cui-llama.rn 1.6.0 → 1.7.0

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 (285) hide show
  1. package/README.md +35 -7
  2. package/android/src/main/CMakeLists.txt +22 -11
  3. package/android/src/main/java/com/rnllama/LlamaContext.java +42 -6
  4. package/android/src/main/java/com/rnllama/RNLlama.java +139 -4
  5. package/android/src/main/jni.cpp +173 -18
  6. package/android/src/main/jniLibs/arm64-v8a/librnllama.so +0 -0
  7. package/android/src/main/jniLibs/arm64-v8a/librnllama_v8.so +0 -0
  8. package/android/src/main/jniLibs/arm64-v8a/librnllama_v8_2.so +0 -0
  9. package/android/src/main/jniLibs/arm64-v8a/librnllama_v8_2_dotprod.so +0 -0
  10. package/android/src/main/jniLibs/arm64-v8a/librnllama_v8_2_dotprod_i8mm.so +0 -0
  11. package/android/src/main/jniLibs/arm64-v8a/librnllama_v8_2_i8mm.so +0 -0
  12. package/android/src/main/jniLibs/x86_64/librnllama.so +0 -0
  13. package/android/src/main/jniLibs/x86_64/librnllama_x86_64.so +0 -0
  14. package/android/src/newarch/java/com/rnllama/RNLlamaModule.java +24 -4
  15. package/android/src/oldarch/java/com/rnllama/RNLlamaModule.java +22 -2
  16. package/cpp/LICENSE +21 -0
  17. package/cpp/chat.cpp +129 -107
  18. package/cpp/chat.h +2 -0
  19. package/cpp/common.cpp +58 -78
  20. package/cpp/common.h +29 -21
  21. package/cpp/ggml-alloc.c +4 -1
  22. package/cpp/ggml-backend.cpp +9 -5
  23. package/cpp/ggml-backend.h +4 -4
  24. package/cpp/ggml-cpp.h +1 -1
  25. package/cpp/ggml-cpu/amx/amx.cpp +221 -0
  26. package/cpp/ggml-cpu/amx/amx.h +8 -0
  27. package/cpp/ggml-cpu/amx/common.h +91 -0
  28. package/cpp/ggml-cpu/amx/mmq.cpp +2511 -0
  29. package/cpp/ggml-cpu/amx/mmq.h +10 -0
  30. package/{ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers → cpp/ggml-cpu}/binary-ops.h +1 -1
  31. package/cpp/ggml-cpu/common.h +72 -0
  32. package/cpp/{ggml-cpu-aarch64.cpp → ggml-cpu/ggml-cpu-aarch64.cpp} +809 -103
  33. package/cpp/{ggml-cpu-quants.c → ggml-cpu/ggml-cpu-quants.c} +306 -6
  34. package/cpp/{ggml-cpu.c → ggml-cpu/ggml-cpu.c} +114 -55
  35. package/cpp/{ggml-cpu.cpp → ggml-cpu/ggml-cpu.cpp} +32 -16
  36. package/cpp/{ops.cpp → ggml-cpu/ops.cpp} +353 -173
  37. package/{ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers → cpp/ggml-cpu}/ops.h +2 -20
  38. package/cpp/{sgemm.cpp → ggml-cpu/sgemm.cpp} +501 -0
  39. package/{ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers → cpp/ggml-cpu}/simd-mappings.h +7 -3
  40. package/{ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers → cpp/ggml-cpu}/unary-ops.h +1 -1
  41. package/cpp/{vec.cpp → ggml-cpu/vec.cpp} +0 -6
  42. package/{ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers → cpp/ggml-cpu}/vec.h +16 -0
  43. package/cpp/ggml-cpu.h +5 -0
  44. package/cpp/ggml-impl.h +16 -9
  45. package/cpp/ggml-llama-sim.metallib +0 -0
  46. package/cpp/ggml-llama.metallib +0 -0
  47. package/cpp/ggml-metal-impl.h +36 -11
  48. package/cpp/ggml-metal.m +810 -176
  49. package/cpp/ggml-opt.cpp +373 -190
  50. package/cpp/ggml-opt.h +49 -28
  51. package/cpp/ggml-quants.c +0 -6
  52. package/cpp/ggml.c +227 -282
  53. package/cpp/ggml.h +82 -101
  54. package/cpp/gguf.cpp +33 -33
  55. package/cpp/json-schema-to-grammar.cpp +3 -0
  56. package/cpp/llama-adapter.cpp +6 -0
  57. package/cpp/llama-arch.cpp +49 -17
  58. package/cpp/llama-arch.h +9 -0
  59. package/cpp/llama-batch.cpp +8 -2
  60. package/cpp/llama-batch.h +2 -1
  61. package/cpp/llama-chat.cpp +39 -16
  62. package/cpp/llama-chat.h +4 -2
  63. package/cpp/llama-context.cpp +440 -611
  64. package/cpp/llama-context.h +44 -33
  65. package/cpp/llama-cparams.h +1 -0
  66. package/cpp/llama-graph.cpp +214 -291
  67. package/cpp/llama-graph.h +69 -21
  68. package/cpp/llama-hparams.cpp +17 -1
  69. package/cpp/llama-hparams.h +39 -5
  70. package/cpp/llama-kv-cache.cpp +2067 -620
  71. package/cpp/llama-kv-cache.h +410 -108
  72. package/cpp/llama-memory.h +12 -1
  73. package/cpp/llama-model-loader.cpp +24 -15
  74. package/cpp/llama-model-saver.cpp +281 -0
  75. package/cpp/llama-model-saver.h +37 -0
  76. package/cpp/llama-model.cpp +1089 -359
  77. package/cpp/llama-model.h +19 -3
  78. package/cpp/llama-sampling.cpp +20 -7
  79. package/cpp/llama-vocab.cpp +54 -9
  80. package/cpp/llama-vocab.h +6 -0
  81. package/cpp/llama.cpp +14 -0
  82. package/cpp/llama.h +86 -142
  83. package/cpp/minja/chat-template.hpp +9 -5
  84. package/cpp/minja/minja.hpp +69 -36
  85. package/cpp/rn-llama.cpp +602 -190
  86. package/cpp/rn-llama.h +34 -8
  87. package/cpp/sampling.cpp +57 -50
  88. package/cpp/tools/mtmd/clip-impl.h +462 -0
  89. package/cpp/tools/mtmd/clip.cpp +4024 -0
  90. package/cpp/tools/mtmd/clip.h +101 -0
  91. package/cpp/tools/mtmd/miniaudio.h +93468 -0
  92. package/cpp/tools/mtmd/mtmd-audio.cpp +855 -0
  93. package/cpp/tools/mtmd/mtmd-audio.h +62 -0
  94. package/cpp/tools/mtmd/mtmd-helper.cpp +297 -0
  95. package/cpp/tools/mtmd/mtmd.cpp +942 -0
  96. package/cpp/tools/mtmd/mtmd.h +362 -0
  97. package/cpp/tools/mtmd/stb_image.h +7988 -0
  98. package/ios/CMakeLists.txt +20 -10
  99. package/ios/RNLlama.h +6 -0
  100. package/ios/RNLlama.mm +82 -3
  101. package/ios/RNLlamaContext.h +5 -1
  102. package/ios/RNLlamaContext.mm +131 -38
  103. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/chat.h +2 -0
  104. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/common.h +29 -21
  105. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ggml-backend.h +4 -4
  106. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ggml-cpp.h +1 -1
  107. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ggml-cpu.h +5 -0
  108. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ggml-impl.h +16 -9
  109. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ggml-metal-impl.h +36 -11
  110. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ggml-opt.h +49 -28
  111. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ggml.h +82 -101
  112. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-arch.h +9 -0
  113. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-batch.h +2 -1
  114. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-chat.h +4 -2
  115. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-context.h +44 -33
  116. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-cparams.h +1 -0
  117. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-graph.h +69 -21
  118. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-hparams.h +39 -5
  119. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-kv-cache.h +410 -108
  120. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-memory.h +12 -1
  121. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-model-saver.h +37 -0
  122. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-model.h +19 -3
  123. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama-vocab.h +6 -0
  124. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/llama.h +86 -142
  125. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/minja/chat-template.hpp +9 -5
  126. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/minja/minja.hpp +69 -36
  127. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/rn-llama.h +34 -8
  128. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Info.plist +0 -0
  129. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/ggml-llama.metallib +0 -0
  130. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/rnllama +0 -0
  131. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/chat.h +2 -0
  132. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/common.h +29 -21
  133. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-backend.h +4 -4
  134. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpp.h +1 -1
  135. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpu.h +5 -0
  136. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-impl.h +16 -9
  137. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-metal-impl.h +36 -11
  138. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-opt.h +49 -28
  139. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/ggml.h +82 -101
  140. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-arch.h +9 -0
  141. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-batch.h +2 -1
  142. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-chat.h +4 -2
  143. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-context.h +44 -33
  144. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-cparams.h +1 -0
  145. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-graph.h +69 -21
  146. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-hparams.h +39 -5
  147. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-kv-cache.h +410 -108
  148. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-memory.h +12 -1
  149. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-model-saver.h +37 -0
  150. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-model.h +19 -3
  151. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama-vocab.h +6 -0
  152. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/llama.h +86 -142
  153. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/minja/chat-template.hpp +9 -5
  154. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/minja/minja.hpp +69 -36
  155. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/rn-llama.h +34 -8
  156. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Info.plist +0 -0
  157. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/_CodeSignature/CodeResources +1 -1
  158. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/ggml-llama-sim.metallib +0 -0
  159. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/rnllama +0 -0
  160. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/chat.h +2 -0
  161. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/common.h +29 -21
  162. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ggml-backend.h +4 -4
  163. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ggml-cpp.h +1 -1
  164. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ggml-cpu.h +5 -0
  165. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ggml-impl.h +16 -9
  166. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ggml-metal-impl.h +36 -11
  167. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ggml-opt.h +49 -28
  168. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ggml.h +82 -101
  169. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-arch.h +9 -0
  170. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-batch.h +2 -1
  171. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-chat.h +4 -2
  172. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-context.h +44 -33
  173. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-cparams.h +1 -0
  174. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-graph.h +69 -21
  175. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-hparams.h +39 -5
  176. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-kv-cache.h +410 -108
  177. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-memory.h +12 -1
  178. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-model-saver.h +37 -0
  179. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-model.h +19 -3
  180. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama-vocab.h +6 -0
  181. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/llama.h +86 -142
  182. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/minja/chat-template.hpp +9 -5
  183. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/minja/minja.hpp +69 -36
  184. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/rn-llama.h +34 -8
  185. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Info.plist +0 -0
  186. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/ggml-llama.metallib +0 -0
  187. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/rnllama +0 -0
  188. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/chat.h +2 -0
  189. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/common.h +29 -21
  190. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-backend.h +4 -4
  191. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpp.h +1 -1
  192. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpu.h +5 -0
  193. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-impl.h +16 -9
  194. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-metal-impl.h +36 -11
  195. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-opt.h +49 -28
  196. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ggml.h +82 -101
  197. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-arch.h +9 -0
  198. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-batch.h +2 -1
  199. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-chat.h +4 -2
  200. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-context.h +44 -33
  201. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-cparams.h +1 -0
  202. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-graph.h +69 -21
  203. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-hparams.h +39 -5
  204. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-kv-cache.h +410 -108
  205. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-memory.h +12 -1
  206. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-model-saver.h +37 -0
  207. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-model.h +19 -3
  208. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama-vocab.h +6 -0
  209. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/llama.h +86 -142
  210. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/minja/chat-template.hpp +9 -5
  211. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/minja/minja.hpp +69 -36
  212. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/rn-llama.h +34 -8
  213. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Info.plist +0 -0
  214. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/_CodeSignature/CodeResources +1 -1
  215. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/ggml-llama-sim.metallib +0 -0
  216. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/rnllama +0 -0
  217. package/jest/mock.js +33 -7
  218. package/lib/commonjs/NativeRNLlama.js.map +1 -1
  219. package/lib/commonjs/index.js +153 -21
  220. package/lib/commonjs/index.js.map +1 -1
  221. package/lib/module/NativeRNLlama.js.map +1 -1
  222. package/lib/module/index.js +152 -20
  223. package/lib/module/index.js.map +1 -1
  224. package/lib/typescript/NativeRNLlama.d.ts +54 -4
  225. package/lib/typescript/NativeRNLlama.d.ts.map +1 -1
  226. package/lib/typescript/index.d.ts +72 -6
  227. package/lib/typescript/index.d.ts.map +1 -1
  228. package/package.json +1 -1
  229. package/src/NativeRNLlama.ts +72 -4
  230. package/src/index.ts +212 -38
  231. package/cpp/binary-ops.h +0 -16
  232. package/cpp/ops.h +0 -128
  233. package/cpp/simd-mappings.h +0 -888
  234. package/cpp/unary-ops.h +0 -28
  235. package/cpp/vec.h +0 -802
  236. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/binary-ops.h +0 -16
  237. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ggml-cpu-aarch64.h +0 -8
  238. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ggml-cpu-impl.h +0 -512
  239. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ggml-cpu-quants.h +0 -63
  240. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ggml-cpu-traits.h +0 -38
  241. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/ops.h +0 -128
  242. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/sgemm.h +0 -14
  243. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/simd-mappings.h +0 -888
  244. package/ios/rnllama.xcframework/ios-arm64/rnllama.framework/Headers/vec.h +0 -802
  245. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpu-aarch64.h +0 -8
  246. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpu-impl.h +0 -512
  247. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpu-quants.h +0 -63
  248. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpu-traits.h +0 -38
  249. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/sgemm.h +0 -14
  250. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/unary-ops.h +0 -28
  251. package/ios/rnllama.xcframework/ios-arm64_x86_64-simulator/rnllama.framework/Headers/vec.h +0 -802
  252. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/binary-ops.h +0 -16
  253. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ggml-cpu-aarch64.h +0 -8
  254. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ggml-cpu-impl.h +0 -512
  255. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ggml-cpu-quants.h +0 -63
  256. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ggml-cpu-traits.h +0 -38
  257. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/ops.h +0 -128
  258. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/sgemm.h +0 -14
  259. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/simd-mappings.h +0 -888
  260. package/ios/rnllama.xcframework/tvos-arm64/rnllama.framework/Headers/unary-ops.h +0 -28
  261. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/binary-ops.h +0 -16
  262. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpu-aarch64.h +0 -8
  263. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpu-impl.h +0 -512
  264. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpu-quants.h +0 -63
  265. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ggml-cpu-traits.h +0 -38
  266. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/ops.h +0 -128
  267. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/sgemm.h +0 -14
  268. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/simd-mappings.h +0 -888
  269. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/unary-ops.h +0 -28
  270. package/ios/rnllama.xcframework/tvos-arm64_x86_64-simulator/rnllama.framework/Headers/vec.h +0 -802
  271. package/lib/commonjs/chat.js +0 -37
  272. package/lib/commonjs/chat.js.map +0 -1
  273. package/lib/module/chat.js +0 -33
  274. package/lib/module/chat.js.map +0 -1
  275. package/lib/typescript/chat.d.ts +0 -10
  276. package/lib/typescript/chat.d.ts.map +0 -1
  277. package/src/chat.ts +0 -44
  278. /package/cpp/{binary-ops.cpp → ggml-cpu/binary-ops.cpp} +0 -0
  279. /package/cpp/{ggml-cpu-aarch64.h → ggml-cpu/ggml-cpu-aarch64.h} +0 -0
  280. /package/cpp/{ggml-cpu-impl.h → ggml-cpu/ggml-cpu-impl.h} +0 -0
  281. /package/cpp/{ggml-cpu-quants.h → ggml-cpu/ggml-cpu-quants.h} +0 -0
  282. /package/cpp/{ggml-cpu-traits.cpp → ggml-cpu/ggml-cpu-traits.cpp} +0 -0
  283. /package/cpp/{ggml-cpu-traits.h → ggml-cpu/ggml-cpu-traits.h} +0 -0
  284. /package/cpp/{sgemm.h → ggml-cpu/sgemm.h} +0 -0
  285. /package/cpp/{unary-ops.cpp → ggml-cpu/unary-ops.cpp} +0 -0
@@ -207,6 +207,10 @@ typedef struct {
207
207
  float attn_factor;
208
208
  float beta_fast;
209
209
  float beta_slow;
210
+ int32_t sect_0;
211
+ int32_t sect_1;
212
+ int32_t sect_2;
213
+ int32_t sect_3;
210
214
  } lm_ggml_metal_kargs_rope;
211
215
 
212
216
  typedef struct {
@@ -299,21 +303,42 @@ typedef struct {
299
303
  } lm_ggml_metal_kargs_mul_mv_ext;
300
304
 
301
305
  typedef struct {
302
- int32_t nei0;
303
- int32_t nei1;
304
- uint64_t nbi1;
306
+ int32_t ne10;
307
+ int32_t ne11; // n_expert_used (bcast)
308
+ uint64_t nb11;
309
+ uint64_t nb12;
310
+ int32_t neh11; // n_tokens
311
+ uint64_t nbh11;
312
+ int32_t ne20; // n_expert_used
313
+ uint64_t nb21;
314
+ } lm_ggml_metal_kargs_mul_mm_id_map0;
315
+
316
+ typedef struct {
317
+ int32_t ne20; // n_expert_used
318
+ int32_t neh0;
319
+ int32_t neh1;
320
+ uint64_t nbh1;
321
+ uint64_t nbh2;
322
+ int32_t ne0;
323
+ uint64_t nb1;
324
+ uint64_t nb2;
325
+ } lm_ggml_metal_kargs_mul_mm_id_map1;
326
+
327
+ typedef struct {
305
328
  int32_t ne00;
306
329
  int32_t ne02;
307
330
  uint64_t nb01;
308
331
  uint64_t nb02;
309
- int32_t ne11;
310
- int32_t ne12;
311
- int32_t ne13;
312
- uint64_t nb10;
313
- uint64_t nb11;
314
- uint64_t nb12;
315
- int32_t ne0;
316
- int32_t ne1;
332
+ uint64_t nb03;
333
+ int32_t neh12;
334
+ uint64_t nbh10;
335
+ uint64_t nbh11;
336
+ uint64_t nbh12;
337
+ uint64_t nbh13;
338
+ int32_t neh0;
339
+ int32_t neh1;
340
+ int16_t r2;
341
+ int16_t r3;
317
342
  } lm_ggml_metal_kargs_mul_mm_id;
318
343
 
319
344
  typedef struct {