ops.c +2223 -0
  232. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/ops-utils.h +149 -0
  233. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/rope-ops.c +418 -0
  234. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/softmax-ops.c +402 -0
  235. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/unary-ops.c +255 -0
  236. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/worker-pool.c +297 -0
  237. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp/worker-pool.h +57 -0
  238. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp-utils.c +448 -0
  239. package/cpp/llama.cpp/ggml/src/ggml-hexagon/htp-utils.h +220 -0
  240. package/cpp/llama.cpp/ggml/src/ggml-hip/CMakeLists.txt +8 -13
  241. package/cpp/llama.cpp/ggml/src/ggml-impl.h +110 -12
  242. package/cpp/llama.cpp/ggml/src/ggml-metal/CMakeLists.txt +6 -5
  243. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal-common.cpp +446 -0
  244. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal-common.h +52 -0
  245. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal-context.h +33 -0
  246. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal-context.m +599 -0
  247. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal-device.cpp +1662 -0
  248. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal-device.h +251 -0
  249. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal-device.m +1527 -0
  250. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal-impl.h +244 -39
  251. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal-ops.cpp +3844 -0
  252. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal-ops.h +90 -0
  253. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal.cpp +723 -0
  254. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal.metal +3453 -1907
  255. package/cpp/llama.cpp/ggml/src/ggml-musa/CMakeLists.txt +3 -1
  256. package/cpp/llama.cpp/ggml/src/ggml-opencl/CMakeLists.txt +10 -0
  257. package/cpp/llama.cpp/ggml/src/ggml-opencl/ggml-opencl.cpp +1331 -109
  258. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/cvt.cl +126 -0
  259. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/flash_attn_f16.cl +31 -4
  260. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/flash_attn_f32.cl +35 -7
  261. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/flash_attn_f32_f16.cl +31 -4
  262. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/gemm_moe_mxfp4_f32.cl +162 -0
  263. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/gemv_moe_mxfp4_f32.cl +156 -0
  264. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/get_rows.cl +36 -12
  265. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/mul_mm_f16_f32_kq_kqv.cl +273 -0
  266. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/mul_mm_f16_f32_l4_lm.cl +24 -10
  267. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/mul_mm_f32_f32_l4_lm.cl +24 -10
  268. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/mul_mm_q8_0_f32_l4_lm.cl +154 -0
  269. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/mul_mv_id_mxfp4_f32_flat.cl +176 -0
  270. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/mul_mv_id_q8_0_f32.cl +140 -0
  271. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/mul_mv_id_q8_0_f32_flat.cl +222 -0
  272. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/mul_mv_mxfp4_f32_flat.cl +167 -0
  273. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/mul_mv_q8_0_f32.cl +125 -0
  274. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/mul_mv_q8_0_f32_flat.cl +202 -0
  275. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/pad.cl +29 -20
  276. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/rms_norm.cl +25 -10
  277. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/rope.cl +50 -24
  278. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/set_rows.cl +123 -10
  279. package/cpp/llama.cpp/ggml/src/ggml-opencl/kernels/tsembd.cl +2 -2
  280. package/cpp/llama.cpp/ggml/src/ggml-quants.c +1 -0
  281. package/cpp/llama.cpp/ggml/src/ggml-rpc/ggml-rpc.cpp +341 -161
  282. package/cpp/llama.cpp/ggml/src/ggml-sycl/backend.hpp +6 -0
  283. package/cpp/llama.cpp/ggml/src/ggml-sycl/binbcast.cpp +6 -5
  284. package/cpp/llama.cpp/ggml/src/ggml-sycl/common.hpp +74 -15
  285. package/cpp/llama.cpp/ggml/src/ggml-sycl/concat.cpp +50 -30
  286. package/cpp/llama.cpp/ggml/src/ggml-sycl/conv.cpp +10 -4
  287. package/cpp/llama.cpp/ggml/src/ggml-sycl/convert.cpp +166 -99
  288. package/cpp/llama.cpp/ggml/src/ggml-sycl/count-equal.cpp +79 -0
  289. package/cpp/llama.cpp/ggml/src/ggml-sycl/count-equal.hpp +9 -0
  290. package/cpp/llama.cpp/ggml/src/ggml-sycl/cpy.cpp +72 -94
  291. package/cpp/llama.cpp/ggml/src/ggml-sycl/dmmv.cpp +67 -49
  292. package/cpp/llama.cpp/ggml/src/ggml-sycl/dpct/helper.hpp +21 -31
  293. package/cpp/llama.cpp/ggml/src/ggml-sycl/element_wise.cpp +252 -316
  294. package/cpp/llama.cpp/ggml/src/ggml-sycl/element_wise.hpp +6 -2
  295. package/cpp/llama.cpp/ggml/src/ggml-sycl/getrows.cpp +9 -6
  296. package/cpp/llama.cpp/ggml/src/ggml-sycl/ggml-sycl.cpp +359 -142
  297. package/cpp/llama.cpp/ggml/src/ggml-sycl/gla.cpp +2 -2
  298. package/cpp/llama.cpp/ggml/src/ggml-sycl/im2col.cpp +1 -1
  299. package/cpp/llama.cpp/ggml/src/ggml-sycl/mmq.cpp +80 -60
  300. package/cpp/llama.cpp/ggml/src/ggml-sycl/mmvq.cpp +201 -132
  301. package/cpp/llama.cpp/ggml/src/ggml-sycl/norm.cpp +230 -55
  302. package/cpp/llama.cpp/ggml/src/ggml-sycl/norm.hpp +2 -0
  303. package/cpp/llama.cpp/ggml/src/ggml-sycl/pad.cpp +97 -0
  304. package/cpp/llama.cpp/ggml/src/ggml-sycl/pad.hpp +24 -0
  305. package/cpp/llama.cpp/ggml/src/ggml-sycl/pad_reflect_1d.cpp +72 -0
  306. package/cpp/llama.cpp/ggml/src/ggml-sycl/pad_reflect_1d.hpp +8 -0
  307. package/cpp/llama.cpp/ggml/src/ggml-sycl/presets.hpp +2 -0
  308. package/cpp/llama.cpp/ggml/src/ggml-sycl/repeat_back.cpp +76 -0
  309. package/cpp/llama.cpp/ggml/src/ggml-sycl/repeat_back.hpp +8 -0
  310. package/cpp/llama.cpp/ggml/src/ggml-sycl/roll.cpp +122 -0
  311. package/cpp/llama.cpp/ggml/src/ggml-sycl/roll.hpp +20 -0
  312. package/cpp/llama.cpp/ggml/src/ggml-sycl/rope.cpp +50 -41
  313. package/cpp/llama.cpp/ggml/src/ggml-sycl/set.cpp +73 -0
  314. package/cpp/llama.cpp/ggml/src/ggml-sycl/set.hpp +5 -0
  315. package/cpp/llama.cpp/ggml/src/ggml-sycl/set_rows.cpp +45 -36
  316. package/cpp/llama.cpp/ggml/src/ggml-sycl/softmax.cpp +330 -165
  317. package/cpp/llama.cpp/ggml/src/ggml-sycl/softmax.hpp +4 -0
  318. package/cpp/llama.cpp/ggml/src/ggml-sycl/ssm_conv.cpp +127 -0
  319. package/cpp/llama.cpp/ggml/src/ggml-sycl/ssm_conv.hpp +5 -0
  320. package/cpp/llama.cpp/ggml/src/ggml-sycl/tsembd.cpp +12 -6
  321. package/cpp/llama.cpp/ggml/src/ggml-sycl/wkv.cpp +16 -12
  322. package/cpp/llama.cpp/ggml/src/ggml-vulkan/CMakeLists.txt +38 -18
  323. package/cpp/llama.cpp/ggml/src/ggml-vulkan/ggml-vulkan.cpp +4184 -2159
  324. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/abs.comp +21 -0
  325. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/acc.comp +2 -2
  326. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/add.comp +2 -2
  327. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/add1.comp +28 -0
  328. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/add_id.comp +1 -1
  329. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/arange.comp +20 -0
  330. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/argmax.comp +2 -2
  331. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/argsort.comp +33 -26
  332. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/argsort_large.comp +114 -0
  333. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/ceil.comp +22 -0
  334. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/clamp.comp +2 -2
  335. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/concat.comp +2 -2
  336. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/contig_copy.comp +2 -2
  337. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_dw.comp +1 -1
  338. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_mm.comp +53 -30
  339. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/conv_transpose_1d.comp +1 -1
  340. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/copy.comp +2 -2
  341. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/copy_from_quant.comp +3 -3
  342. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/copy_to_quant.comp +13 -6
  343. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/copy_transpose.comp +67 -0
  344. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/cos.comp +2 -2
  345. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/count_equal.comp +2 -2
  346. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_f32.comp +1 -1
  347. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{dequant_funcs.comp → dequant_funcs.glsl} +138 -2
  348. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{dequant_funcs_cm2.comp → dequant_funcs_cm2.glsl} +18 -4
  349. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{dequant_head.comp → dequant_head.glsl} +1 -1
  350. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_m.comp +1 -1
  351. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_s.comp +1 -1
  352. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_s.comp +2 -2
  353. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xs.comp +1 -1
  354. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xxs.comp +3 -2
  355. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_s.comp +7 -6
  356. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_xxs.comp +5 -3
  357. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp +1 -1
  358. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_xs.comp +1 -1
  359. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_mxfp4.comp +3 -3
  360. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q2_k.comp +3 -3
  361. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q3_k.comp +1 -1
  362. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_0.comp +1 -1
  363. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_1.comp +1 -1
  364. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_k.comp +3 -3
  365. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_0.comp +1 -1
  366. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_1.comp +1 -1
  367. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_k.comp +3 -3
  368. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q6_k.comp +1 -1
  369. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q8_0.comp +1 -1
  370. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/diag_mask_inf.comp +1 -1
  371. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/div.comp +2 -2
  372. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/exp.comp +3 -2
  373. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/fill.comp +19 -0
  374. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp +52 -14
  375. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{flash_attn_base.comp → flash_attn_base.glsl} +50 -12
  376. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp +61 -12
  377. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp +54 -12
  378. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_split_k_reduce.comp +5 -1
  379. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/floor.comp +22 -0
  380. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/geglu.comp +2 -2
  381. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/geglu_erf.comp +2 -2
  382. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/geglu_quick.comp +2 -2
  383. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/gelu.comp +2 -2
  384. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/gelu_erf.comp +2 -2
  385. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/gelu_quick.comp +2 -2
  386. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{generic_binary_head.comp → generic_binary_head.glsl} +10 -2
  387. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/get_rows.comp +21 -12
  388. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/get_rows_quant.comp +28 -18
  389. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{glu_head.comp → glu_head.glsl} +1 -1
  390. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/group_norm.comp +2 -2
  391. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/hardsigmoid.comp +22 -0
  392. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/hardswish.comp +22 -0
  393. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/im2col.comp +15 -7
  394. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/im2col_3d.comp +125 -0
  395. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/l2_norm.comp +2 -2
  396. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/leaky_relu.comp +2 -2
  397. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/log.comp +18 -0
  398. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul.comp +2 -2
  399. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp +1 -1
  400. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.glsl +229 -0
  401. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iface.glsl +33 -0
  402. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_m.comp +1 -1
  403. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_s.comp +1 -1
  404. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_s.comp +1 -1
  405. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xs.comp +1 -1
  406. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xxs.comp +1 -1
  407. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_s.comp +1 -1
  408. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_xxs.comp +1 -1
  409. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_nc.comp +9 -7
  410. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_p021.comp +9 -7
  411. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp +3 -5
  412. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp +1 -1
  413. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp +3 -5
  414. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp +3 -5
  415. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp +1 -1
  416. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vecq.comp +140 -0
  417. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp +106 -634
  418. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp +118 -9
  419. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_funcs.glsl +556 -0
  420. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_id_funcs.glsl +70 -0
  421. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq.comp +77 -214
  422. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.glsl +589 -0
  423. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_shmem_types.glsl +78 -0
  424. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/multi_add.comp +97 -13
  425. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/neg.comp +20 -0
  426. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/norm.comp +2 -2
  427. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_adamw.comp +2 -2
  428. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_sgd.comp +1 -1
  429. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/pad.comp +25 -4
  430. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/pool2d.comp +1 -1
  431. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/quantize_q8_1.comp +55 -5
  432. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/reglu.comp +2 -2
  433. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/relu.comp +2 -2
  434. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/repeat.comp +2 -2
  435. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/repeat_back.comp +2 -2
  436. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp +45 -3
  437. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_back.comp +2 -2
  438. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_partials.comp +2 -2
  439. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/roll.comp +2 -2
  440. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_funcs.glsl +227 -0
  441. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.glsl +20 -0
  442. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_multi.comp +5 -52
  443. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_neox.comp +5 -35
  444. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_norm.comp +5 -35
  445. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_params.glsl +27 -0
  446. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_vision.comp +5 -41
  447. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/round.comp +29 -0
  448. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp +2 -2
  449. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/sigmoid.comp +2 -2
  450. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/silu.comp +2 -2
  451. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/silu_back.comp +2 -2
  452. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp +2 -2
  453. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp +1 -1
  454. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/soft_max_back.comp +6 -2
  455. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/softplus.comp +23 -0
  456. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/sqrt.comp +2 -2
  457. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/square.comp +2 -2
  458. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/ssm_conv.comp +44 -0
  459. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/ssm_scan.comp +140 -0
  460. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/step.comp +22 -0
  461. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/sub.comp +2 -2
  462. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/sum_rows.comp +1 -1
  463. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/swiglu.comp +2 -2
  464. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/swiglu_oai.comp +2 -2
  465. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/tanh.comp +2 -2
  466. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/timestep_embedding.comp +5 -4
  467. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/topk_moe.comp +171 -0
  468. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/trunc.comp +22 -0
  469. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{types.comp → types.glsl} +79 -29
  470. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/upscale.comp +36 -12
  471. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp +471 -196
  472. package/cpp/llama.cpp/ggml/src/ggml-webgpu/CMakeLists.txt +8 -0
  473. package/cpp/llama.cpp/ggml/src/ggml-webgpu/ggml-webgpu.cpp +1690 -383
  474. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/bin_op.tmpl.wgsl +188 -0
  475. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/binary_head.tmpl +45 -0
  476. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/common_decls.tmpl +930 -0
  477. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/cpy.tmpl.wgsl +101 -0
  478. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/embed_wgsl.py +57 -10
  479. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/get_rows.tmpl.wgsl +874 -0
  480. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/glu.tmpl.wgsl +323 -0
  481. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat.tmpl.wgsl +25 -912
  482. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_decls.tmpl +97 -0
  483. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_reg_tile.tmpl.wgsl +247 -0
  484. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_subgroup_matrix.tmpl.wgsl +302 -0
  485. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/mul_mat_vec.tmpl.wgsl +267 -0
  486. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/rms_norm.wgsl +123 -0
  487. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/rope.tmpl.wgsl +295 -0
  488. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/scale.tmpl.wgsl +90 -0
  489. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/{set_rows.wgsl → set_rows.tmpl.wgsl} +38 -8
  490. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/soft_max.tmpl.wgsl +345 -0
  491. package/cpp/llama.cpp/ggml/src/ggml-zdnn/common.hpp +59 -0
  492. package/cpp/llama.cpp/ggml/src/ggml-zdnn/ggml-zdnn.cpp +96 -314
  493. package/cpp/llama.cpp/ggml/src/ggml-zdnn/mmf.cpp +80 -0
  494. package/cpp/llama.cpp/ggml/src/ggml-zdnn/mmf.hpp +12 -0
  495. package/cpp/llama.cpp/ggml/src/ggml-zdnn/utils.cpp +79 -0
  496. package/cpp/llama.cpp/ggml/src/ggml-zdnn/utils.hpp +19 -0
  497. package/cpp/llama.cpp/ggml/src/ggml.c +440 -17
  498. package/cpp/llama.cpp/ggml/src/gguf.cpp +104 -29
  499. package/cpp/llama.cpp/gguf-py/gguf/constants.py +363 -13
  500. package/cpp/llama.cpp/gguf-py/gguf/gguf_writer.py +64 -0
  501. package/cpp/llama.cpp/gguf-py/gguf/lazy.py +8 -3
  502. package/cpp/llama.cpp/gguf-py/gguf/scripts/gguf_convert_endian.py +6 -0
  503. package/cpp/llama.cpp/gguf-py/gguf/tensor_mapping.py +156 -18
  504. package/cpp/llama.cpp/gguf-py/gguf/utility.py +80 -0
  505. package/cpp/llama.cpp/gguf-py/gguf/vocab.py +4 -4
  506. package/cpp/llama.cpp/include/llama.h +44 -21
  507. package/cpp/llama.cpp/media/llama1-icon-transparent.png +0 -0
  508. package/cpp/llama.cpp/media/llama1-icon-transparent.svg +77 -0
  509. package/cpp/llama.cpp/media/llama1-icon.png +0 -0
  510. package/cpp/llama.cpp/media/llama1-icon.svg +87 -0
  511. package/cpp/llama.cpp/requirements/requirements-all.txt +2 -0
  512. package/cpp/llama.cpp/requirements/requirements-convert_hf_to_gguf.txt +3 -3
  513. package/cpp/llama.cpp/requirements/requirements-convert_legacy_llama.txt +3 -1
  514. package/cpp/llama.cpp/requirements/requirements-tool_bench.txt +1 -1
  515. package/cpp/llama.cpp/src/CMakeLists.txt +101 -0
  516. package/cpp/llama.cpp/src/llama-adapter.cpp +33 -0
  517. package/cpp/llama.cpp/src/llama-adapter.h +3 -0
  518. package/cpp/llama.cpp/src/llama-arch.cpp +344 -14
  519. package/cpp/llama.cpp/src/llama-arch.h +50 -0
  520. package/cpp/llama.cpp/src/llama-batch.cpp +63 -31
  521. package/cpp/llama.cpp/src/llama-batch.h +13 -2
  522. package/cpp/llama.cpp/src/llama-chat.cpp +85 -3
  523. package/cpp/llama.cpp/src/llama-chat.h +4 -0
  524. package/cpp/llama.cpp/src/llama-context.cpp +300 -45
  525. package/cpp/llama.cpp/src/llama-context.h +16 -6
  526. package/cpp/llama.cpp/src/llama-cparams.h +2 -1
  527. package/cpp/llama.cpp/src/llama-grammar.cpp +17 -9
  528. package/cpp/llama.cpp/src/llama-graph.cpp +226 -64
  529. package/cpp/llama.cpp/src/llama-graph.h +27 -5
  530. package/cpp/llama.cpp/src/llama-hparams.cpp +53 -2
  531. package/cpp/llama.cpp/src/llama-hparams.h +48 -8
  532. package/cpp/llama.cpp/src/llama-impl.cpp +3 -3
  533. package/cpp/llama.cpp/src/llama-impl.h +2 -0
  534. package/cpp/llama.cpp/src/llama-kv-cache-iswa.cpp +13 -3
  535. package/cpp/llama.cpp/src/llama-kv-cache-iswa.h +2 -0
  536. package/cpp/llama.cpp/src/llama-kv-cache.cpp +120 -62
  537. package/cpp/llama.cpp/src/llama-kv-cache.h +13 -4
  538. package/cpp/llama.cpp/src/llama-kv-cells.h +44 -2
  539. package/cpp/llama.cpp/src/llama-memory-hybrid.cpp +19 -9
  540. package/cpp/llama.cpp/src/llama-memory-hybrid.h +2 -0
  541. package/cpp/llama.cpp/src/llama-memory-recurrent.cpp +38 -17
  542. package/cpp/llama.cpp/src/llama-memory-recurrent.h +5 -2
  543. package/cpp/llama.cpp/src/llama-memory.h +3 -0
  544. package/cpp/llama.cpp/src/llama-model-loader.cpp +2 -0
  545. package/cpp/llama.cpp/src/llama-model.cpp +1070 -12614
  546. package/cpp/llama.cpp/src/llama-model.h +40 -4
  547. package/cpp/llama.cpp/src/llama-quant.cpp +14 -6
  548. package/cpp/llama.cpp/src/llama-sampling.cpp +243 -136
  549. package/cpp/llama.cpp/src/llama-vocab.cpp +43 -3
  550. package/cpp/llama.cpp/src/llama-vocab.h +43 -39
  551. package/cpp/llama.cpp/src/llama.cpp +69 -10
  552. package/cpp/llama.cpp/src/models/afmoe.cpp +187 -0
  553. package/cpp/llama.cpp/src/models/apertus.cpp +125 -0
  554. package/cpp/llama.cpp/src/models/arcee.cpp +135 -0
  555. package/cpp/llama.cpp/src/models/arctic.cpp +138 -0
  556. package/cpp/llama.cpp/src/models/arwkv7.cpp +86 -0
  557. package/cpp/llama.cpp/src/models/baichuan.cpp +122 -0
  558. package/cpp/llama.cpp/src/models/bailingmoe.cpp +144 -0
  559. package/cpp/llama.cpp/src/models/bailingmoe2.cpp +135 -0
  560. package/cpp/llama.cpp/src/models/bert.cpp +176 -0
  561. package/cpp/llama.cpp/src/models/bitnet.cpp +160 -0
  562. package/cpp/llama.cpp/src/models/bloom.cpp +101 -0
  563. package/cpp/llama.cpp/src/models/chameleon.cpp +178 -0
  564. package/cpp/llama.cpp/src/models/chatglm.cpp +132 -0
  565. package/cpp/llama.cpp/src/models/codeshell.cpp +111 -0
  566. package/cpp/llama.cpp/src/models/cogvlm.cpp +100 -0
  567. package/cpp/llama.cpp/src/models/cohere2-iswa.cpp +131 -0
  568. package/cpp/llama.cpp/src/models/command-r.cpp +122 -0
  569. package/cpp/llama.cpp/src/models/dbrx.cpp +123 -0
  570. package/cpp/llama.cpp/src/models/deci.cpp +135 -0
  571. package/cpp/llama.cpp/src/models/deepseek.cpp +144 -0
  572. package/cpp/llama.cpp/src/models/deepseek2.cpp +237 -0
  573. package/cpp/llama.cpp/src/models/dots1.cpp +134 -0
  574. package/cpp/llama.cpp/src/models/dream.cpp +105 -0
  575. package/cpp/llama.cpp/src/models/ernie4-5-moe.cpp +150 -0
  576. package/cpp/llama.cpp/src/models/ernie4-5.cpp +110 -0
  577. package/cpp/llama.cpp/src/models/exaone.cpp +114 -0
  578. package/cpp/llama.cpp/src/models/exaone4.cpp +123 -0
  579. package/cpp/llama.cpp/src/models/falcon-h1.cpp +113 -0
  580. package/cpp/llama.cpp/src/models/falcon.cpp +120 -0
  581. package/cpp/llama.cpp/src/models/gemma-embedding.cpp +120 -0
  582. package/cpp/llama.cpp/src/models/gemma.cpp +112 -0
  583. package/cpp/llama.cpp/src/models/gemma2-iswa.cpp +125 -0
  584. package/cpp/llama.cpp/src/models/gemma3-iswa.cpp +131 -0
  585. package/cpp/llama.cpp/src/models/gemma3n-iswa.cpp +377 -0
  586. package/cpp/llama.cpp/src/models/glm4-moe.cpp +153 -0
  587. package/cpp/llama.cpp/src/models/glm4.cpp +127 -0
  588. package/cpp/llama.cpp/src/models/gpt2.cpp +105 -0
  589. package/cpp/llama.cpp/src/models/gptneox.cpp +144 -0
  590. package/cpp/llama.cpp/src/models/granite-hybrid.cpp +196 -0
  591. package/cpp/llama.cpp/src/models/granite.cpp +211 -0
  592. package/cpp/llama.cpp/src/models/graph-context-mamba.cpp +283 -0
  593. package/cpp/llama.cpp/src/models/grok.cpp +159 -0
  594. package/cpp/llama.cpp/src/models/grovemoe.cpp +141 -0
  595. package/cpp/llama.cpp/src/models/hunyuan-dense.cpp +132 -0
  596. package/cpp/llama.cpp/src/models/hunyuan-moe.cpp +154 -0
  597. package/cpp/llama.cpp/src/models/internlm2.cpp +120 -0
  598. package/cpp/llama.cpp/src/models/jais.cpp +86 -0
  599. package/cpp/llama.cpp/src/models/jamba.cpp +106 -0
  600. package/cpp/llama.cpp/src/models/lfm2.cpp +173 -0
  601. package/cpp/llama.cpp/src/models/llada-moe.cpp +122 -0
  602. package/cpp/llama.cpp/src/models/llada.cpp +99 -0
  603. package/cpp/llama.cpp/src/models/llama-iswa.cpp +174 -0
  604. package/cpp/llama.cpp/src/models/llama.cpp +155 -0
  605. package/cpp/llama.cpp/src/models/mamba.cpp +55 -0
  606. package/cpp/llama.cpp/src/models/minicpm3.cpp +199 -0
  607. package/cpp/llama.cpp/src/models/minimax-m2.cpp +124 -0
  608. package/cpp/llama.cpp/src/models/models.h +485 -0
  609. package/cpp/llama.cpp/src/models/mpt.cpp +126 -0
  610. package/cpp/llama.cpp/src/models/nemotron-h.cpp +121 -0
  611. package/cpp/llama.cpp/src/models/nemotron.cpp +122 -0
  612. package/cpp/llama.cpp/src/models/neo-bert.cpp +104 -0
  613. package/cpp/llama.cpp/src/models/olmo.cpp +121 -0
  614. package/cpp/llama.cpp/src/models/olmo2.cpp +150 -0
  615. package/cpp/llama.cpp/src/models/olmoe.cpp +124 -0
  616. package/cpp/llama.cpp/src/models/openai-moe-iswa.cpp +124 -0
  617. package/cpp/llama.cpp/src/models/openelm.cpp +124 -0
  618. package/cpp/llama.cpp/src/models/orion.cpp +123 -0
  619. package/cpp/llama.cpp/src/models/pangu-embedded.cpp +121 -0
  620. package/cpp/llama.cpp/src/models/phi2.cpp +121 -0
  621. package/cpp/llama.cpp/src/models/phi3.cpp +152 -0
  622. package/cpp/llama.cpp/src/models/plamo.cpp +110 -0
  623. package/cpp/llama.cpp/src/models/plamo2.cpp +316 -0
  624. package/cpp/llama.cpp/src/models/plm.cpp +168 -0
  625. package/cpp/llama.cpp/src/models/qwen.cpp +108 -0
  626. package/cpp/llama.cpp/src/models/qwen2.cpp +117 -0
  627. package/cpp/llama.cpp/src/models/qwen2moe.cpp +151 -0
  628. package/cpp/llama.cpp/src/models/qwen2vl.cpp +117 -0
  629. package/cpp/llama.cpp/src/models/qwen3.cpp +117 -0
  630. package/cpp/llama.cpp/src/models/qwen3moe.cpp +124 -0
  631. package/cpp/llama.cpp/src/models/qwen3vl-moe.cpp +149 -0
  632. package/cpp/llama.cpp/src/models/qwen3vl.cpp +141 -0
  633. package/cpp/llama.cpp/src/models/refact.cpp +94 -0
  634. package/cpp/llama.cpp/src/models/rwkv6-base.cpp +162 -0
  635. package/cpp/llama.cpp/src/models/rwkv6.cpp +94 -0
  636. package/cpp/llama.cpp/src/models/rwkv6qwen2.cpp +86 -0
  637. package/cpp/llama.cpp/src/models/rwkv7-base.cpp +135 -0
  638. package/cpp/llama.cpp/src/models/rwkv7.cpp +90 -0
  639. package/cpp/llama.cpp/src/models/seed-oss.cpp +124 -0
  640. package/cpp/llama.cpp/src/models/smallthinker.cpp +120 -0
  641. package/cpp/llama.cpp/src/models/smollm3.cpp +128 -0
  642. package/cpp/llama.cpp/src/models/stablelm.cpp +146 -0
  643. package/cpp/llama.cpp/src/models/starcoder.cpp +100 -0
  644. package/cpp/llama.cpp/src/models/starcoder2.cpp +121 -0
  645. package/cpp/llama.cpp/src/models/t5-dec.cpp +166 -0
  646. package/cpp/llama.cpp/src/models/t5-enc.cpp +96 -0
  647. package/cpp/llama.cpp/src/models/wavtokenizer-dec.cpp +149 -0
  648. package/cpp/llama.cpp/src/models/xverse.cpp +108 -0
  649. package/cpp/llama.cpp/src/unicode.cpp +77 -0
  650. package/cpp/llama.cpp/src/unicode.h +43 -0
  651. package/cpp/llama.cpp/vendor/cpp-httplib/CMakeLists.txt +94 -0
  652. package/cpp/llama.cpp/vendor/cpp-httplib/httplib.cpp +9339 -0
  653. package/cpp/llama.cpp/vendor/cpp-httplib/httplib.h +433 -8222
  654. package/cpp/llama.cpp/vendor/cpp-httplib/patch-boringssl.cmake +6 -0
  655. package/cpp/llama.cpp/vendor/miniaudio/miniaudio.h +4179 -1900
  656. package/cpp/llama.cpp/vendor/minja/chat-template.hpp +9 -2
  657. package/cpp/llama.cpp/vendor/minja/minja.hpp +101 -22
  658. package/cpp/rn-completion.cpp +3 -27
  659. package/ios/include/chat.h +16 -3
  660. package/ios/include/common/minja/chat-template.hpp +9 -2
  661. package/ios/include/common/minja/minja.hpp +101 -22
  662. package/ios/include/common.h +57 -19
  663. package/ios/include/json-schema-to-grammar.h +2 -0
  664. package/ios/include/llama.h +44 -21
  665. package/ios/include/log.h +12 -4
  666. package/ios/include/sampling.h +3 -1
  667. package/ios/libs/llama.xcframework/Info.plist +20 -20
  668. package/ios/libs/llama.xcframework/ios-arm64/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
  669. package/ios/libs/llama.xcframework/ios-arm64/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +6399 -5557
  670. package/ios/libs/llama.xcframework/ios-arm64/llama.framework/Headers/ggml-backend.h +19 -1
  671. package/ios/libs/llama.xcframework/ios-arm64/llama.framework/Headers/ggml-cpu.h +1 -1
  672. package/ios/libs/llama.xcframework/ios-arm64/llama.framework/Headers/ggml-metal.h +1 -6
  673. package/ios/libs/llama.xcframework/ios-arm64/llama.framework/Headers/ggml.h +199 -6
  674. package/ios/libs/llama.xcframework/ios-arm64/llama.framework/Headers/llama.h +44 -21
  675. package/ios/libs/llama.xcframework/ios-arm64/llama.framework/llama +0 -0
  676. package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
  677. package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +6362 -5520
  678. package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/x86_64/llama.yml +4813 -4241
  679. package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/llama.framework/Headers/ggml-backend.h +19 -1
  680. package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/llama.framework/Headers/ggml-cpu.h +1 -1
  681. package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/llama.framework/Headers/ggml-metal.h +1 -6
  682. package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/llama.framework/Headers/ggml.h +199 -6
  683. package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/llama.framework/Headers/llama.h +44 -21
  684. package/ios/libs/llama.xcframework/ios-arm64_x86_64-simulator/llama.framework/llama +0 -0
  685. package/package.json +10 -4
  686. package/cpp/llama.cpp/ggml/src/ggml-cann/Doxyfile +0 -2579
  687. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-tile-f16.cu +0 -371
  688. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-tile-f16.cuh +0 -3
  689. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-tile-f32.cu +0 -379
  690. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-tile-f32.cuh +0 -3
  691. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-vec-f16.cuh +0 -495
  692. package/cpp/llama.cpp/ggml/src/ggml-cuda/fattn-vec-f32.cuh +0 -486
  693. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-f16.cu +0 -5
  694. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q4_0.cu +0 -5
  695. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q4_1.cu +0 -5
  696. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q5_0.cu +0 -5
  697. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q5_1.cu +0 -5
  698. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-f16-q8_0.cu +0 -5
  699. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-f16.cu +0 -5
  700. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q4_0.cu +0 -5
  701. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q4_1.cu +0 -5
  702. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q5_0.cu +0 -5
  703. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q5_1.cu +0 -5
  704. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_0-q8_0.cu +0 -5
  705. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-f16.cu +0 -5
  706. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q4_0.cu +0 -5
  707. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q4_1.cu +0 -5
  708. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q5_0.cu +0 -5
  709. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q5_1.cu +0 -5
  710. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q4_1-q8_0.cu +0 -5
  711. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-f16.cu +0 -5
  712. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q4_0.cu +0 -5
  713. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q4_1.cu +0 -5
  714. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q5_0.cu +0 -5
  715. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q5_1.cu +0 -5
  716. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_0-q8_0.cu +0 -5
  717. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-f16.cu +0 -5
  718. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q4_0.cu +0 -5
  719. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q4_1.cu +0 -5
  720. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q5_0.cu +0 -5
  721. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q5_1.cu +0 -5
  722. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q5_1-q8_0.cu +0 -5
  723. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-f16.cu +0 -5
  724. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q4_0.cu +0 -5
  725. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q4_1.cu +0 -5
  726. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q5_0.cu +0 -5
  727. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q5_1.cu +0 -5
  728. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs128-q8_0-q8_0.cu +0 -5
  729. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs256-f16-f16.cu +0 -5
  730. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-f16.cu +0 -5
  731. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q4_0.cu +0 -5
  732. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q4_1.cu +0 -5
  733. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q5_0.cu +0 -5
  734. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q5_1.cu +0 -5
  735. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f16-instance-hs64-f16-q8_0.cu +0 -5
  736. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-f16.cu +0 -5
  737. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q4_0.cu +0 -5
  738. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q4_1.cu +0 -5
  739. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q5_0.cu +0 -5
  740. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q5_1.cu +0 -5
  741. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-f16-q8_0.cu +0 -5
  742. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-f16.cu +0 -5
  743. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q4_0.cu +0 -5
  744. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q4_1.cu +0 -5
  745. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q5_0.cu +0 -5
  746. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q5_1.cu +0 -5
  747. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_0-q8_0.cu +0 -5
  748. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-f16.cu +0 -5
  749. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q4_0.cu +0 -5
  750. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q4_1.cu +0 -5
  751. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q5_0.cu +0 -5
  752. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q5_1.cu +0 -5
  753. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q4_1-q8_0.cu +0 -5
  754. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-f16.cu +0 -5
  755. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q4_0.cu +0 -5
  756. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q4_1.cu +0 -5
  757. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q5_0.cu +0 -5
  758. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q5_1.cu +0 -5
  759. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_0-q8_0.cu +0 -5
  760. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-f16.cu +0 -5
  761. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q4_0.cu +0 -5
  762. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q4_1.cu +0 -5
  763. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q5_0.cu +0 -5
  764. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q5_1.cu +0 -5
  765. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q5_1-q8_0.cu +0 -5
  766. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-f16.cu +0 -5
  767. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q4_0.cu +0 -5
  768. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q4_1.cu +0 -5
  769. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q5_0.cu +0 -5
  770. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q5_1.cu +0 -5
  771. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs128-q8_0-q8_0.cu +0 -5
  772. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs256-f16-f16.cu +0 -5
  773. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-f16.cu +0 -5
  774. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q4_0.cu +0 -5
  775. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q4_1.cu +0 -5
  776. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q5_0.cu +0 -5
  777. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q5_1.cu +0 -5
  778. package/cpp/llama.cpp/ggml/src/ggml-cuda/template-instances/fattn-vec-f32-instance-hs64-f16-q8_0.cu +0 -5
  779. package/cpp/llama.cpp/ggml/src/ggml-metal/ggml-metal.m +0 -6886
  780. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.comp +0 -154
  781. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.comp +0 -105
  782. package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.comp +0 -55
  783. package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/cpy.wgsl +0 -60
  784. package/cpp/llama.cpp/ggml/src/ggml-zdnn/ggml-zdnn-impl.h +0 -97
  785. package/cpp/llama.cpp/models/ggml-vocab-aquila.gguf +0 -0
  786. package/cpp/llama.cpp/models/ggml-vocab-baichuan.gguf +0 -0
  787. package/cpp/llama.cpp/models/ggml-vocab-bert-bge.gguf +0 -0
  788. package/cpp/llama.cpp/models/ggml-vocab-bert-bge.gguf.inp +0 -112
  789. package/cpp/llama.cpp/models/ggml-vocab-bert-bge.gguf.out +0 -46
  790. package/cpp/llama.cpp/models/ggml-vocab-command-r.gguf +0 -0
  791. package/cpp/llama.cpp/models/ggml-vocab-command-r.gguf.inp +0 -112
  792. package/cpp/llama.cpp/models/ggml-vocab-command-r.gguf.out +0 -46
  793. package/cpp/llama.cpp/models/ggml-vocab-deepseek-coder.gguf +0 -0
  794. package/cpp/llama.cpp/models/ggml-vocab-deepseek-coder.gguf.inp +0 -112
  795. package/cpp/llama.cpp/models/ggml-vocab-deepseek-coder.gguf.out +0 -46
  796. package/cpp/llama.cpp/models/ggml-vocab-deepseek-llm.gguf +0 -0
  797. package/cpp/llama.cpp/models/ggml-vocab-deepseek-llm.gguf.inp +0 -112
  798. package/cpp/llama.cpp/models/ggml-vocab-deepseek-llm.gguf.out +0 -46
  799. package/cpp/llama.cpp/models/ggml-vocab-falcon.gguf +0 -0
  800. package/cpp/llama.cpp/models/ggml-vocab-falcon.gguf.inp +0 -112
  801. package/cpp/llama.cpp/models/ggml-vocab-falcon.gguf.out +0 -46
  802. package/cpp/llama.cpp/models/ggml-vocab-gpt-2.gguf +0 -0
  803. package/cpp/llama.cpp/models/ggml-vocab-gpt-2.gguf.inp +0 -112
  804. package/cpp/llama.cpp/models/ggml-vocab-gpt-2.gguf.out +0 -46
  805. package/cpp/llama.cpp/models/ggml-vocab-gpt-neox.gguf +0 -0
  806. package/cpp/llama.cpp/models/ggml-vocab-llama-bpe.gguf +0 -0
  807. package/cpp/llama.cpp/models/ggml-vocab-llama-bpe.gguf.inp +0 -112
  808. package/cpp/llama.cpp/models/ggml-vocab-llama-bpe.gguf.out +0 -46
  809. package/cpp/llama.cpp/models/ggml-vocab-llama-spm.gguf +0 -0
  810. package/cpp/llama.cpp/models/ggml-vocab-llama-spm.gguf.inp +0 -112
  811. package/cpp/llama.cpp/models/ggml-vocab-llama-spm.gguf.out +0 -46
  812. package/cpp/llama.cpp/models/ggml-vocab-mpt.gguf +0 -0
  813. package/cpp/llama.cpp/models/ggml-vocab-mpt.gguf.inp +0 -112
  814. package/cpp/llama.cpp/models/ggml-vocab-mpt.gguf.out +0 -46
  815. package/cpp/llama.cpp/models/ggml-vocab-nomic-bert-moe.gguf +0 -0
  816. package/cpp/llama.cpp/models/ggml-vocab-phi-3.gguf +0 -0
  817. package/cpp/llama.cpp/models/ggml-vocab-phi-3.gguf.inp +0 -112
  818. package/cpp/llama.cpp/models/ggml-vocab-phi-3.gguf.out +0 -46
  819. package/cpp/llama.cpp/models/ggml-vocab-qwen2.gguf +0 -0
  820. package/cpp/llama.cpp/models/ggml-vocab-qwen2.gguf.inp +0 -112
  821. package/cpp/llama.cpp/models/ggml-vocab-qwen2.gguf.out +0 -46
  822. package/cpp/llama.cpp/models/ggml-vocab-refact.gguf +0 -0
  823. package/cpp/llama.cpp/models/ggml-vocab-refact.gguf.inp +0 -112
  824. package/cpp/llama.cpp/models/ggml-vocab-refact.gguf.out +0 -46
  825. package/cpp/llama.cpp/models/ggml-vocab-starcoder.gguf +0 -0
  826. package/cpp/llama.cpp/models/ggml-vocab-starcoder.gguf.inp +0 -112
  827. package/cpp/llama.cpp/models/ggml-vocab-starcoder.gguf.out +0 -46
  828. package/cpp/llama.cpp/models/templates/ByteDance-Seed-OSS.jinja +0 -171
  829. package/cpp/llama.cpp/models/templates/CohereForAI-c4ai-command-r-plus-tool_use.jinja +0 -202
  830. package/cpp/llama.cpp/models/templates/CohereForAI-c4ai-command-r7b-12-2024-tool_use.jinja +0 -156
  831. package/cpp/llama.cpp/models/templates/Mistral-Small-3.2-24B-Instruct-2506.jinja +0 -124
  832. package/cpp/llama.cpp/models/templates/NousResearch-Hermes-2-Pro-Llama-3-8B-tool_use.jinja +0 -152
  833. package/cpp/llama.cpp/models/templates/NousResearch-Hermes-3-Llama-3.1-8B-tool_use.jinja +0 -152
  834. package/cpp/llama.cpp/models/templates/Qwen-QwQ-32B.jinja +0 -62
  835. package/cpp/llama.cpp/models/templates/Qwen-Qwen2.5-7B-Instruct.jinja +0 -54
  836. package/cpp/llama.cpp/models/templates/Qwen-Qwen3-0.6B.jinja +0 -85
  837. package/cpp/llama.cpp/models/templates/README.md +0 -25
  838. package/cpp/llama.cpp/models/templates/deepseek-ai-DeepSeek-R1-Distill-Llama-8B.jinja +0 -1
  839. package/cpp/llama.cpp/models/templates/deepseek-ai-DeepSeek-R1-Distill-Qwen-32B.jinja +0 -1
  840. package/cpp/llama.cpp/models/templates/fireworks-ai-llama-3-firefunction-v2.jinja +0 -57
  841. package/cpp/llama.cpp/models/templates/google-gemma-2-2b-it.jinja +0 -4
  842. package/cpp/llama.cpp/models/templates/ibm-granite-granite-3.3-2B-Instruct.jinja +0 -59
  843. package/cpp/llama.cpp/models/templates/llama-cpp-deepseek-r1.jinja +0 -76
  844. package/cpp/llama.cpp/models/templates/llama-cpp-rwkv-world.jinja +0 -34
  845. package/cpp/llama.cpp/models/templates/meetkai-functionary-medium-v3.1.jinja +0 -58
  846. package/cpp/llama.cpp/models/templates/meetkai-functionary-medium-v3.2.jinja +0 -287
  847. package/cpp/llama.cpp/models/templates/meta-llama-Llama-3.1-8B-Instruct.jinja +0 -109
  848. package/cpp/llama.cpp/models/templates/meta-llama-Llama-3.2-3B-Instruct.jinja +0 -93
  849. package/cpp/llama.cpp/models/templates/meta-llama-Llama-3.3-70B-Instruct.jinja +0 -109
  850. package/cpp/llama.cpp/models/templates/microsoft-Phi-3.5-mini-instruct.jinja +0 -8
  851. package/cpp/llama.cpp/models/templates/mistralai-Mistral-Nemo-Instruct-2407.jinja +0 -87
  852. package/cpp/llama.cpp/models/templates/moonshotai-Kimi-K2.jinja +0 -43
  853. package/cpp/llama.cpp/models/templates/openai-gpt-oss-120b.jinja +0 -331
  854. package/cpp/llama.cpp/models/templates/unsloth-mistral-Devstral-Small-2507.jinja +0 -105
  855. package/cpp/llama.cpp/prompts/LLM-questions.txt +0 -49
  856. package/cpp/llama.cpp/prompts/alpaca.txt +0 -1
  857. package/cpp/llama.cpp/prompts/assistant.txt +0 -31
  858. package/cpp/llama.cpp/prompts/chat-with-baichuan.txt +0 -4
  859. package/cpp/llama.cpp/prompts/chat-with-bob.txt +0 -7
  860. package/cpp/llama.cpp/prompts/chat-with-qwen.txt +0 -1
  861. package/cpp/llama.cpp/prompts/chat-with-vicuna-v0.txt +0 -7
  862. package/cpp/llama.cpp/prompts/chat-with-vicuna-v1.txt +0 -7
  863. package/cpp/llama.cpp/prompts/chat.txt +0 -28
  864. package/cpp/llama.cpp/prompts/dan-modified.txt +0 -1
  865. package/cpp/llama.cpp/prompts/dan.txt +0 -1
  866. package/cpp/llama.cpp/prompts/mnemonics.txt +0 -93
  867. package/cpp/llama.cpp/prompts/parallel-questions.txt +0 -43
  868. package/cpp/llama.cpp/prompts/reason-act.txt +0 -18
  869. package/ios/libs/llama.xcframework/macos-arm64_x86_64/dSYMs/llama.dSYM/Contents/Info.plist +0 -20
  870. package/ios/libs/llama.xcframework/macos-arm64_x86_64/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
  871. package/ios/libs/llama.xcframework/macos-arm64_x86_64/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +0 -5524
  872. package/ios/libs/llama.xcframework/macos-arm64_x86_64/dSYMs/llama.dSYM/Contents/Resources/Relocations/x86_64/llama.yml +0 -4247
  873. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/ggml-alloc.h +0 -76
  874. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/ggml-backend.h +0 -354
  875. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/ggml-blas.h +0 -25
  876. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/ggml-cpu.h +0 -145
  877. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/ggml-metal.h +0 -66
  878. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/ggml-opt.h +0 -256
  879. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/ggml.h +0 -2492
  880. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/gguf.h +0 -202
  881. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Headers/llama.h +0 -1391
  882. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Modules/module.modulemap +0 -17
  883. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Resources/Info.plist +0 -32
  884. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/ggml-alloc.h +0 -76
  885. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/ggml-backend.h +0 -354
  886. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/ggml-blas.h +0 -25
  887. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/ggml-cpu.h +0 -145
  888. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/ggml-metal.h +0 -66
  889. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/ggml-opt.h +0 -256
  890. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/ggml.h +0 -2492
  891. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/gguf.h +0 -202
  892. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Headers/llama.h +0 -1391
  893. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Modules/module.modulemap +0 -17
  894. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/Resources/Info.plist +0 -32
  895. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/A/llama +0 -0
  896. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/ggml-alloc.h +0 -76
  897. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/ggml-backend.h +0 -354
  898. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/ggml-blas.h +0 -25
  899. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/ggml-cpu.h +0 -145
  900. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/ggml-metal.h +0 -66
  901. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/ggml-opt.h +0 -256
  902. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/ggml.h +0 -2492
  903. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/gguf.h +0 -202
  904. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Headers/llama.h +0 -1391
  905. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Modules/module.modulemap +0 -17
  906. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/Resources/Info.plist +0 -32
  907. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/Versions/Current/llama +0 -0
  908. package/ios/libs/llama.xcframework/macos-arm64_x86_64/llama.framework/llama +0 -0
  909. package/ios/libs/llama.xcframework/tvos-arm64/dSYMs/llama.dSYM/Contents/Info.plist +0 -20
  910. package/ios/libs/llama.xcframework/tvos-arm64/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
  911. package/ios/libs/llama.xcframework/tvos-arm64/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +0 -5561
  912. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/ggml-alloc.h +0 -76
  913. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/ggml-backend.h +0 -354
  914. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/ggml-blas.h +0 -25
  915. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/ggml-cpu.h +0 -145
  916. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/ggml-metal.h +0 -66
  917. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/ggml-opt.h +0 -256
  918. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/ggml.h +0 -2492
  919. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/gguf.h +0 -202
  920. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Headers/llama.h +0 -1391
  921. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Info.plist +0 -35
  922. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/Modules/module.modulemap +0 -17
  923. package/ios/libs/llama.xcframework/tvos-arm64/llama.framework/llama +0 -0
  924. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Info.plist +0 -20
  925. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
  926. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +0 -5524
  927. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/x86_64/llama.yml +0 -4246
  928. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/ggml-alloc.h +0 -76
  929. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/ggml-backend.h +0 -354
  930. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/ggml-blas.h +0 -25
  931. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/ggml-cpu.h +0 -145
  932. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/ggml-metal.h +0 -66
  933. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/ggml-opt.h +0 -256
  934. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/ggml.h +0 -2492
  935. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/gguf.h +0 -202
  936. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Headers/llama.h +0 -1391
  937. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Info.plist +0 -35
  938. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/Modules/module.modulemap +0 -17
  939. package/ios/libs/llama.xcframework/tvos-arm64_x86_64-simulator/llama.framework/llama +0 -0
  940. package/ios/libs/llama.xcframework/xros-arm64/dSYMs/llama.dSYM/Contents/Info.plist +0 -20
  941. package/ios/libs/llama.xcframework/xros-arm64/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
  942. package/ios/libs/llama.xcframework/xros-arm64/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +0 -5558
  943. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/ggml-alloc.h +0 -76
  944. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/ggml-backend.h +0 -354
  945. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/ggml-blas.h +0 -25
  946. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/ggml-cpu.h +0 -145
  947. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/ggml-metal.h +0 -66
  948. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/ggml-opt.h +0 -256
  949. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/ggml.h +0 -2492
  950. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/gguf.h +0 -202
  951. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Headers/llama.h +0 -1391
  952. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Info.plist +0 -32
  953. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/Modules/module.modulemap +0 -17
  954. package/ios/libs/llama.xcframework/xros-arm64/llama.framework/llama +0 -0
  955. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Info.plist +0 -20
  956. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/DWARF/llama +0 -0
  957. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/aarch64/llama.yml +0 -5520
  958. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/dSYMs/llama.dSYM/Contents/Resources/Relocations/x86_64/llama.yml +0 -4243
  959. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/ggml-alloc.h +0 -76
  960. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/ggml-backend.h +0 -354
  961. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/ggml-blas.h +0 -25
  962. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/ggml-cpu.h +0 -145
  963. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/ggml-metal.h +0 -66
  964. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/ggml-opt.h +0 -256
  965. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/ggml.h +0 -2492
  966. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/gguf.h +0 -202
  967. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Headers/llama.h +0 -1391
  968. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Info.plist +0 -32
  969. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/Modules/module.modulemap +0 -17
  970. package/ios/libs/llama.xcframework/xros-arm64_x86_64-simulator/llama.framework/llama +0 -0
  971. /package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{test_bfloat16_support.comp → feature-tests/bfloat16.comp} +0 -0
  972. /package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{test_coopmat_support.comp → feature-tests/coopmat.comp} +0 -0
  973. /package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{test_coopmat2_support.comp → feature-tests/coopmat2.comp} +0 -0
  974. /package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{test_integer_dot_support.comp → feature-tests/integer_dot.comp} +0 -0
  975. /package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{generic_head.comp → generic_head.glsl} +0 -0
  976. /package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{generic_unary_head.comp → generic_unary_head.glsl} +0 -0
  977. /package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{glu_main.comp → glu_main.glsl} +0 -0
  978. /package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{rte.comp → rte.glsl} +0 -0
  979. /package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/{utils.comp → utils.glsl} +0 -0
@@ -0,0 +1,448 @@
1
+
2
+ #pragma clang diagnostic ignored "-Wgnu-anonymous-struct"
3
+ #pragma clang diagnostic ignored "-Wmissing-prototypes"
4
+ #pragma clang diagnostic ignored "-Wsign-compare"
5
+
6
+ #define GGML_COMMON_IMPL_C
7
+ #include "ggml-backend-impl.h"
8
+ #include "ggml-common.h"
9
+ #include "ggml-hexagon.h"
10
+ #include "ggml-impl.h"
11
+
12
+ #include "htp-utils.h"
13
+
14
+ #include <domain.h>
15
+ #include <remote.h>
16
+ #include <stdbool.h>
17
+ #include <stdint.h>
18
+ #include <stdio.h>
19
+ #include <stdlib.h>
20
+ #include <string.h>
21
+
22
+ domain * get_domain(int domain_id) {
23
+ int i = 0;
24
+ int size = sizeof(supported_domains) / sizeof(domain);
25
+
26
+ for (i = 0; i < size; i++) {
27
+ if (supported_domains[i].id == domain_id) {
28
+ return &supported_domains[i];
29
+ }
30
+ }
31
+
32
+ return NULL;
33
+ }
34
+
35
+ bool is_valid_domain_id(int domain_id, int compute_only) {
36
+ int i = 0;
37
+ int size = sizeof(supported_domains) / sizeof(domain);
38
+
39
+ if (compute_only) {
40
+ return is_CDSP(domain_id);
41
+ }
42
+
43
+ for (i = 0; i < size; i++) {
44
+ if (supported_domains[i].id == domain_id) {
45
+ return true;
46
+ }
47
+ }
48
+
49
+ return false;
50
+ }
51
+
52
+ int get_domains_info(char * domain_type, int * num_domains, fastrpc_domain ** domains_info) {
53
+ int nErr = AEE_SUCCESS;
54
+ int ss_info = 0;
55
+ if (domain_type != NULL) {
56
+ if (strcmp(domain_type, "LPASS") == 0) {
57
+ ss_info = FASTRPC_LPASS;
58
+ } else if (strcmp(domain_type, "HPASS") == 0) {
59
+ ss_info = FASTRPC_HPASS;
60
+ } else {
61
+ ss_info = FASTRPC_NSP;
62
+ }
63
+ }
64
+ system_req_payload req = { 0 };
65
+ req.id = FASTRPC_GET_DOMAINS;
66
+ req.sys.domains = NULL;
67
+ fastrpc_domain * domain = NULL;
68
+ if (ss_info != 0) {
69
+ req.sys.flags = DOMAINS_LIST_FLAGS_SET_TYPE(req.sys.flags, ss_info);
70
+ } else {
71
+ req.sys.flags = 0;
72
+ }
73
+ #ifdef _WIN32
74
+ nErr = AEE_EUNSUPPORTED;
75
+ goto bail;
76
+ #endif
77
+ if (remote_system_request) {
78
+ nErr = remote_system_request(&req);
79
+ if (nErr != AEE_SUCCESS) {
80
+ GGML_LOG_ERROR("Failure in remote_system_request call: %d.\n", nErr);
81
+ goto bail;
82
+ }
83
+ // Allocate memory for domain-info array
84
+ req.sys.max_domains = req.sys.num_domains;
85
+ if ((req.sys.domains = calloc(req.sys.num_domains, sizeof(fastrpc_domain))) == NULL) {
86
+ nErr = AEE_ENOMEMORY;
87
+ GGML_LOG_ERROR("Unable to allocate memory for req.sys.domains");
88
+ goto bail;
89
+ }
90
+
91
+ nErr = remote_system_request(&req);
92
+ if (nErr != AEE_SUCCESS) {
93
+ GGML_LOG_ERROR("Failure in remote_system_request call: %d.\n", nErr);
94
+ goto bail;
95
+ }
96
+
97
+ for (int i = 0; i < req.sys.num_domains; i++) {
98
+ // Verify that only requested type domains were returned
99
+ domain = &req.sys.domains[i];
100
+ if (domain->type != ss_info && domain_type != NULL) {
101
+ nErr = -1;
102
+ GGML_LOG_ERROR("Incorrect data received from remote_system_request.\n");
103
+ goto bail;
104
+ }
105
+ }
106
+ *domains_info = req.sys.domains;
107
+ *num_domains = req.sys.num_domains;
108
+ } else {
109
+ nErr = AEE_EUNSUPPORTED;
110
+ goto bail;
111
+ }
112
+ bail:
113
+ if (nErr && !req.sys.domains) {
114
+ free(req.sys.domains);
115
+ }
116
+ return nErr;
117
+ }
118
+
119
+ int get_effective_domain_id(char * domain_name, int session_id, int * effec_domain_id) {
120
+ int err = 0;
121
+ remote_rpc_effective_domain_id_t sess = { 0 };
122
+
123
+ sess.domain_name = domain_name;
124
+ sess.domain_name_len = strlen(domain_name);
125
+ sess.session_id = session_id;
126
+
127
+ err = remote_session_control(FASTRPC_GET_EFFECTIVE_DOMAIN_ID, &sess, sizeof(sess));
128
+ if (err) {
129
+ GGML_LOG_ERROR("Error 0x%x: failed to get effective domain id for %s, session id %d\n", err, sess.domain_name,
130
+ session_id);
131
+ return err;
132
+ }
133
+
134
+ *effec_domain_id = sess.effective_domain_id;
135
+ return err;
136
+ }
137
+
138
+ int get_dsp_support(int * domain) {
139
+ int nErr = AEE_SUCCESS;
140
+ *domain = CDSP_DOMAIN_ID; // DSP domain default value is CDSP_DOMAIN_ID
141
+
142
+ if (remote_handle_control) {
143
+ struct remote_dsp_capability dsp_capability_domain = { CDSP_DOMAIN_ID, DOMAIN_SUPPORT, 0 };
144
+ nErr = remote_handle_control(DSPRPC_GET_DSP_INFO, &dsp_capability_domain, sizeof(struct remote_dsp_capability));
145
+ if ((nErr & 0xFF) == (AEE_EUNSUPPORTEDAPI & 0xFF)) {
146
+ GGML_LOG_ERROR("\nFastRPC Capability API is not supported on this device\n");
147
+ goto bail;
148
+ }
149
+
150
+ if (dsp_capability_domain.capability == 0) {
151
+ dsp_capability_domain.domain = ADSP_DOMAIN_ID; // Check for ADSP support.
152
+ dsp_capability_domain.attribute_ID = DOMAIN_SUPPORT;
153
+ dsp_capability_domain.capability = 0;
154
+ nErr = remote_handle_control(DSPRPC_GET_DSP_INFO, &dsp_capability_domain,
155
+ sizeof(struct remote_dsp_capability));
156
+ if (dsp_capability_domain.capability) {
157
+ *domain = ADSP_DOMAIN_ID; // For targets like Agatti (not having cDSP), domain is ADSP_DOMAIN_ID
158
+ }
159
+ }
160
+
161
+ if (nErr != AEE_SUCCESS) {
162
+ GGML_LOG_ERROR("\nget_dsp_support failed with Error 0x%x\n", nErr);
163
+ goto bail;
164
+ }
165
+ } else {
166
+ nErr = AEE_EUNSUPPORTEDAPI;
167
+ GGML_LOG_ERROR("remote_dsp_capability interface is not supported on this device\n");
168
+ }
169
+
170
+ bail:
171
+ return nErr;
172
+ }
173
+
174
+ int get_vtcm_info(int domain, uint32_t * capability, uint32_t attr) {
175
+ int nErr = AEE_SUCCESS;
176
+ *capability = 0;
177
+
178
+ if (attr == VTCM_PAGE || attr == VTCM_COUNT) {
179
+ } else {
180
+ nErr = AEE_EBADPARM;
181
+ GGML_LOG_ERROR("Unsupported attr. Only VTCM_PAGE and VTCM_COUNT supported\n");
182
+ goto bail;
183
+ }
184
+ if (remote_handle_control) {
185
+ if (domain == ADSP_DOMAIN_ID || domain == CDSP_DOMAIN_ID) {
186
+ /*
187
+ * Query the DSP for VTCM information
188
+ * Since the ADSP does not have a dedicated VTCM, we expect the output to be 0
189
+ */
190
+ struct remote_dsp_capability dsp_capability_vtcm_dsp;
191
+ dsp_capability_vtcm_dsp.domain = (uint32_t) domain;
192
+ dsp_capability_vtcm_dsp.attribute_ID = attr;
193
+ dsp_capability_vtcm_dsp.capability = (uint32_t) 0;
194
+ nErr = remote_handle_control(DSPRPC_GET_DSP_INFO, &dsp_capability_vtcm_dsp,
195
+ sizeof(struct remote_dsp_capability));
196
+ if ((nErr & 0xFF) == (AEE_EUNSUPPORTEDAPI & 0xFF)) {
197
+ GGML_LOG_ERROR("\nFastRPC Capability API is not supported on this device\n");
198
+ GGML_LOG_ERROR("Running the usecase without checking the capability\n");
199
+ nErr = AEE_SUCCESS;
200
+ goto bail;
201
+ } else if (nErr == AEE_SUCCESS) {
202
+ *capability = dsp_capability_vtcm_dsp.capability;
203
+ } else {
204
+ GGML_LOG_ERROR("\nget_vtcm_info failed with Error 0x%x\n", nErr);
205
+ goto bail;
206
+ }
207
+ } else {
208
+ nErr = AEE_EUNSUPPORTED;
209
+ GGML_LOG_ERROR("Unsupported domain %d\n", domain);
210
+ goto bail;
211
+ }
212
+ } else {
213
+ nErr = AEE_EUNSUPPORTEDAPI;
214
+ GGML_LOG_ERROR("remote_dsp_capability interface is not supported on this device\n");
215
+ }
216
+
217
+ bail:
218
+ return nErr;
219
+ }
220
+
221
+ bool is_unsignedpd_supported(int domain_id) {
222
+ int nErr = AEE_SUCCESS;
223
+ if (remote_handle_control) {
224
+ struct remote_dsp_capability dsp_capability_domain = { domain_id, UNSIGNED_PD_SUPPORT, 0 };
225
+ nErr = remote_handle_control(DSPRPC_GET_DSP_INFO, &dsp_capability_domain, sizeof(struct remote_dsp_capability));
226
+ if ((nErr & 0xFF) == (AEE_EUNSUPPORTEDAPI & 0xFF)) {
227
+ GGML_LOG_ERROR("\nFastRPC Capability API is not supported on this device. Falling back to signed pd.\n");
228
+ return false;
229
+ }
230
+ if (nErr) {
231
+ GGML_LOG_ERROR("\nERROR 0x%x: FastRPC Capability API failed. Falling back to signed pd.", nErr);
232
+ return false;
233
+ }
234
+ if (dsp_capability_domain.capability == 1) {
235
+ return true;
236
+ }
237
+ } else {
238
+ nErr = AEE_EUNSUPPORTEDAPI;
239
+ GGML_LOG_ERROR("remote_dsp_capability interface is not supported on this device. Falling back to signed pd.\n");
240
+ return false;
241
+ }
242
+ return false;
243
+ }
244
+
245
+ bool get_unsignedpd_support(void) {
246
+ return is_unsignedpd_supported(CDSP_DOMAIN_ID);
247
+ }
248
+
249
+ bool is_async_fastrpc_supported(int domain) {
250
+ int nErr = AEE_SUCCESS;
251
+ if (remote_handle_control) {
252
+ if (domain == CDSP_DOMAIN_ID) {
253
+ /*
254
+ * Query the DSP for ASYNC_FASTRPC_SUPPORT information
255
+ * Async fastrpc is supported only on CDSP
256
+ */
257
+ struct remote_dsp_capability dsp_capability_async_support;
258
+ dsp_capability_async_support.domain = (uint32_t) domain;
259
+ dsp_capability_async_support.attribute_ID = ASYNC_FASTRPC_SUPPORT;
260
+ dsp_capability_async_support.capability = (uint32_t) 0;
261
+ nErr = remote_handle_control(DSPRPC_GET_DSP_INFO, &dsp_capability_async_support,
262
+ sizeof(struct remote_dsp_capability));
263
+ if ((nErr & 0xFF) == (AEE_EUNSUPPORTEDAPI & 0xFF)) {
264
+ GGML_LOG_ERROR("\nFastRPC Capability API is not supported on this device\n");
265
+ GGML_LOG_ERROR("Running the usecase without checking the capability\n");
266
+ nErr = AEE_SUCCESS;
267
+ goto bail;
268
+ } else if (dsp_capability_async_support.capability == 1) {
269
+ return true;
270
+ }
271
+ if (nErr != AEE_SUCCESS) {
272
+ GGML_LOG_ERROR("\nis_async_fastrpc_supported failed with Error 0x%x\n", nErr);
273
+ goto bail;
274
+ }
275
+ } else {
276
+ nErr = AEE_EUNSUPPORTED;
277
+ GGML_LOG_ERROR("Async fastrpc is not supported on domain %d\n", domain);
278
+ goto bail;
279
+ }
280
+ } else {
281
+ nErr = AEE_EUNSUPPORTEDAPI;
282
+ GGML_LOG_ERROR("remote_dsp_capability interface is not supported on this device\n");
283
+ }
284
+
285
+ bail:
286
+ return false;
287
+ }
288
+
289
+ bool is_status_notification_supported(int domain) {
290
+ int nErr = AEE_SUCCESS;
291
+
292
+ if (remote_handle_control) {
293
+ /*
294
+ * Query the DSP for STATUS_NOTIFICATION_SUPPORT information
295
+ * DSP User PD status notification Support
296
+ */
297
+ struct remote_dsp_capability dsp_capability_status_notification_support;
298
+ dsp_capability_status_notification_support.domain = (uint32_t) domain;
299
+ dsp_capability_status_notification_support.attribute_ID = STATUS_NOTIFICATION_SUPPORT;
300
+ dsp_capability_status_notification_support.capability = (uint32_t) 0;
301
+ nErr = remote_handle_control(DSPRPC_GET_DSP_INFO, &dsp_capability_status_notification_support,
302
+ sizeof(struct remote_dsp_capability));
303
+ if ((nErr & 0xFF) == (AEE_EUNSUPPORTEDAPI & 0xFF)) {
304
+ GGML_LOG_ERROR("\nFastRPC Capability API is not supported on this device\n");
305
+ GGML_LOG_ERROR("Running the usecase without checking the capability\n");
306
+ nErr = AEE_SUCCESS;
307
+ goto bail;
308
+ } else if (dsp_capability_status_notification_support.capability == 1) {
309
+ return true;
310
+ }
311
+ if (nErr != AEE_SUCCESS) {
312
+ GGML_LOG_ERROR("\nis_status_notification_supported failed with Error 0x%x\n", nErr);
313
+ goto bail;
314
+ }
315
+ } else {
316
+ nErr = AEE_EUNSUPPORTEDAPI;
317
+ GGML_LOG_ERROR("remote_dsp_capability interface is not supported on this device\n");
318
+ }
319
+
320
+ bail:
321
+ return false;
322
+ }
323
+
324
+ int get_hmx_support_info(int domain, uint32_t * capability, uint32_t attr) {
325
+ int nErr = AEE_SUCCESS;
326
+ *capability = 0;
327
+
328
+ if (attr != HMX_SUPPORT_SPATIAL && attr != HMX_SUPPORT_DEPTH) {
329
+ nErr = AEE_EBADPARM;
330
+ GGML_LOG_ERROR("Unsupported attr. Only HMX_SUPPORT_SPATIAL and HMX_SUPPORT_DEPTH supported\n");
331
+ goto bail;
332
+ }
333
+ if (remote_handle_control) {
334
+ if (domain == CDSP_DOMAIN_ID) {
335
+ /*
336
+ * Query the DSP for HMX SUPPORT information
337
+ * HMX is supported on CDSP only
338
+ */
339
+ struct remote_dsp_capability dsp_capability_hmx_dsp;
340
+ dsp_capability_hmx_dsp.domain = (uint32_t) domain;
341
+ dsp_capability_hmx_dsp.attribute_ID = attr;
342
+ dsp_capability_hmx_dsp.capability = (uint32_t) 0;
343
+ nErr = remote_handle_control(DSPRPC_GET_DSP_INFO, &dsp_capability_hmx_dsp,
344
+ sizeof(struct remote_dsp_capability));
345
+ if ((nErr & 0xFF) == (AEE_EUNSUPPORTEDAPI & 0xFF)) {
346
+ GGML_LOG_ERROR("\nFastRPC Capability API is not supported on this device\n");
347
+ GGML_LOG_ERROR("Running the usecase without checking the capability\n");
348
+ nErr = AEE_SUCCESS;
349
+ goto bail;
350
+ } else if (nErr == AEE_SUCCESS) {
351
+ *capability = dsp_capability_hmx_dsp.capability;
352
+ } else {
353
+ GGML_LOG_ERROR("\nget_hmx_support_info failed with Error 0x%x\n", nErr);
354
+ goto bail;
355
+ }
356
+ } else {
357
+ nErr = AEE_EUNSUPPORTED;
358
+ GGML_LOG_ERROR("HMX support is not there for domain %d\n", domain);
359
+ goto bail;
360
+ }
361
+ } else {
362
+ nErr = AEE_EUNSUPPORTEDAPI;
363
+ GGML_LOG_ERROR("remote_dsp_capability interface is not supported on this device\n");
364
+ }
365
+
366
+ bail:
367
+ return nErr;
368
+ }
369
+
370
+ int get_hex_arch_ver(int domain, int * arch) {
371
+ if (!remote_handle_control) {
372
+ GGML_LOG_ERROR("ggml-hex: remote_handle_control is not supported on this device\n");
373
+ return AEE_EUNSUPPORTEDAPI;
374
+ }
375
+
376
+ struct remote_dsp_capability arch_ver;
377
+ arch_ver.domain = (uint32_t) domain;
378
+ arch_ver.attribute_ID = ARCH_VER;
379
+ arch_ver.capability = (uint32_t) 0;
380
+
381
+ int err = remote_handle_control(DSPRPC_GET_DSP_INFO, &arch_ver, sizeof(arch_ver));
382
+ if ((err & 0xff) == (AEE_EUNSUPPORTEDAPI & 0xff)) {
383
+ GGML_LOG_ERROR("ggml-hex: FastRPC capability API is not supported on this device\n");
384
+ return AEE_EUNSUPPORTEDAPI;
385
+ }
386
+
387
+ if (err != AEE_SUCCESS) {
388
+ GGML_LOG_ERROR("ggml-hex: FastRPC capability query failed (err %d)\n", err);
389
+ return err;
390
+ }
391
+
392
+ switch (arch_ver.capability & 0xff) {
393
+ case 0x73:
394
+ *arch = 73;
395
+ return 0;
396
+ case 0x75:
397
+ *arch = 75;
398
+ return 0;
399
+ case 0x79:
400
+ *arch = 79;
401
+ return 0;
402
+ case 0x81:
403
+ *arch = 81;
404
+ return 0;
405
+ }
406
+ return -1;
407
+ }
408
+
409
+ int get_hvx_support_info(int domain, uint32_t * capability, uint32_t attr) {
410
+ int nErr = AEE_SUCCESS;
411
+ *capability = 0;
412
+
413
+ if (remote_handle_control) {
414
+ if (domain == CDSP_DOMAIN_ID) {
415
+ /*
416
+ * Query the DSP for HVX SUPPORT information
417
+ * HVX is supported on CDSP only
418
+ */
419
+ struct remote_dsp_capability dsp_capability_hvx_dsp;
420
+ dsp_capability_hvx_dsp.domain = (uint32_t) domain;
421
+ dsp_capability_hvx_dsp.attribute_ID = attr;
422
+ dsp_capability_hvx_dsp.capability = (uint32_t) 0;
423
+ nErr = remote_handle_control(DSPRPC_GET_DSP_INFO, &dsp_capability_hvx_dsp,
424
+ sizeof(struct remote_dsp_capability));
425
+ if ((nErr & 0xFF) == (AEE_EUNSUPPORTEDAPI & 0xFF)) {
426
+ GGML_LOG_ERROR("\nFastRPC Capability API is not supported on this device\n");
427
+ GGML_LOG_ERROR("Running the usecase without checking the capability\n");
428
+ nErr = AEE_SUCCESS;
429
+ goto bail;
430
+ } else if (nErr == AEE_SUCCESS) {
431
+ *capability = dsp_capability_hvx_dsp.capability;
432
+ } else {
433
+ GGML_LOG_ERROR("\nget_hvx_support_info failed with Error 0x%x\n", nErr);
434
+ goto bail;
435
+ }
436
+ } else {
437
+ nErr = AEE_EUNSUPPORTED;
438
+ GGML_LOG_ERROR("HVX support is not available on domain %d\n", domain);
439
+ goto bail;
440
+ }
441
+ } else {
442
+ nErr = AEE_EUNSUPPORTEDAPI;
443
+ GGML_LOG_ERROR("remote_dsp_capability interface is not supported on this device\n");
444
+ }
445
+
446
+ bail:
447
+ return nErr;
448
+ }
@@ -0,0 +1,220 @@
1
+ #ifndef HTP_UTILS_H
2
+ #define HTP_UTILS_H
3
+
4
+ #ifdef __cplusplus
5
+ extern "C" {
6
+ #endif
7
+
8
+ #include <AEEStdErr.h>
9
+ #include <inttypes.h>
10
+ #include <remote.h>
11
+ #include <stdbool.h>
12
+
13
+ /* Offset to differentiate HLOS and Hexagon error codes.
14
+ Stores the value of AEE_EOFFSET for Hexagon. */
15
+ #ifndef DSP_OFFSET
16
+ # define DSP_OFFSET 0x80000400
17
+ #endif
18
+
19
+ /* Errno for connection reset by peer. */
20
+ #ifndef ECONNRESET
21
+ # ifdef __hexagon__
22
+ # define ECONNRESET 104
23
+ # endif
24
+ #endif
25
+
26
+ /* Abstraction of different OS specific sleep APIs.
27
+ SLEEP accepts input in seconds. */
28
+ #ifndef SLEEP
29
+ # ifdef __hexagon__
30
+ # define SLEEP(x) \
31
+ { /* Do nothing for simulator. */ \
32
+ }
33
+ # else
34
+ # ifdef _WINDOWS
35
+ # define SLEEP(x) Sleep(1000 * x) /* Sleep accepts input in milliseconds. */
36
+ # else
37
+ # define SLEEP(x) sleep(x) /* sleep accepts input in seconds. */
38
+ # endif
39
+ # endif
40
+ #endif
41
+
42
+ /* Include windows specific header files. */
43
+ #ifdef _WINDOWS
44
+ # include <sysinfoapi.h>
45
+ # include <windows.h>
46
+ # define _CRT_SECURE_NO_WARNINGS 1
47
+ # define _WINSOCK_DEPRECATED_NO_WARNINGS 1
48
+ /* Including this file for custom implementation of getopt function. */
49
+ # include "getopt_custom.h"
50
+ #endif
51
+
52
+ /* Includes and defines for all HLOS except windows */
53
+ #if !defined(__hexagon__) && !defined(_WINDOWS)
54
+ # include "unistd.h"
55
+
56
+ # include <sys/time.h>
57
+ #endif
58
+
59
+ /* Includes and defines for Hexagon and all HLOS except Windows. */
60
+ #if !defined(_WINDOWS)
61
+ /* Weak reference to remote symbol for compilation. */
62
+ # pragma weak remote_session_control
63
+ # pragma weak remote_handle_control
64
+ # pragma weak remote_handle64_control
65
+ # pragma weak fastrpc_mmap
66
+ # pragma weak fastrpc_munmap
67
+ # pragma weak rpcmem_alloc2
68
+ #endif
69
+
70
+ #if !defined(_WINDOWS)
71
+ # pragma weak remote_system_request
72
+ #endif
73
+ /**
74
+ * Wrapper for FastRPC Capability API: query DSP support.
75
+ *
76
+ * @param[out] domain pointer to supported domain.
77
+ * @return 0 if query is successful.
78
+ * non-zero if error, return value points to the error.
79
+ */
80
+ int get_dsp_support(int * domain);
81
+
82
+ /**
83
+ * Wrapper for FastRPC Capability API: query VTCM information.
84
+ *
85
+ * @param[in] domain value of domain in the queried.
86
+ * @param[out] capability capability value of the attribute queried.
87
+ * @param[in] attr value of the attribute to the queried.
88
+ * @return 0 if query is successful.
89
+ * non-zero if error, return value points to the error.
90
+ */
91
+ int get_vtcm_info(int domain, uint32_t * capability, uint32_t attr);
92
+
93
+ /**
94
+ * Wrapper for FastRPC Capability API: query unsigned pd support on CDSP domain.
95
+ *
96
+ * @return true if unsigned pd is supported.
97
+ * false if unsigned pd is not supported, capability query failed.
98
+ */
99
+
100
+ bool get_unsignedpd_support(void);
101
+
102
+ /**
103
+ * Wrapper for FastRPC Capability API: query unsigned pd support.
104
+ *
105
+ * @param[in] domain value of domain in the queried.
106
+ * @return true if unsigned pd is supported.
107
+ * false if unsigned pd is not supported, capability query failed.
108
+ */
109
+
110
+ bool is_unsignedpd_supported(int domain_id);
111
+
112
+ /**
113
+ * is_valid_domain_id API: query a domain id is valid.
114
+ *
115
+ * @param[in] domain value of domain in the queried.
116
+ * @param[in] compute_only value of domain is only compared with CDSP domains supported by the target when enabled.
117
+ * @return true if value of domain is valid.
118
+ * false if value of domain is not valid.
119
+ */
120
+
121
+ bool is_valid_domain_id(int domain_id, int compute_only);
122
+
123
+ /**
124
+ * get_domain API: get domain struct from domain value.
125
+ *
126
+ * @param[in] domain value of a domain
127
+ * @return Returns domain struct of the domain if it is supported or else
128
+ * returns NULL.
129
+ *
130
+ */
131
+
132
+ domain * get_domain(int domain_id);
133
+
134
+ /**
135
+ * get_domains_info API: get information for all the domains available on the device
136
+ *
137
+ * @param[in] domain_type pointer to domain type
138
+ * @param[in] num_domains pointer to number of domains
139
+ * @param[in] domains_info pointer to save discovered domains information.
140
+ * @return 0 if query is successful.
141
+ * non-zero if error, return value points to the error.
142
+ *
143
+ * It is user's responsibility to free the memory used to store the domains info whose address is present in domains_info before closing the application.
144
+ *
145
+ */
146
+
147
+ int get_domains_info(char * domain_type, int * num_domains, fastrpc_domain ** domains_info);
148
+
149
+ /**
150
+ * get_effective_domain_id API: get effective domain id for given session id
151
+ *
152
+ * @param[in] domain_name pointer to domain name
153
+ * @param[in] session_id
154
+ * @param[in] effec_domain_id pointer to save obtained effective domain id.
155
+ * @return 0 if query is successful.
156
+ * non-zero if error, return value points to the error.
157
+ *
158
+ */
159
+
160
+ int get_effective_domain_id(char * domain_name, int session_id, int * effec_domain_id);
161
+
162
+ /**
163
+ * is_async_fastrpc_supported API: query a domain id has async fastrpc supported or not
164
+ *
165
+ * @param[in] domain_id value of a domain
166
+ * @return Returns true or false stating support of Async FastRPC
167
+ *
168
+ */
169
+
170
+ bool is_async_fastrpc_supported(int domain_id);
171
+
172
+ /**
173
+ * is_status_notification_supported API: query the DSP for STATUS_NOTIFICATION_SUPPORT information
174
+ *
175
+ * @param[in] domain_id value of a domain
176
+ * @return Returns true or false stating status notification support information
177
+ *
178
+ */
179
+ bool is_status_notification_supported(int domain_id);
180
+
181
+ /**
182
+ * get_hmx_support_info API: query the DSP for HMX SUPPORT information
183
+ *
184
+ * @param[in] domain_id value of a domain
185
+ * @param[out] capability capability value of the attribute queried.
186
+ * @param[in] attr value of the attribute to the queried.
187
+ * @return 0 if query is successful.
188
+ * non-zero if error, return value points to the error.
189
+ *
190
+ */
191
+ int get_hmx_support_info(int domain, uint32_t * capability, uint32_t attr);
192
+
193
+ /**
194
+ * get_hex_arch_ver API: query the Hexagon processor architecture version information
195
+ *
196
+ * @param[in] domain_id value of a domain
197
+ * @param[out] Arch version (73, 75, ...)
198
+ * @return 0 if query is successful.
199
+ * non-zero if error, return value points to the error.
200
+ *
201
+ */
202
+ int get_hex_arch_ver(int domain, int * arch);
203
+
204
+ /**
205
+ * get_hvx_support_info API: query the DSP for HVX SUPPORT information
206
+ *
207
+ * @param[in] domain_id value of a domain
208
+ * @param[out] capability capability value of the attribute queried.
209
+ * @param[in] attr value of the attribute to the queried.
210
+ * @return 0 if query is successful.
211
+ * non-zero if error, return value points to the error.
212
+ *
213
+ */
214
+ int get_hvx_support_info(int domain, uint32_t * capability, uint32_t attr);
215
+
216
+ #ifdef __cplusplus
217
+ }
218
+ #endif
219
+
220
+ #endif //DSP_CAPABILITIES_UTILS_H