@fugood/llama.node 0.6.2 → 1.0.0-beta.1

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 (378) hide show
  1. package/CMakeLists.txt +40 -30
  2. package/README.md +4 -1
  3. package/lib/binding.js +41 -29
  4. package/lib/binding.ts +26 -25
  5. package/package.json +45 -10
  6. package/scripts/build.js +47 -0
  7. package/scripts/llama.cpp.patch +109 -0
  8. package/src/anyascii.c +22223 -0
  9. package/src/anyascii.h +42 -0
  10. package/src/tts_utils.cpp +20 -7
  11. package/src/tts_utils.h +2 -0
  12. package/bin/darwin/arm64/llama-node.node +0 -0
  13. package/bin/darwin/x64/llama-node.node +0 -0
  14. package/bin/linux/arm64/llama-node.node +0 -0
  15. package/bin/linux/x64/llama-node.node +0 -0
  16. package/bin/linux-cuda/arm64/llama-node.node +0 -0
  17. package/bin/linux-cuda/x64/llama-node.node +0 -0
  18. package/bin/linux-vulkan/arm64/llama-node.node +0 -0
  19. package/bin/linux-vulkan/x64/llama-node.node +0 -0
  20. package/bin/win32/x64/llama-node.node +0 -0
  21. package/bin/win32/x64/node.lib +0 -0
  22. package/bin/win32-vulkan/arm64/llama-node.node +0 -0
  23. package/bin/win32-vulkan/arm64/node.lib +0 -0
  24. package/bin/win32-vulkan/x64/llama-node.node +0 -0
  25. package/bin/win32-vulkan/x64/node.lib +0 -0
  26. package/patches/node-api-headers+1.1.0.patch +0 -26
  27. package/src/llama.cpp/.github/workflows/build-linux-cross.yml +0 -233
  28. package/src/llama.cpp/.github/workflows/build.yml +0 -1078
  29. package/src/llama.cpp/.github/workflows/close-issue.yml +0 -28
  30. package/src/llama.cpp/.github/workflows/docker.yml +0 -178
  31. package/src/llama.cpp/.github/workflows/editorconfig.yml +0 -29
  32. package/src/llama.cpp/.github/workflows/gguf-publish.yml +0 -44
  33. package/src/llama.cpp/.github/workflows/labeler.yml +0 -17
  34. package/src/llama.cpp/.github/workflows/python-check-requirements.yml +0 -33
  35. package/src/llama.cpp/.github/workflows/python-lint.yml +0 -30
  36. package/src/llama.cpp/.github/workflows/python-type-check.yml +0 -40
  37. package/src/llama.cpp/.github/workflows/release.yml +0 -739
  38. package/src/llama.cpp/.github/workflows/server.yml +0 -237
  39. package/src/llama.cpp/.github/workflows/winget.yml +0 -42
  40. package/src/llama.cpp/cmake/arm64-apple-clang.cmake +0 -16
  41. package/src/llama.cpp/cmake/arm64-windows-llvm.cmake +0 -16
  42. package/src/llama.cpp/cmake/build-info.cmake +0 -64
  43. package/src/llama.cpp/cmake/common.cmake +0 -35
  44. package/src/llama.cpp/cmake/git-vars.cmake +0 -22
  45. package/src/llama.cpp/cmake/x64-windows-llvm.cmake +0 -5
  46. package/src/llama.cpp/common/build-info.cpp.in +0 -4
  47. package/src/llama.cpp/docs/build.md +0 -561
  48. package/src/llama.cpp/examples/CMakeLists.txt +0 -43
  49. package/src/llama.cpp/examples/batched/CMakeLists.txt +0 -5
  50. package/src/llama.cpp/examples/batched/batched.cpp +0 -246
  51. package/src/llama.cpp/examples/chat-13B.bat +0 -57
  52. package/src/llama.cpp/examples/convert-llama2c-to-ggml/CMakeLists.txt +0 -5
  53. package/src/llama.cpp/examples/convert-llama2c-to-ggml/convert-llama2c-to-ggml.cpp +0 -941
  54. package/src/llama.cpp/examples/deprecation-warning/deprecation-warning.cpp +0 -35
  55. package/src/llama.cpp/examples/embedding/CMakeLists.txt +0 -5
  56. package/src/llama.cpp/examples/embedding/embedding.cpp +0 -323
  57. package/src/llama.cpp/examples/eval-callback/CMakeLists.txt +0 -10
  58. package/src/llama.cpp/examples/eval-callback/eval-callback.cpp +0 -194
  59. package/src/llama.cpp/examples/gen-docs/CMakeLists.txt +0 -5
  60. package/src/llama.cpp/examples/gen-docs/gen-docs.cpp +0 -83
  61. package/src/llama.cpp/examples/gguf/CMakeLists.txt +0 -5
  62. package/src/llama.cpp/examples/gguf/gguf.cpp +0 -265
  63. package/src/llama.cpp/examples/gguf-hash/CMakeLists.txt +0 -22
  64. package/src/llama.cpp/examples/gguf-hash/deps/rotate-bits/rotate-bits.h +0 -46
  65. package/src/llama.cpp/examples/gguf-hash/deps/sha1/sha1.c +0 -295
  66. package/src/llama.cpp/examples/gguf-hash/deps/sha1/sha1.h +0 -52
  67. package/src/llama.cpp/examples/gguf-hash/deps/sha256/sha256.c +0 -221
  68. package/src/llama.cpp/examples/gguf-hash/deps/sha256/sha256.h +0 -24
  69. package/src/llama.cpp/examples/gguf-hash/deps/xxhash/xxhash.c +0 -42
  70. package/src/llama.cpp/examples/gguf-hash/deps/xxhash/xxhash.h +0 -7093
  71. package/src/llama.cpp/examples/gguf-hash/gguf-hash.cpp +0 -694
  72. package/src/llama.cpp/examples/gritlm/CMakeLists.txt +0 -5
  73. package/src/llama.cpp/examples/gritlm/gritlm.cpp +0 -229
  74. package/src/llama.cpp/examples/jeopardy/questions.txt +0 -100
  75. package/src/llama.cpp/examples/llama.android/app/build.gradle.kts +0 -65
  76. package/src/llama.cpp/examples/llama.android/build.gradle.kts +0 -6
  77. package/src/llama.cpp/examples/llama.android/llama/build.gradle.kts +0 -71
  78. package/src/llama.cpp/examples/llama.android/llama/src/main/cpp/CMakeLists.txt +0 -53
  79. package/src/llama.cpp/examples/llama.android/llama/src/main/cpp/llama-android.cpp +0 -452
  80. package/src/llama.cpp/examples/llama.android/settings.gradle.kts +0 -18
  81. package/src/llama.cpp/examples/lookahead/CMakeLists.txt +0 -5
  82. package/src/llama.cpp/examples/lookahead/lookahead.cpp +0 -472
  83. package/src/llama.cpp/examples/lookup/CMakeLists.txt +0 -23
  84. package/src/llama.cpp/examples/lookup/lookup-create.cpp +0 -40
  85. package/src/llama.cpp/examples/lookup/lookup-merge.cpp +0 -47
  86. package/src/llama.cpp/examples/lookup/lookup-stats.cpp +0 -157
  87. package/src/llama.cpp/examples/lookup/lookup.cpp +0 -242
  88. package/src/llama.cpp/examples/parallel/CMakeLists.txt +0 -5
  89. package/src/llama.cpp/examples/parallel/parallel.cpp +0 -492
  90. package/src/llama.cpp/examples/passkey/CMakeLists.txt +0 -5
  91. package/src/llama.cpp/examples/passkey/passkey.cpp +0 -277
  92. package/src/llama.cpp/examples/retrieval/CMakeLists.txt +0 -5
  93. package/src/llama.cpp/examples/retrieval/retrieval.cpp +0 -304
  94. package/src/llama.cpp/examples/save-load-state/CMakeLists.txt +0 -5
  95. package/src/llama.cpp/examples/save-load-state/save-load-state.cpp +0 -246
  96. package/src/llama.cpp/examples/simple/CMakeLists.txt +0 -5
  97. package/src/llama.cpp/examples/simple/simple.cpp +0 -206
  98. package/src/llama.cpp/examples/simple-chat/CMakeLists.txt +0 -5
  99. package/src/llama.cpp/examples/simple-chat/simple-chat.cpp +0 -206
  100. package/src/llama.cpp/examples/simple-cmake-pkg/CMakeLists.txt +0 -11
  101. package/src/llama.cpp/examples/speculative/CMakeLists.txt +0 -5
  102. package/src/llama.cpp/examples/speculative/speculative.cpp +0 -644
  103. package/src/llama.cpp/examples/speculative-simple/CMakeLists.txt +0 -5
  104. package/src/llama.cpp/examples/speculative-simple/speculative-simple.cpp +0 -261
  105. package/src/llama.cpp/examples/sycl/CMakeLists.txt +0 -9
  106. package/src/llama.cpp/examples/sycl/build.sh +0 -23
  107. package/src/llama.cpp/examples/sycl/ls-sycl-device.cpp +0 -13
  108. package/src/llama.cpp/examples/sycl/run-llama2.sh +0 -27
  109. package/src/llama.cpp/examples/sycl/run-llama3.sh +0 -28
  110. package/src/llama.cpp/examples/sycl/win-build-sycl.bat +0 -33
  111. package/src/llama.cpp/examples/sycl/win-run-llama2.bat +0 -9
  112. package/src/llama.cpp/examples/sycl/win-run-llama3.bat +0 -9
  113. package/src/llama.cpp/examples/training/CMakeLists.txt +0 -5
  114. package/src/llama.cpp/examples/training/finetune.cpp +0 -96
  115. package/src/llama.cpp/ggml/cmake/GitVars.cmake +0 -22
  116. package/src/llama.cpp/ggml/cmake/common.cmake +0 -26
  117. package/src/llama.cpp/ggml/src/ggml-alloc.c +0 -1042
  118. package/src/llama.cpp/ggml/src/ggml-backend-impl.h +0 -255
  119. package/src/llama.cpp/ggml/src/ggml-backend-reg.cpp +0 -586
  120. package/src/llama.cpp/ggml/src/ggml-backend.cpp +0 -2008
  121. package/src/llama.cpp/ggml/src/ggml-blas/CMakeLists.txt +0 -87
  122. package/src/llama.cpp/ggml/src/ggml-blas/ggml-blas.cpp +0 -517
  123. package/src/llama.cpp/ggml/src/ggml-cann/CMakeLists.txt +0 -74
  124. package/src/llama.cpp/ggml/src/ggml-cann/acl_tensor.cpp +0 -179
  125. package/src/llama.cpp/ggml/src/ggml-cann/acl_tensor.h +0 -258
  126. package/src/llama.cpp/ggml/src/ggml-cann/aclnn_ops.cpp +0 -2863
  127. package/src/llama.cpp/ggml/src/ggml-cann/aclnn_ops.h +0 -1110
  128. package/src/llama.cpp/ggml/src/ggml-cann/common.h +0 -420
  129. package/src/llama.cpp/ggml/src/ggml-cann/ggml-cann.cpp +0 -2570
  130. package/src/llama.cpp/ggml/src/ggml-common.h +0 -1857
  131. package/src/llama.cpp/ggml/src/ggml-cpu/cmake/FindSIMD.cmake +0 -100
  132. package/src/llama.cpp/ggml/src/ggml-cuda/CMakeLists.txt +0 -184
  133. package/src/llama.cpp/ggml/src/ggml-cuda/vendors/cuda.h +0 -15
  134. package/src/llama.cpp/ggml/src/ggml-cuda/vendors/hip.h +0 -243
  135. package/src/llama.cpp/ggml/src/ggml-cuda/vendors/musa.h +0 -140
  136. package/src/llama.cpp/ggml/src/ggml-hip/CMakeLists.txt +0 -131
  137. package/src/llama.cpp/ggml/src/ggml-impl.h +0 -601
  138. package/src/llama.cpp/ggml/src/ggml-kompute/CMakeLists.txt +0 -166
  139. package/src/llama.cpp/ggml/src/ggml-kompute/ggml-kompute.cpp +0 -2251
  140. package/src/llama.cpp/ggml/src/ggml-metal/CMakeLists.txt +0 -120
  141. package/src/llama.cpp/ggml/src/ggml-metal/ggml-metal-impl.h +0 -622
  142. package/src/llama.cpp/ggml/src/ggml-musa/CMakeLists.txt +0 -113
  143. package/src/llama.cpp/ggml/src/ggml-opencl/CMakeLists.txt +0 -96
  144. package/src/llama.cpp/ggml/src/ggml-opencl/ggml-opencl.cpp +0 -5124
  145. package/src/llama.cpp/ggml/src/ggml-opt.cpp +0 -1037
  146. package/src/llama.cpp/ggml/src/ggml-quants.c +0 -5232
  147. package/src/llama.cpp/ggml/src/ggml-quants.h +0 -100
  148. package/src/llama.cpp/ggml/src/ggml-rpc/CMakeLists.txt +0 -9
  149. package/src/llama.cpp/ggml/src/ggml-rpc/ggml-rpc.cpp +0 -1813
  150. package/src/llama.cpp/ggml/src/ggml-sycl/CMakeLists.txt +0 -189
  151. package/src/llama.cpp/ggml/src/ggml-sycl/backend.hpp +0 -37
  152. package/src/llama.cpp/ggml/src/ggml-sycl/binbcast.cpp +0 -239
  153. package/src/llama.cpp/ggml/src/ggml-sycl/binbcast.hpp +0 -39
  154. package/src/llama.cpp/ggml/src/ggml-sycl/common.cpp +0 -83
  155. package/src/llama.cpp/ggml/src/ggml-sycl/common.hpp +0 -493
  156. package/src/llama.cpp/ggml/src/ggml-sycl/concat.cpp +0 -197
  157. package/src/llama.cpp/ggml/src/ggml-sycl/concat.hpp +0 -20
  158. package/src/llama.cpp/ggml/src/ggml-sycl/conv.cpp +0 -100
  159. package/src/llama.cpp/ggml/src/ggml-sycl/conv.hpp +0 -20
  160. package/src/llama.cpp/ggml/src/ggml-sycl/convert.cpp +0 -623
  161. package/src/llama.cpp/ggml/src/ggml-sycl/convert.hpp +0 -34
  162. package/src/llama.cpp/ggml/src/ggml-sycl/cpy.cpp +0 -701
  163. package/src/llama.cpp/ggml/src/ggml-sycl/cpy.hpp +0 -11
  164. package/src/llama.cpp/ggml/src/ggml-sycl/dequantize.hpp +0 -791
  165. package/src/llama.cpp/ggml/src/ggml-sycl/dmmv.cpp +0 -1160
  166. package/src/llama.cpp/ggml/src/ggml-sycl/dmmv.hpp +0 -27
  167. package/src/llama.cpp/ggml/src/ggml-sycl/dpct/helper.hpp +0 -2957
  168. package/src/llama.cpp/ggml/src/ggml-sycl/element_wise.cpp +0 -1536
  169. package/src/llama.cpp/ggml/src/ggml-sycl/element_wise.hpp +0 -75
  170. package/src/llama.cpp/ggml/src/ggml-sycl/gemm.hpp +0 -99
  171. package/src/llama.cpp/ggml/src/ggml-sycl/getrows.cpp +0 -311
  172. package/src/llama.cpp/ggml/src/ggml-sycl/getrows.hpp +0 -20
  173. package/src/llama.cpp/ggml/src/ggml-sycl/ggml-sycl.cpp +0 -4443
  174. package/src/llama.cpp/ggml/src/ggml-sycl/gla.cpp +0 -105
  175. package/src/llama.cpp/ggml/src/ggml-sycl/gla.hpp +0 -8
  176. package/src/llama.cpp/ggml/src/ggml-sycl/im2col.cpp +0 -136
  177. package/src/llama.cpp/ggml/src/ggml-sycl/im2col.hpp +0 -21
  178. package/src/llama.cpp/ggml/src/ggml-sycl/mmq.cpp +0 -3030
  179. package/src/llama.cpp/ggml/src/ggml-sycl/mmq.hpp +0 -33
  180. package/src/llama.cpp/ggml/src/ggml-sycl/mmvq.cpp +0 -1108
  181. package/src/llama.cpp/ggml/src/ggml-sycl/mmvq.hpp +0 -27
  182. package/src/llama.cpp/ggml/src/ggml-sycl/norm.cpp +0 -474
  183. package/src/llama.cpp/ggml/src/ggml-sycl/norm.hpp +0 -26
  184. package/src/llama.cpp/ggml/src/ggml-sycl/outprod.cpp +0 -46
  185. package/src/llama.cpp/ggml/src/ggml-sycl/outprod.hpp +0 -10
  186. package/src/llama.cpp/ggml/src/ggml-sycl/presets.hpp +0 -74
  187. package/src/llama.cpp/ggml/src/ggml-sycl/quants.hpp +0 -83
  188. package/src/llama.cpp/ggml/src/ggml-sycl/rope.cpp +0 -362
  189. package/src/llama.cpp/ggml/src/ggml-sycl/rope.hpp +0 -20
  190. package/src/llama.cpp/ggml/src/ggml-sycl/softmax.cpp +0 -264
  191. package/src/llama.cpp/ggml/src/ggml-sycl/softmax.hpp +0 -20
  192. package/src/llama.cpp/ggml/src/ggml-sycl/sycl_hw.cpp +0 -13
  193. package/src/llama.cpp/ggml/src/ggml-sycl/sycl_hw.hpp +0 -23
  194. package/src/llama.cpp/ggml/src/ggml-sycl/tsembd.cpp +0 -73
  195. package/src/llama.cpp/ggml/src/ggml-sycl/tsembd.hpp +0 -20
  196. package/src/llama.cpp/ggml/src/ggml-sycl/vecdotq.hpp +0 -1215
  197. package/src/llama.cpp/ggml/src/ggml-sycl/wkv.cpp +0 -305
  198. package/src/llama.cpp/ggml/src/ggml-sycl/wkv.hpp +0 -10
  199. package/src/llama.cpp/ggml/src/ggml-threading.cpp +0 -12
  200. package/src/llama.cpp/ggml/src/ggml-threading.h +0 -14
  201. package/src/llama.cpp/ggml/src/ggml-vulkan/CMakeLists.txt +0 -196
  202. package/src/llama.cpp/ggml/src/ggml-vulkan/ggml-vulkan.cpp +0 -10699
  203. package/src/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/CMakeLists.txt +0 -39
  204. package/src/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp +0 -751
  205. package/src/llama.cpp/ggml/src/ggml.c +0 -6550
  206. package/src/llama.cpp/ggml/src/gguf.cpp +0 -1330
  207. package/src/llama.cpp/models/.editorconfig +0 -1
  208. package/src/llama.cpp/models/ggml-vocab-aquila.gguf +0 -0
  209. package/src/llama.cpp/models/ggml-vocab-baichuan.gguf +0 -0
  210. package/src/llama.cpp/models/ggml-vocab-bert-bge.gguf +0 -0
  211. package/src/llama.cpp/models/ggml-vocab-bert-bge.gguf.inp +0 -112
  212. package/src/llama.cpp/models/ggml-vocab-bert-bge.gguf.out +0 -46
  213. package/src/llama.cpp/models/ggml-vocab-chameleon.gguf.inp +0 -112
  214. package/src/llama.cpp/models/ggml-vocab-chameleon.gguf.out +0 -46
  215. package/src/llama.cpp/models/ggml-vocab-command-r.gguf +0 -0
  216. package/src/llama.cpp/models/ggml-vocab-command-r.gguf.inp +0 -112
  217. package/src/llama.cpp/models/ggml-vocab-command-r.gguf.out +0 -46
  218. package/src/llama.cpp/models/ggml-vocab-deepseek-coder.gguf +0 -0
  219. package/src/llama.cpp/models/ggml-vocab-deepseek-coder.gguf.inp +0 -112
  220. package/src/llama.cpp/models/ggml-vocab-deepseek-coder.gguf.out +0 -46
  221. package/src/llama.cpp/models/ggml-vocab-deepseek-llm.gguf +0 -0
  222. package/src/llama.cpp/models/ggml-vocab-deepseek-llm.gguf.inp +0 -112
  223. package/src/llama.cpp/models/ggml-vocab-deepseek-llm.gguf.out +0 -46
  224. package/src/llama.cpp/models/ggml-vocab-deepseek-r1-qwen.gguf.inp +0 -112
  225. package/src/llama.cpp/models/ggml-vocab-deepseek-r1-qwen.gguf.out +0 -46
  226. package/src/llama.cpp/models/ggml-vocab-falcon.gguf +0 -0
  227. package/src/llama.cpp/models/ggml-vocab-falcon.gguf.inp +0 -112
  228. package/src/llama.cpp/models/ggml-vocab-falcon.gguf.out +0 -46
  229. package/src/llama.cpp/models/ggml-vocab-gpt-2.gguf +0 -0
  230. package/src/llama.cpp/models/ggml-vocab-gpt-2.gguf.inp +0 -112
  231. package/src/llama.cpp/models/ggml-vocab-gpt-2.gguf.out +0 -46
  232. package/src/llama.cpp/models/ggml-vocab-gpt-4o.gguf.inp +0 -112
  233. package/src/llama.cpp/models/ggml-vocab-gpt-4o.gguf.out +0 -46
  234. package/src/llama.cpp/models/ggml-vocab-gpt-neox.gguf +0 -0
  235. package/src/llama.cpp/models/ggml-vocab-llama-bpe.gguf +0 -0
  236. package/src/llama.cpp/models/ggml-vocab-llama-bpe.gguf.inp +0 -112
  237. package/src/llama.cpp/models/ggml-vocab-llama-bpe.gguf.out +0 -46
  238. package/src/llama.cpp/models/ggml-vocab-llama-spm.gguf +0 -0
  239. package/src/llama.cpp/models/ggml-vocab-llama-spm.gguf.inp +0 -112
  240. package/src/llama.cpp/models/ggml-vocab-llama-spm.gguf.out +0 -46
  241. package/src/llama.cpp/models/ggml-vocab-llama4.gguf.inp +0 -112
  242. package/src/llama.cpp/models/ggml-vocab-llama4.gguf.out +0 -46
  243. package/src/llama.cpp/models/ggml-vocab-mpt.gguf +0 -0
  244. package/src/llama.cpp/models/ggml-vocab-mpt.gguf.inp +0 -112
  245. package/src/llama.cpp/models/ggml-vocab-mpt.gguf.out +0 -46
  246. package/src/llama.cpp/models/ggml-vocab-phi-3.gguf +0 -0
  247. package/src/llama.cpp/models/ggml-vocab-phi-3.gguf.inp +0 -112
  248. package/src/llama.cpp/models/ggml-vocab-phi-3.gguf.out +0 -46
  249. package/src/llama.cpp/models/ggml-vocab-pixtral.gguf.inp +0 -112
  250. package/src/llama.cpp/models/ggml-vocab-pixtral.gguf.out +0 -46
  251. package/src/llama.cpp/models/ggml-vocab-qwen2.gguf +0 -0
  252. package/src/llama.cpp/models/ggml-vocab-qwen2.gguf.inp +0 -112
  253. package/src/llama.cpp/models/ggml-vocab-qwen2.gguf.out +0 -46
  254. package/src/llama.cpp/models/ggml-vocab-refact.gguf +0 -0
  255. package/src/llama.cpp/models/ggml-vocab-refact.gguf.inp +0 -112
  256. package/src/llama.cpp/models/ggml-vocab-refact.gguf.out +0 -46
  257. package/src/llama.cpp/models/ggml-vocab-roberta-bpe.gguf.inp +0 -112
  258. package/src/llama.cpp/models/ggml-vocab-roberta-bpe.gguf.out +0 -46
  259. package/src/llama.cpp/models/ggml-vocab-starcoder.gguf +0 -0
  260. package/src/llama.cpp/models/ggml-vocab-starcoder.gguf.inp +0 -112
  261. package/src/llama.cpp/models/ggml-vocab-starcoder.gguf.out +0 -46
  262. package/src/llama.cpp/pocs/CMakeLists.txt +0 -14
  263. package/src/llama.cpp/pocs/vdot/CMakeLists.txt +0 -9
  264. package/src/llama.cpp/pocs/vdot/q8dot.cpp +0 -173
  265. package/src/llama.cpp/pocs/vdot/vdot.cpp +0 -311
  266. package/src/llama.cpp/prompts/LLM-questions.txt +0 -49
  267. package/src/llama.cpp/prompts/alpaca.txt +0 -1
  268. package/src/llama.cpp/prompts/assistant.txt +0 -31
  269. package/src/llama.cpp/prompts/chat-with-baichuan.txt +0 -4
  270. package/src/llama.cpp/prompts/chat-with-bob.txt +0 -7
  271. package/src/llama.cpp/prompts/chat-with-qwen.txt +0 -1
  272. package/src/llama.cpp/prompts/chat-with-vicuna-v0.txt +0 -7
  273. package/src/llama.cpp/prompts/chat-with-vicuna-v1.txt +0 -7
  274. package/src/llama.cpp/prompts/chat.txt +0 -28
  275. package/src/llama.cpp/prompts/dan-modified.txt +0 -1
  276. package/src/llama.cpp/prompts/dan.txt +0 -1
  277. package/src/llama.cpp/prompts/mnemonics.txt +0 -93
  278. package/src/llama.cpp/prompts/parallel-questions.txt +0 -43
  279. package/src/llama.cpp/prompts/reason-act.txt +0 -18
  280. package/src/llama.cpp/requirements/requirements-all.txt +0 -15
  281. package/src/llama.cpp/requirements/requirements-compare-llama-bench.txt +0 -2
  282. package/src/llama.cpp/requirements/requirements-convert_hf_to_gguf.txt +0 -7
  283. package/src/llama.cpp/requirements/requirements-convert_hf_to_gguf_update.txt +0 -7
  284. package/src/llama.cpp/requirements/requirements-convert_legacy_llama.txt +0 -5
  285. package/src/llama.cpp/requirements/requirements-convert_llama_ggml_to_gguf.txt +0 -1
  286. package/src/llama.cpp/requirements/requirements-convert_lora_to_gguf.txt +0 -4
  287. package/src/llama.cpp/requirements/requirements-gguf_editor_gui.txt +0 -3
  288. package/src/llama.cpp/requirements/requirements-pydantic.txt +0 -3
  289. package/src/llama.cpp/requirements/requirements-test-tokenizer-random.txt +0 -1
  290. package/src/llama.cpp/requirements/requirements-tool_bench.txt +0 -12
  291. package/src/llama.cpp/requirements.txt +0 -13
  292. package/src/llama.cpp/scripts/build-info.sh +0 -30
  293. package/src/llama.cpp/scripts/install-oneapi.bat +0 -19
  294. package/src/llama.cpp/scripts/xxd.cmake +0 -16
  295. package/src/llama.cpp/tests/CMakeLists.txt +0 -177
  296. package/src/llama.cpp/tests/get-model.cpp +0 -21
  297. package/src/llama.cpp/tests/get-model.h +0 -2
  298. package/src/llama.cpp/tests/test-arg-parser.cpp +0 -178
  299. package/src/llama.cpp/tests/test-autorelease.cpp +0 -24
  300. package/src/llama.cpp/tests/test-backend-ops.cpp +0 -4793
  301. package/src/llama.cpp/tests/test-barrier.cpp +0 -94
  302. package/src/llama.cpp/tests/test-c.c +0 -7
  303. package/src/llama.cpp/tests/test-chat-template.cpp +0 -417
  304. package/src/llama.cpp/tests/test-chat.cpp +0 -985
  305. package/src/llama.cpp/tests/test-double-float.cpp +0 -57
  306. package/src/llama.cpp/tests/test-gbnf-validator.cpp +0 -109
  307. package/src/llama.cpp/tests/test-gguf.cpp +0 -1338
  308. package/src/llama.cpp/tests/test-grammar-integration.cpp +0 -1308
  309. package/src/llama.cpp/tests/test-grammar-llguidance.cpp +0 -1201
  310. package/src/llama.cpp/tests/test-grammar-parser.cpp +0 -519
  311. package/src/llama.cpp/tests/test-json-schema-to-grammar.cpp +0 -1304
  312. package/src/llama.cpp/tests/test-llama-grammar.cpp +0 -408
  313. package/src/llama.cpp/tests/test-log.cpp +0 -39
  314. package/src/llama.cpp/tests/test-model-load-cancel.cpp +0 -27
  315. package/src/llama.cpp/tests/test-mtmd-c-api.c +0 -63
  316. package/src/llama.cpp/tests/test-opt.cpp +0 -904
  317. package/src/llama.cpp/tests/test-quantize-fns.cpp +0 -186
  318. package/src/llama.cpp/tests/test-quantize-perf.cpp +0 -365
  319. package/src/llama.cpp/tests/test-quantize-stats.cpp +0 -424
  320. package/src/llama.cpp/tests/test-regex-partial.cpp +0 -288
  321. package/src/llama.cpp/tests/test-rope.cpp +0 -262
  322. package/src/llama.cpp/tests/test-sampling.cpp +0 -399
  323. package/src/llama.cpp/tests/test-tokenizer-0.cpp +0 -312
  324. package/src/llama.cpp/tests/test-tokenizer-1-bpe.cpp +0 -155
  325. package/src/llama.cpp/tests/test-tokenizer-1-spm.cpp +0 -125
  326. package/src/llama.cpp/tools/CMakeLists.txt +0 -39
  327. package/src/llama.cpp/tools/batched-bench/CMakeLists.txt +0 -5
  328. package/src/llama.cpp/tools/batched-bench/batched-bench.cpp +0 -204
  329. package/src/llama.cpp/tools/cvector-generator/CMakeLists.txt +0 -5
  330. package/src/llama.cpp/tools/cvector-generator/completions.txt +0 -582
  331. package/src/llama.cpp/tools/cvector-generator/cvector-generator.cpp +0 -508
  332. package/src/llama.cpp/tools/cvector-generator/mean.hpp +0 -48
  333. package/src/llama.cpp/tools/cvector-generator/negative.txt +0 -4
  334. package/src/llama.cpp/tools/cvector-generator/pca.hpp +0 -315
  335. package/src/llama.cpp/tools/cvector-generator/positive.txt +0 -4
  336. package/src/llama.cpp/tools/export-lora/CMakeLists.txt +0 -5
  337. package/src/llama.cpp/tools/export-lora/export-lora.cpp +0 -434
  338. package/src/llama.cpp/tools/gguf-split/CMakeLists.txt +0 -5
  339. package/src/llama.cpp/tools/gguf-split/gguf-split.cpp +0 -583
  340. package/src/llama.cpp/tools/imatrix/CMakeLists.txt +0 -5
  341. package/src/llama.cpp/tools/imatrix/imatrix.cpp +0 -667
  342. package/src/llama.cpp/tools/llama-bench/CMakeLists.txt +0 -5
  343. package/src/llama.cpp/tools/llama-bench/llama-bench.cpp +0 -2024
  344. package/src/llama.cpp/tools/main/CMakeLists.txt +0 -5
  345. package/src/llama.cpp/tools/main/main.cpp +0 -977
  346. package/src/llama.cpp/tools/mtmd/CMakeLists.txt +0 -58
  347. package/src/llama.cpp/tools/mtmd/clip-impl.h +0 -462
  348. package/src/llama.cpp/tools/mtmd/clip.cpp +0 -4024
  349. package/src/llama.cpp/tools/mtmd/clip.h +0 -101
  350. package/src/llama.cpp/tools/mtmd/deprecation-warning.cpp +0 -22
  351. package/src/llama.cpp/tools/mtmd/miniaudio.h +0 -93468
  352. package/src/llama.cpp/tools/mtmd/mtmd-audio.cpp +0 -855
  353. package/src/llama.cpp/tools/mtmd/mtmd-audio.h +0 -62
  354. package/src/llama.cpp/tools/mtmd/mtmd-cli.cpp +0 -377
  355. package/src/llama.cpp/tools/mtmd/mtmd-helper.cpp +0 -297
  356. package/src/llama.cpp/tools/mtmd/mtmd.cpp +0 -942
  357. package/src/llama.cpp/tools/mtmd/mtmd.h +0 -362
  358. package/src/llama.cpp/tools/mtmd/requirements.txt +0 -5
  359. package/src/llama.cpp/tools/perplexity/CMakeLists.txt +0 -5
  360. package/src/llama.cpp/tools/perplexity/perplexity.cpp +0 -2063
  361. package/src/llama.cpp/tools/quantize/CMakeLists.txt +0 -6
  362. package/src/llama.cpp/tools/quantize/quantize.cpp +0 -519
  363. package/src/llama.cpp/tools/rpc/CMakeLists.txt +0 -4
  364. package/src/llama.cpp/tools/rpc/rpc-server.cpp +0 -322
  365. package/src/llama.cpp/tools/run/CMakeLists.txt +0 -16
  366. package/src/llama.cpp/tools/run/linenoise.cpp/linenoise.cpp +0 -1995
  367. package/src/llama.cpp/tools/run/linenoise.cpp/linenoise.h +0 -137
  368. package/src/llama.cpp/tools/run/run.cpp +0 -1261
  369. package/src/llama.cpp/tools/server/CMakeLists.txt +0 -51
  370. package/src/llama.cpp/tools/server/bench/requirements.txt +0 -2
  371. package/src/llama.cpp/tools/server/httplib.h +0 -10506
  372. package/src/llama.cpp/tools/server/server.cpp +0 -4966
  373. package/src/llama.cpp/tools/server/tests/requirements.txt +0 -8
  374. package/src/llama.cpp/tools/server/utils.hpp +0 -1337
  375. package/src/llama.cpp/tools/tokenize/CMakeLists.txt +0 -5
  376. package/src/llama.cpp/tools/tokenize/tokenize.cpp +0 -416
  377. package/src/llama.cpp/tools/tts/CMakeLists.txt +0 -5
  378. package/src/llama.cpp/tools/tts/tts.cpp +0 -1092
@@ -1,94 +0,0 @@
1
- #include "ggml.h"
2
- #include "ggml-cpu.h"
3
- #include "ggml-backend.h"
4
-
5
- #include <chrono>
6
- #include <iostream>
7
- #include <cstdio>
8
- #include <cstdlib>
9
- #include <cassert>
10
- #include <vector>
11
-
12
- #define MAX_NARGS 2
13
-
14
- int main(int argc, char *argv[]) {
15
-
16
- int n_threads = 4;
17
- int n_rounds = 100;
18
-
19
- if (argc > 1) {
20
- n_threads = std::atoi(argv[1]);
21
- }
22
-
23
- if (argc > 2) {
24
- n_rounds = std::atoi(argv[2]);
25
- }
26
-
27
- struct ggml_init_params params = {
28
- /* .mem_size = */ 1024*1024*1024,
29
- /* .mem_buffer = */ NULL,
30
- /* .no_alloc = */ false,
31
- };
32
-
33
- struct ggml_context * ctx = ggml_init(params);
34
-
35
- // Create graph
36
- struct ggml_cgraph * gf = ggml_new_graph(ctx);
37
-
38
- // Lots of small, parallel ops where barriers in between will dominate
39
- struct ggml_tensor * out = ggml_new_tensor_1d(ctx, GGML_TYPE_F32, 64);
40
- for (int i = 0; i < 1000; i++) {
41
- struct ggml_tensor * a = ggml_new_tensor_2d(ctx, GGML_TYPE_Q4_0, 64, 128);
42
- out = ggml_mul_mat(ctx, a, out);
43
-
44
- struct ggml_tensor * d = ggml_new_tensor_2d(ctx, GGML_TYPE_Q4_0, 128, 64);
45
- out = ggml_mul_mat(ctx, d, out);
46
- }
47
-
48
- ggml_build_forward_expand(gf, out);
49
- int n_nodes = ggml_graph_n_nodes(gf);
50
-
51
- // Create threadpool
52
- struct ggml_threadpool_params tpp = ggml_threadpool_params_default(n_threads);
53
- struct ggml_threadpool* threadpool = ggml_threadpool_new(&tpp);
54
- if (!threadpool) {
55
- fprintf(stderr, "threadpool create failed : n_threads %d\n", n_threads);
56
- exit(1);
57
- }
58
-
59
- // Create compute plan
60
- struct ggml_cplan cplan = ggml_graph_plan(gf, n_threads, threadpool);
61
-
62
- std::vector<uint8_t> work_data(cplan.work_size);
63
- cplan.work_data = work_data.data();
64
-
65
- std::cerr << "graph-compute with"
66
- << "\n n_threads: " << n_threads
67
- << "\n n_nodes: " << n_nodes
68
- << "\n n_rounds: " << n_rounds
69
- << "\n";
70
- // ggml_graph_print(gf);
71
-
72
- // Warmup
73
- ggml_graph_compute(gf, &cplan);
74
-
75
- auto t0 = std::chrono::high_resolution_clock::now();
76
-
77
- for (int i=0; i < n_rounds; i++) {
78
- ggml_graph_compute(gf, &cplan);
79
- }
80
-
81
- auto t1 = std::chrono::high_resolution_clock::now();
82
-
83
- auto usec = std::chrono::duration_cast<std::chrono::microseconds>(t1-t0).count();
84
- auto nsec = std::chrono::duration_cast<std::chrono::nanoseconds>(t1-t0).count();
85
- std::cerr << "graph-compute took " << usec << " usec "
86
- << "\n " << (float) usec / n_rounds << " usec per-iter"
87
- << "\n " << (float) nsec / (n_rounds * n_nodes) << " nsec per-node"
88
- << "\n";
89
-
90
- ggml_threadpool_free(threadpool);
91
- ggml_free(ctx);
92
-
93
- return 0;
94
- }
@@ -1,7 +0,0 @@
1
- #include "llama.h"
2
-
3
- #ifdef GGML_USE_KOMPUTE
4
- #include "ggml-kompute.h"
5
- #endif
6
-
7
- int main(void) {}
@@ -1,417 +0,0 @@
1
- #include <string>
2
- #include <vector>
3
- #include <sstream>
4
- #include <regex>
5
-
6
- #undef NDEBUG
7
- #include <cassert>
8
-
9
- #include "llama.h"
10
- #include "common.h"
11
- #include "chat.h"
12
-
13
- static std::string normalize_newlines(const std::string & s) {
14
- #ifdef _WIN32
15
- static const std::regex nl_regex("\r\n");
16
- return std::regex_replace(s, nl_regex, "\n");
17
- #else
18
- return s;
19
- #endif
20
- }
21
-
22
- #define U8C(x) (const char*)(u8##x)
23
-
24
- static common_chat_msg simple_msg(const std::string & role, const std::string & content) {
25
- common_chat_msg msg;
26
- msg.role = role;
27
- msg.content = content;
28
- return msg;
29
- }
30
-
31
- int main(void) {
32
- std::vector<llama_chat_message> conversation {
33
- {"system", "You are a helpful assistant"},
34
- {"user", "Hello"},
35
- {"assistant", "Hi there"},
36
- {"user", "Who are you"},
37
- {"assistant", " I am an assistant "},
38
- {"user", "Another question"},
39
- };
40
-
41
- // std::string wrong = /* .template_str= */ u8"[gMASK]<sop>{% for item in messages %}{% if item['tools'] is defined %}<|system|>\n你是一个名为 ChatGLM 的人工智能助手。你是基于智谱AI训练的语言模型 GLM-4 模型开发的,你的任务是针对用户的问题和要求提供适当的答复和支持。\n\n# 可用工具{% set tools = item['tools'] %}{% for tool in tools %}{% if tool['type'] == 'function' %}\n\n## {{ tool['function']['name'] }}\n\n{{ tool['function'] | tojson(indent=4) }}\n......{% endif %}{% endfor %}{% endif %}{% if item['content'] %}<|{{ item['role'] }}|>{{ item['metadata'] }}\n{{ item['content'] }}{% endif %}{% endfor %}{% if add_generation_prompt %}<|assistant|>{% endif %}";
42
- struct TestCase {
43
- std::string name;
44
- std::string template_str;
45
- std::string expected_output;
46
- std::string expected_output_jinja;
47
- std::string bos_token = "";
48
- std::string eos_token = "";
49
- bool supported_with_jinja = true;
50
- };
51
- std::vector<TestCase> test_cases {
52
- {
53
- /* .name= */ "teknium/OpenHermes-2.5-Mistral-7B",
54
- /* .template_str= */ "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\\n' + message['content'] + '<|im_end|>' + '\\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\\n' }}{% endif %}",
55
- /* .expected_output= */ "<|im_start|>system\nYou are a helpful assistant<|im_end|>\n<|im_start|>user\nHello<|im_end|>\n<|im_start|>assistant\nHi there<|im_end|>\n<|im_start|>user\nWho are you<|im_end|>\n<|im_start|>assistant\n I am an assistant <|im_end|>\n<|im_start|>user\nAnother question<|im_end|>\n<|im_start|>assistant\n",
56
- /* .expected_output_jinja= */ "",
57
- /* .bos_token= */ "",
58
- /* .eos_token= */ "",
59
- },
60
- {
61
- /* .name= */ "mistralai/Mistral-7B-Instruct-v0.2 (NOTE: Old pre-v1 without a system prompt)",
62
- /* .template_str= */ "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + ' [/INST]' }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token}}{% else %}{{ raise_exception('Only user and assistant roles are supported!') }}{% endif %}{% endfor %}",
63
- /* .expected_output= */ "[INST] You are a helpful assistant\nHello [/INST]Hi there</s>[INST] Who are you [/INST] I am an assistant </s>[INST] Another question [/INST]",
64
- /* .expected_output_jinja= */ "",
65
- /* .bos_token= */ "<s>",
66
- /* .eos_token= */ "</s>",
67
- },
68
- {
69
- /* .name= */ "TheBloke/FusionNet_34Bx2_MoE-AWQ",
70
- /* .template_str= */ "{%- for idx in range(0, messages|length) -%}\n{%- if messages[idx]['role'] == 'user' -%}\n{%- if idx > 1 -%}\n{{- bos_token + '[INST] ' + messages[idx]['content'] + ' [/INST]' -}}\n{%- else -%}\n{{- messages[idx]['content'] + ' [/INST]' -}}\n{%- endif -%}\n{% elif messages[idx]['role'] == 'system' %}\n{{- '[INST] <<SYS>>\\n' + messages[idx]['content'] + '\\n<</SYS>>\\n\\n' -}}\n{%- elif messages[idx]['role'] == 'assistant' -%}\n{{- ' ' + messages[idx]['content'] + ' ' + eos_token -}}\n{% endif %}\n{% endfor %}",
71
- /* .expected_output= */ "[INST] <<SYS>>\nYou are a helpful assistant\n<</SYS>>\n\nHello [/INST]Hi there</s><s>[INST] Who are you [/INST] I am an assistant </s><s>[INST] Another question [/INST]",
72
- /* .expected_output_jinja= */ "[INST] <<SYS>>\nYou are a helpful assistant\n<</SYS>>\n\nHello [/INST] Hi there </s><s>[INST] Who are you [/INST] I am an assistant </s><s>[INST] Another question [/INST]",
73
- /* .bos_token= */ "<s>",
74
- /* .eos_token= */ "</s>",
75
- },
76
- {
77
- /* .name= */ "bofenghuang/vigogne-2-70b-chat",
78
- /* .template_str= */ "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% elif true == true and not '<<SYS>>' in messages[0]['content'] %}{% set loop_messages = messages %}{% set system_message = 'Vous êtes Vigogne, un assistant IA créé par Zaion Lab. Vous suivez extrêmement bien les instructions. Aidez autant que vous le pouvez.' %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if loop.index0 == 0 and system_message != false %}{% set content = '<<SYS>>\\n' + system_message + '\\n<</SYS>>\\n\\n' + message['content'] %}{% else %}{% set content = message['content'] %}{% endif %}{% if message['role'] == 'user' %}{{ '[INST] ' + content.strip() + ' [/INST]' }}{% elif message['role'] == 'system' %}{{ '<<SYS>>\\n' + content.strip() + '\\n<</SYS>>\\n\\n' }}{% elif message['role'] == 'assistant' %}{{ ' ' + content.strip() + ' ' + eos_token }}{% endif %}{% endfor %}",
79
- /* .expected_output= */ "[INST] <<SYS>>\nYou are a helpful assistant\n<</SYS>>\n\nHello [/INST]Hi there</s>[INST] Who are you [/INST]I am an assistant</s>[INST] Another question [/INST]",
80
- /* .expected_output_jinja= */ "[INST] <<SYS>>\nYou are a helpful assistant\n<</SYS>>\n\nHello [/INST] Hi there </s>[INST] Who are you [/INST] I am an assistant </s>[INST] Another question [/INST]",
81
- /* .bos_token= */ "",
82
- /* .eos_token= */ "</s>",
83
- },
84
- {
85
- /* .name= */ "mlabonne/AlphaMonarch-7B",
86
- /* .template_str= */ "{% for message in messages %}{{bos_token + message['role'] + '\\n' + message['content'] + eos_token + '\\n'}}{% endfor %}{% if add_generation_prompt %}{{ bos_token + 'assistant\\n' }}{% endif %}",
87
- /* .expected_output= */ "system\nYou are a helpful assistant</s>\n<s>user\nHello</s>\n<s>assistant\nHi there</s>\n<s>user\nWho are you</s>\n<s>assistant\n I am an assistant </s>\n<s>user\nAnother question</s>\n<s>assistant\n",
88
- /* .expected_output_jinja= */ "",
89
- /* .bos_token= */ "<s>",
90
- /* .eos_token= */ "</s>",
91
- },
92
- {
93
- /* .name= */ "google/gemma-7b-it",
94
- /* .template_str= */ "{% if messages[0]['role'] == 'system' %}{{ raise_exception('System role not supported') }}{% endif %}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if (message['role'] == 'assistant') %}{% set role = 'model' %}{% else %}{% set role = message['role'] %}{% endif %}{{ '<start_of_turn>' + role + '\\n' + message['content'] | trim + '<end_of_turn>\\n' }}{% endfor %}{% if add_generation_prompt %}{{'<start_of_turn>model\\n'}}{% endif %}",
95
- /* .expected_output= */ "<start_of_turn>user\nYou are a helpful assistant\n\nHello<end_of_turn>\n<start_of_turn>model\nHi there<end_of_turn>\n<start_of_turn>user\nWho are you<end_of_turn>\n<start_of_turn>model\nI am an assistant<end_of_turn>\n<start_of_turn>user\nAnother question<end_of_turn>\n<start_of_turn>model\n",
96
- /* .expected_output_jinja= */ "<start_of_turn>user\nYou are a helpful assistant\nHello<end_of_turn>\n<start_of_turn>model\nHi there<end_of_turn>\n<start_of_turn>user\nWho are you<end_of_turn>\n<start_of_turn>model\nI am an assistant<end_of_turn>\n<start_of_turn>user\nAnother question<end_of_turn>\n<start_of_turn>model\n",
97
- },
98
- {
99
- /* .name= */ "OrionStarAI/Orion-14B-Chat",
100
- /* .template_str= */ "{% for message in messages %}{% if loop.first %}{{ bos_token }}{% endif %}{% if message['role'] == 'user' %}{{ 'Human: ' + message['content'] + '\\n\\nAssistant: ' + eos_token }}{% elif message['role'] == 'assistant' %}{{ message['content'] + eos_token }}{% endif %}{% endfor %}",
101
- /* .expected_output= */ "Human: You are a helpful assistant\n\nHello\n\nAssistant: </s>Hi there</s>Human: Who are you\n\nAssistant: </s> I am an assistant </s>Human: Another question\n\nAssistant: </s>",
102
- /* .expected_output_jinja= */ "Human: You are a helpful assistant\nHello\n\nAssistant: </s>Hi there</s>Human: Who are you\n\nAssistant: </s> I am an assistant </s>Human: Another question\n\nAssistant: ",
103
- /* .bos_token= */ "",
104
- /* .eos_token= */ "</s>",
105
- },
106
- {
107
- /* .name= */ "openchat/openchat-3.5-0106",
108
- // The included chat_template differs from the author's suggestions here: https://huggingface.co/openchat/openchat_3.5/discussions/5#65448109b4a3f3a2f486fd9d
109
- // So we match against the included template but implement the suggested version.
110
- /* .template_str= */ "{{ bos_token }}{% for message in messages %}{{ 'GPT4 Correct ' + message['role'].title() + ': ' + message['content'] + '<|end_of_turn|>'}}{% endfor %}{% if add_generation_prompt %}{{ 'GPT4 Correct Assistant:' }}{% endif %}",
111
- /* .expected_output= */ "You are a helpful assistant<|end_of_turn|>GPT4 Correct User: Hello<|end_of_turn|>GPT4 Correct Assistant: Hi there<|end_of_turn|>GPT4 Correct User: Who are you<|end_of_turn|>GPT4 Correct Assistant: I am an assistant <|end_of_turn|>GPT4 Correct User: Another question<|end_of_turn|>GPT4 Correct Assistant:",
112
- /* .expected_output_jinja= */ "GPT4 Correct System: You are a helpful assistant<|end_of_turn|>GPT4 Correct User: Hello<|end_of_turn|>GPT4 Correct Assistant: Hi there<|end_of_turn|>GPT4 Correct User: Who are you<|end_of_turn|>GPT4 Correct Assistant: I am an assistant <|end_of_turn|>GPT4 Correct User: Another question<|end_of_turn|>GPT4 Correct Assistant:",
113
- },
114
- {
115
- /* .name= */ "deepseek-ai/deepseek-coder-33b-instruct",
116
- /* .template_str= */ "{% if not add_generation_prompt is defined %}\n{% set add_generation_prompt = false %}\n{% endif %}\n{%- set ns = namespace(found=false) -%}\n{%- for message in messages -%}\n {%- if message['role'] == 'system' -%}\n {%- set ns.found = true -%}\n {%- endif -%}\n{%- endfor -%}\n{{bos_token}}{%- if not ns.found -%}\n{{'You are an AI programming assistant, utilizing the Deepseek Coder model, developed by Deepseek Company, and you only answer questions related to computer science. For politically sensitive questions, security and privacy issues, and other non-computer science questions, you will refuse to answer\\n'}}\n{%- endif %}\n{%- for message in messages %}\n {%- if message['role'] == 'system' %}\n{{ message['content'] }}\n {%- else %}\n {%- if message['role'] == 'user' %}\n{{'### Instruction:\\n' + message['content'] + '\\n'}}\n {%- else %}\n{{'### Response:\\n' + message['content'] + '\\n<|EOT|>\\n'}}\n {%- endif %}\n {%- endif %}\n{%- endfor %}\n{% if add_generation_prompt %}\n{{'### Response:'}}\n{% endif %}",
117
- /* .expected_output= */ "You are a helpful assistant### Instruction:\nHello\n### Response:\nHi there\n<|EOT|>\n### Instruction:\nWho are you\n### Response:\n I am an assistant \n<|EOT|>\n### Instruction:\nAnother question\n### Response:\n",
118
- /* .expected_output_jinja= */ "",
119
- },
120
- {
121
- /* .name= */ "eachadea/vicuna-13b-1.1",
122
- // No template included in tokenizer_config.json, so this template likely needs to be manually set.
123
- /* .template_str= */ "{%- for message in messages %}{%- if message['role'] == 'system' -%}{{- '' + message['content'] + '\n\n' -}}{%- else -%}{%- if message['role'] == 'user' -%}{{-'USER: ' + message['content'] + '\n'-}}{%- else -%}{{-'ASSISTANT: ' + message['content'] + '</s>\n' -}}{%- endif -%}{%- endif -%}{%- endfor -%}{%- if add_generation_prompt -%}{{-'ASSISTANT:'-}}{%- endif -%}",
124
- /* .expected_output= */ "You are a helpful assistant\n\nUSER: Hello\nASSISTANT: Hi there</s>\nUSER: Who are you\nASSISTANT: I am an assistant </s>\nUSER: Another question\nASSISTANT:",
125
- /* .expected_output_jinja= */ "",
126
- /* .bos_token= */ "",
127
- /* .eos_token= */ "",
128
- },
129
- {
130
- /* .name= */ "Orca-Vicuna",
131
- // No template included in tokenizer_config.json, so this template likely needs to be manually set.
132
- /* .template_str= */ "{%- for message in messages %}{%- if message['role'] == 'system' -%}{{-'SYSTEM: ' + message['content'] + '\n' -}}{%- else -%}{%- if message['role'] == 'user' -%}{{-'USER: ' + message['content'] + '\n'-}}{%- else -%}{{-'ASSISTANT: ' + message['content'] + '</s>\n' -}}{%- endif -%}{%- endif -%}{%- endfor -%}{%- if add_generation_prompt -%}{{-'ASSISTANT:'-}}{%- endif -%}",
133
- /* .expected_output= */ "SYSTEM: You are a helpful assistant\nUSER: Hello\nASSISTANT: Hi there</s>\nUSER: Who are you\nASSISTANT: I am an assistant </s>\nUSER: Another question\nASSISTANT:",
134
- /* .expected_output_jinja= */ "",
135
- /* .bos_token= */ "",
136
- /* .eos_token= */ "",
137
- },
138
- {
139
- /* .name= */ "CohereForAI/c4ai-command-r-plus",
140
- /* .template_str= */ "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% elif false == true %}{% set loop_messages = messages %}{% set system_message = 'You are Command-R, a brilliant, sophisticated, AI-assistant trained to assist human users by providing thorough responses. You are trained by Cohere.' %}{% else %}{% set loop_messages = messages %}{% set system_message = false %}{% endif %}{% if system_message != false %}{{ '<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>' + system_message + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% for message in loop_messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|START_OF_TURN_TOKEN|><|USER_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% elif message['role'] == 'assistant' %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' + content.strip() + '<|END_OF_TURN_TOKEN|>' }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>' }}{% endif %}",
141
- /* .expected_output= */ "<|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>You are a helpful assistant<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Hello<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>Hi there<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Who are you<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>I am an assistant<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Another question<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|>",
142
- /* .expected_output_jinja= */ "",
143
- },
144
- {
145
- /* .name= */ "Llama-3",
146
- /* .template_str= */ "{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}",
147
- /* .expected_output= */ "<|start_header_id|>system<|end_header_id|>\n\nYou are a helpful assistant<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nHello<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\nHi there<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nWho are you<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\nI am an assistant<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nAnother question<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n",
148
- /* .expected_output_jinja= */ "",
149
- },
150
- {
151
- /* .name= */ "Phi-3-mini",
152
- /* .template_str= */ "{{ bos_token }}{% for message in messages %}{% if (message['role'] == 'user') %}{{'<|user|>' + '\n' + message['content'] + '<|end|>' + '\n' + '<|assistant|>' + '\n'}}{% elif (message['role'] == 'assistant') %}{{message['content'] + '<|end|>' + '\n'}}{% endif %}{% endfor %}",
153
- /* .expected_output= */ "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n I am an assistant <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
154
- /* .expected_output_jinja= */ "<|user|>\nYou are a helpful assistant\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n I am an assistant <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
155
- },
156
- {
157
- /* .name= */ "Phi-3-small",
158
- /* .template_str= */ "{{ bos_token }}{% for message in messages %}{{'<|' + message['role'] + '|>' + '\n' + message['content'] + '<|end|>\n' }}{% endfor %}{% if add_generation_prompt %}{{ '<|assistant|>\n' }}{% else %}{{ eos_token }}{% endif %}",
159
- /* .expected_output= */ "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n I am an assistant <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
160
- /* .expected_output_jinja= */ "",
161
- },
162
- {
163
- /* .name= */ "Phi-3-medium",
164
- /* .template_str= */ "{% for message in messages %}{% if (message['role'] == 'user') %}{{'<|user|>' + '\n' + message['content'] + '<|end|>' + '\n' + '<|assistant|>' + '\n'}}{% elif (message['role'] == 'assistant') %}{{message['content'] + '<|end|>' + '\n'}}{% endif %}{% endfor %}",
165
- /* .expected_output= */ "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n I am an assistant <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
166
- /* .expected_output_jinja= */ "<|user|>\nYou are a helpful assistant\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n I am an assistant <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
167
- },
168
- {
169
- /* .name= */ "Phi-3-vision",
170
- /* .template_str= */ "{% for message in messages %}{{'<|' + message['role'] + '|>' + '\n' + message['content'] + '<|end|>\n' }}{% endfor %}{% if add_generation_prompt and messages[-1]['role'] != 'assistant' %}{{- '<|assistant|>\n' -}}{% endif %}",
171
- /* .expected_output= */ "<|system|>\nYou are a helpful assistant<|end|>\n<|user|>\nHello<|end|>\n<|assistant|>\nHi there<|end|>\n<|user|>\nWho are you<|end|>\n<|assistant|>\n I am an assistant <|end|>\n<|user|>\nAnother question<|end|>\n<|assistant|>\n",
172
- /* .expected_output_jinja= */ "",
173
- /* .bos_token= */ "",
174
- /* .eos_token= */ "",
175
- },
176
- {
177
- /* .name= */ "ChatGLM3",
178
- /* .template_str= */ "{% for message in messages %}{% if loop.first %}[gMASK]sop<|{{ message['role'] }}|>\n {{ message['content'] }}{% else %}<|{{ message['role'] }}|>\n {{ message['content'] }}{% endif %}{% endfor %}{% if add_generation_prompt %}<|assistant|>{% endif %}",
179
- /* .expected_output= */ "[gMASK]sop<|system|>\n You are a helpful assistant<|user|>\n Hello<|assistant|>\n Hi there<|user|>\n Who are you<|assistant|>\n I am an assistant <|user|>\n Another question<|assistant|>",
180
- /* .expected_output_jinja= */ "[gMASK]sop<|system|>\nYou are a helpful assistant<|user|>\nHello<|assistant|>\nHi there<|user|>\nWho are you<|assistant|>\n I am an assistant <|user|>\nAnother question<|assistant|>",
181
- },
182
- {
183
- /* .name= */ "ChatGLM4",
184
- /* .template_str= */ U8C("[gMASK]<sop>{% for item in messages %}{% if item['tools'] is defined %}<|system|>\n你是一个名为 ChatGLM 的人工智能助手。你是基于智谱AI训练的语言模型 GLM-4 模型开发的,你的任务是针对用户的问题和要求提供适当的答复和支持。\n\n# 可用工具{% set tools = item['tools'] %}{% for tool in tools %}{% if tool['type'] == 'function' %}\n\n## {{ tool['function']['name'] }}\n\n{{ tool['function'] | tojson(indent=4) }}\n......{% endif %}{% endfor %}{% endif %}{% if item['content'] %}<|{{ item['role'] }}|>{{ item['metadata'] }}\n{{ item['content'] }}{% endif %}{% endfor %}{% if add_generation_prompt %}<|assistant|>\n{% endif %}"),
185
- /* .expected_output= */ "[gMASK]<sop><|system|>\nYou are a helpful assistant<|user|>\nHello<|assistant|>\nHi there<|user|>\nWho are you<|assistant|>\n I am an assistant <|user|>\nAnother question<|assistant|>\n",
186
- /* .expected_output_jinja= */ "",
187
- /* .bos_token= */ "",
188
- /* .eos_token= */ "",
189
- },
190
- {
191
- /* .name= */ "GLMEdge",
192
- /* .template_str= */ "{% for item in messages %}{% if item['role'] == 'system' %}<|system|>\n{{ item['content'] }}{% elif item['role'] == 'user' %}<|user|>\n{{ item['content'] }}{% elif item['role'] == 'assistant' %}<|assistant|>\n{{ item['content'] }}{% endif %}{% endfor %}<|assistant|>",
193
- /* .expected_output= */ "<|system|>\nYou are a helpful assistant<|user|>\nHello<|assistant|>\nHi there<|user|>\nWho are you<|assistant|>\n I am an assistant <|user|>\nAnother question<|assistant|>",
194
- /* .expected_output_jinja= */ "<|system|>\nYou are a helpful assistant<|user|>\nHello<|assistant|>\nHi there<|user|>\nWho are you<|assistant|>\n I am an assistant <|user|>\nAnother question<|assistant|>",
195
- /* .bos_token= */ "",
196
- /* .eos_token= */ "",
197
- },
198
- {
199
- /* .name= */ "MiniCPM-3B-OpenHermes-2.5-v2-GGUF",
200
- /* .template_str= */ U8C("{% for message in messages %}{% if message['role'] == 'user' %}{{'<用户>' + message['content'].strip() + '<AI>'}}{% else %}{{message['content'].strip()}}{% endif %}{% endfor %}"),
201
- /* .expected_output= */ U8C("You are a helpful assistant<用户>Hello<AI>Hi there<用户>Who are you<AI>I am an assistant<用户>Another question<AI>"),
202
- /* .expected_output_jinja= */ "",
203
- /* .bos_token= */ "",
204
- /* .eos_token= */ "",
205
- },
206
- {
207
- /* .name= */ "DeepSeek-V2",
208
- /* .template_str= */ "{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{{ bos_token }}{% for message in messages %}{% if message['role'] == 'user' %}{{ 'User: ' + message['content'] + '\n\n' }}{% elif message['role'] == 'assistant' %}{{ 'Assistant: ' + message['content'] + eos_token }}{% elif message['role'] == 'system' %}{{ message['content'] + '\n\n' }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}",
209
- /* .expected_output= */ U8C("You are a helpful assistant\n\nUser: Hello\n\nAssistant: Hi there<|end▁of▁sentence|>User: Who are you\n\nAssistant: I am an assistant <|end▁of▁sentence|>User: Another question\n\nAssistant:"),
210
- /* .expected_output_jinja= */ "",
211
- /* .bos_token= */ "",
212
- /* .eos_token= */ "<|end▁of▁sentence|>",
213
- },
214
- {
215
- /* .name= */ "ibm-granite/granite-3.0-8b-instruct",
216
- /* .template_str= */ "{%- if tools %}\n {{- '<|start_of_role|>available_tools<|end_of_role|>\n' }}\n {%- for tool in tools %}\n {{- tool | tojson(indent=4) }}\n {%- if not loop.last %}\n {{- '\n\n' }}\n {%- endif %}\n {%- endfor %}\n {{- '<|end_of_text|>\n' }}\n{%- endif %}\n{%- for message in messages %}\n {%- if message['role'] == 'system' %}\n {{- '<|start_of_role|>system<|end_of_role|>' + message['content'] + '<|end_of_text|>\n' }}\n {%- elif message['role'] == 'user' %}\n {{- '<|start_of_role|>user<|end_of_role|>' + message['content'] + '<|end_of_text|>\n' }}\n {%- elif message['role'] == 'assistant' %}\n {{- '<|start_of_role|>assistant<|end_of_role|>' + message['content'] + '<|end_of_text|>\n' }}\n {%- elif message['role'] == 'assistant_tool_call' %}\n {{- '<|start_of_role|>assistant<|end_of_role|><|tool_call|>' + message['content'] + '<|end_of_text|>\n' }}\n {%- elif message['role'] == 'tool_response' %}\n {{- '<|start_of_role|>tool_response<|end_of_role|>' + message['content'] + '<|end_of_text|>\n' }}\n {%- endif %}\n {%- if loop.last and add_generation_prompt %}\n {{- '<|start_of_role|>assistant<|end_of_role|>' }}\n {%- endif %}\n{%- endfor %}",
217
- /* .expected_output= */ "<|start_of_role|>system<|end_of_role|>You are a helpful assistant<|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Hello<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|>Hi there<|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Who are you<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|> I am an assistant <|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Another question<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|>\n",
218
- /* .expected_output_jinja= */ "<|start_of_role|>system<|end_of_role|>You are a helpful assistant<|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Hello<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|>Hi there<|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Who are you<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|> I am an assistant <|end_of_text|>\n<|start_of_role|>user<|end_of_role|>Another question<|end_of_text|>\n<|start_of_role|>assistant<|end_of_role|>",
219
- },
220
- {
221
- /* .name= */ "mistralai/Mistral-7B-Instruct-v0.2 (mistralai 'v1' template with a system prompt)",
222
- /* .template_str= */ "{%- if messages[0]['role'] == 'system' %}\n {%- set system_message = messages[0]['content'] %}\n {%- set loop_messages = messages[1:] %}\n{%- else %}\n {%- set loop_messages = messages %}\n{%- endif %}\n\n{{- bos_token }}\n{%- for message in loop_messages %}\n {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}\n {{- raise_exception('After the optional system message, conversation roles must alternate user/assistant/user/assistant/...') }}\n {%- endif %}\n {%- if message['role'] == 'user' %}\n {%- if loop.first and system_message is defined %}\n {{- ' [INST] ' + system_message + '\\n\\n' + message['content'] + ' [/INST]' }}\n {%- else %}\n {{- ' [INST] ' + message['content'] + ' [/INST]' }}\n {%- endif %}\n {%- elif message['role'] == 'assistant' %}\n {{- ' ' + message['content'] + eos_token}}\n {%- else %}\n {{- raise_exception('Only user and assistant roles are supported, with the exception of an initial optional system message!') }}\n {%- endif %}\n{%- endfor %}\n",
223
- /* .expected_output= */ " [INST] You are a helpful assistant\n\nHello [/INST] Hi there</s> [INST] Who are you [/INST] I am an assistant </s> [INST] Another question [/INST]",
224
- /* .expected_output_jinja= */ "",
225
- /* .bos_token= */ "",
226
- /* .eos_token= */ "</s>",
227
- },
228
- {
229
- /* .name= */ "Mistral-Large-Instruct-2407 (mistralai 'v3' template; modified to have system prompt at start)",
230
- /* .template_str= */ "{%- if messages[0][\"role\"] == \"system\" %}\n {%- set system_message = messages[0][\"content\"] %}\n {%- set loop_messages = messages[1:] %}\n{%- else %}\n {%- set loop_messages = messages %}\n{%- endif %}\n{%- if not tools is defined %}\n {%- set tools = none %}\n{%- endif %}\n{%- set user_messages = loop_messages | selectattr(\"role\", \"equalto\", \"user\") | list %}\n\n{#- This block checks for alternating user/assistant messages, skipping tool calling messages #}\n{%- set ns = namespace() %}\n{%- set ns.index = 0 %}\n{%- for message in loop_messages %}\n {%- if not (message.role == \"tool\" or message.role == \"tool_results\" or (message.tool_calls is defined and message.tool_calls is not none)) %}\n {%- if (message[\"role\"] == \"user\") != (ns.index % 2 == 0) %}\n {{- raise_exception(\"After the optional system message, conversation roles must alternate user/assistant/user/assistant/...\") }}\n {%- endif %}\n {%- set ns.index = ns.index + 1 %}\n {%- endif %}\n{%- endfor %}\n\n{{- bos_token }}\n{%- for message in loop_messages %}\n {%- if message[\"role\"] == \"user\" %}\n {%- if tools is not none and (message == user_messages[-1]) %}\n {{- \"[AVAILABLE_TOOLS] [\" }}\n {%- for tool in tools %}\n {%- set tool = tool.function %}\n {{- '{\"type\": \"function\", \"function\": {' }}\n {%- for key, val in tool.items() if key != \"return\" %}\n {%- if val is string %}\n {{- '\"' + key + '\": \"' + val + '\"' }}\n {%- else %}\n {{- '\"' + key + '\": ' + val|tojson }}\n {%- endif %}\n {%- if not loop.last %}\n {{- \", \" }}\n {%- endif %}\n {%- endfor %}\n {{- \"}}\" }}\n {%- if not loop.last %}\n {{- \", \" }}\n {%- else %}\n {{- \"]\" }}\n {%- endif %}\n {%- endfor %}\n {{- \"[/AVAILABLE_TOOLS]\" }}\n {%- endif %}\n {%- if loop.last and system_message is defined %}\n {{- \"[INST] \" + system_message + \"\\n\\n\" + message[\"content\"] + \"[/INST]\" }}\n {%- else %}\n {{- \"[INST] \" + message[\"content\"] + \"[/INST]\" }}\n {%- endif %}\n {%- elif message.tool_calls is defined and message.tool_calls is not none %}\n {{- \"[TOOL_CALLS] [\" }}\n {%- for tool_call in message.tool_calls %}\n {%- set out = tool_call.function|tojson %}\n {{- out[:-1] }}\n {%- if not tool_call.id is defined or tool_call.id|length != 9 %}\n {{- raise_exception(\"Tool call IDs should be alphanumeric strings with length 9!\") }}\n {%- endif %}\n {{- ', \"id\": \"' + tool_call.id + '\"}' }}\n {%- if not loop.last %}\n {{- \", \" }}\n {%- else %}\n {{- \"]\" + eos_token }}\n {%- endif %}\n {%- endfor %}\n {%- elif message[\"role\"] == \"assistant\" %}\n {{- \" \" + message[\"content\"]|trim + eos_token}}\n {%- elif message[\"role\"] == \"tool_results\" or message[\"role\"] == \"tool\" %}\n {%- if message.content is defined and message.content.content is defined %}\n {%- set content = message.content.content %}\n {%- else %}\n {%- set content = message.content %}\n {%- endif %}\n {{- '[TOOL_RESULTS] {\"content\": ' + content|string + \", \" }}\n {%- if not message.tool_call_id is defined or message.tool_call_id|length != 9 %}\n {{- raise_exception(\"Tool call IDs should be alphanumeric strings with length 9!\") }}\n {%- endif %}\n {{- '\"call_id\": \"' + message.tool_call_id + '\"}[/TOOL_RESULTS]' }}\n {%- else %}\n {{- raise_exception(\"Only user and assistant roles are supported, with the exception of an initial optional system message!\") }}\n {%- endif %}\n{%- endfor %}\n",
231
- /* .expected_output= */ "[INST] You are a helpful assistant\n\nHello[/INST] Hi there</s>[INST] Who are you[/INST] I am an assistant</s>[INST] Another question[/INST]",
232
- /* .expected_output_jinja= */ "[INST] Hello[/INST] Hi there</s>[INST] Who are you[/INST] I am an assistant</s>[INST] You are a helpful assistant\n\nAnother question[/INST]",
233
- /* .bos_token= */ "",
234
- /* .eos_token= */ "</s>",
235
- },
236
- {
237
- /* .name= */ "Mistral-Nemo-Instruct-2407 (mistralai 'v3-tekken' template; modified to have system prompt at start)",
238
- /* .template_str= */ "{%- if messages[0][\"role\"] == \"system\" %}\n {%- set system_message = messages[0][\"content\"] %}\n {%- set loop_messages = messages[1:] %}\n{%- else %}\n {%- set loop_messages = messages %}\n{%- endif %}\n{%- if not tools is defined %}\n {%- set tools = none %}\n{%- endif %}\n{%- set user_messages = loop_messages | selectattr(\"role\", \"equalto\", \"user\") | list %}\n\n{#- This block checks for alternating user/assistant messages, skipping tool calling messages #}\n{%- set ns = namespace() %}\n{%- set ns.index = 0 %}\n{%- for message in loop_messages %}\n {%- if not (message.role == \"tool\" or message.role == \"tool_results\" or (message.tool_calls is defined and message.tool_calls is not none)) %}\n {%- if (message[\"role\"] == \"user\") != (ns.index % 2 == 0) %}\n {{- raise_exception(\"After the optional system message, conversation roles must alternate user/assistant/user/assistant/...\") }}\n {%- endif %}\n {%- set ns.index = ns.index + 1 %}\n {%- endif %}\n{%- endfor %}\n\n{{- bos_token }}\n{%- for message in loop_messages %}\n {%- if message[\"role\"] == \"user\" %}\n {%- if tools is not none and (message == user_messages[-1]) %}\n {{- \"[AVAILABLE_TOOLS][\" }}\n {%- for tool in tools %}\n {%- set tool = tool.function %}\n {{- '{\"type\": \"function\", \"function\": {' }}\n {%- for key, val in tool.items() if key != \"return\" %}\n {%- if val is string %}\n {{- '\"' + key + '\": \"' + val + '\"' }}\n {%- else %}\n {{- '\"' + key + '\": ' + val|tojson }}\n {%- endif %}\n {%- if not loop.last %}\n {{- \", \" }}\n {%- endif %}\n {%- endfor %}\n {{- \"}}\" }}\n {%- if not loop.last %}\n {{- \", \" }}\n {%- else %}\n {{- \"]\" }}\n {%- endif %}\n {%- endfor %}\n {{- \"[/AVAILABLE_TOOLS]\" }}\n {%- endif %}\n {%- if loop.last and system_message is defined %}\n {{- \"[INST]\" + system_message + \"\\n\\n\" + message[\"content\"] + \"[/INST]\" }}\n {%- else %}\n {{- \"[INST]\" + message[\"content\"] + \"[/INST]\" }}\n {%- endif %}\n {%- elif (message.tool_calls is defined and message.tool_calls is not none) %}\n {{- \"[TOOL_CALLS][\" }}\n {%- for tool_call in message.tool_calls %}\n {%- set out = tool_call.function|tojson %}\n {{- out[:-1] }}\n {%- if not tool_call.id is defined or tool_call.id|length != 9 %}\n {{- raise_exception(\"Tool call IDs should be alphanumeric strings with length 9!\") }}\n {%- endif %}\n {{- ', \"id\": \"' + tool_call.id + '\"}' }}\n {%- if not loop.last %}\n {{- \", \" }}\n {%- else %}\n {{- \"]\" + eos_token }}\n {%- endif %}\n {%- endfor %}\n {%- elif message[\"role\"] == \"assistant\" %}\n {{- message[\"content\"] + eos_token}}\n {%- elif message[\"role\"] == \"tool_results\" or message[\"role\"] == \"tool\" %}\n {%- if message.content is defined and message.content.content is defined %}\n {%- set content = message.content.content %}\n {%- else %}\n {%- set content = message.content %}\n {%- endif %}\n {{- '[TOOL_RESULTS]{\"content\": ' + content|string + \", \" }}\n {%- if not message.tool_call_id is defined or message.tool_call_id|length != 9 %}\n {{- raise_exception(\"Tool call IDs should be alphanumeric strings with length 9!\") }}\n {%- endif %}\n {{- '\"call_id\": \"' + message.tool_call_id + '\"}[/TOOL_RESULTS]' }}\n {%- else %}\n {{- raise_exception(\"Only user and assistant roles are supported, with the exception of an initial optional system message!\") }}\n {%- endif %}\n{%- endfor %}\n",
239
- /* .expected_output= */ "[INST]You are a helpful assistant\n\nHello[/INST]Hi there</s>[INST]Who are you[/INST] I am an assistant </s>[INST]Another question[/INST]",
240
- /* .expected_output_jinja= */ "[INST]Hello[/INST]Hi there</s>[INST]Who are you[/INST] I am an assistant </s>[INST]You are a helpful assistant\n\nAnother question[/INST]",
241
- /* .bos_token= */ "",
242
- /* .eos_token= */ "</s>",
243
- },
244
- {
245
- /* .name= */ "mistralai/Mistral-Large-Instruct-2411 (mistralai 'v7' template)",
246
- /* .template_str= */ "{{ bos_token }}{% for message in messages %}{% if message['role'] == 'user' %}{{ '[INST] ' + message['content'] + '[/INST]' }}{% elif message['role'] == 'system' %}{{ '[SYSTEM_PROMPT] ' + message['content'] + '[/SYSTEM_PROMPT]' }}{% elif message['role'] == 'assistant' %}{{ ' ' + message['content'] + eos_token }}{% else %}{{ raise_exception('Only user, system and assistant roles are supported!') }}{% endif %}{% endfor %}",
247
- /* .expected_output= */ "[SYSTEM_PROMPT] You are a helpful assistant[/SYSTEM_PROMPT][INST] Hello[/INST] Hi there</s>[INST] Who are you[/INST] I am an assistant </s>[INST] Another question[/INST]",
248
- /* .expected_output_jinja= */ "",
249
- /* .bos_token= */ "",
250
- /* .eos_token= */ "</s>",
251
- },
252
- {
253
- /* .name= */ "ai-sage/GigaChat-20B-A3B-instruct",
254
- /* .template_str= */ "{% if messages[0]['role'] == 'system' -%}\n {%- set loop_messages = messages[1:] -%}\n {%- set system_message = bos_token + messages[0]['content'] + additional_special_tokens[1] -%}\n{%- else -%}\n {%- set loop_messages = messages -%}\n {%- set system_message = bos_token + '' -%}\n{%- endif -%}\n{%- for message in loop_messages %}\n {% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}\n {{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}\n {% endif %}\n \n {%- if loop.index0 == 0 -%}\n {{ system_message -}}\n {%- endif -%}\n {%- if message['role'] == 'user' -%}\n {{ message['role'] + additional_special_tokens[0] + message['content'] + additional_special_tokens[1] -}}\n {{ 'available functions' + additional_special_tokens[0] + additional_special_tokens[2] + additional_special_tokens[3] + additional_special_tokens[1] -}}\n {%- endif -%}\n {%- if message['role'] == 'assistant' -%}\n {{ message['role'] + additional_special_tokens[0] + message['content'] + additional_special_tokens[1] -}}\n {%- endif -%}\n {%- if loop.last and add_generation_prompt -%}\n {{ 'assistant' + additional_special_tokens[0] -}}\n {%- endif -%}\n{%- endfor %}",
255
- /* .expected_output= */ "<s>You are a helpful assistant<|message_sep|>user<|role_sep|>Hello<|message_sep|>available functions<|role_sep|>[]<|message_sep|>assistant<|role_sep|>Hi there<|message_sep|>user<|role_sep|>Who are you<|message_sep|>available functions<|role_sep|>[]<|message_sep|>assistant<|role_sep|> I am an assistant <|message_sep|>user<|role_sep|>Another question<|message_sep|>available functions<|role_sep|>[]<|message_sep|>assistant<|role_sep|>",
256
- /* .expected_output_jinja= */ "",
257
- /* .bos_token= */ "",
258
- /* .eos_token= */ "",
259
- /* .supported_with_jinja= */ false, // Requires additional_special_tokens as extra context
260
- },
261
- {
262
- /* .name= */ "Infinigence/Megrez-3B-Instruct",
263
- /* .template_str= */ U8C("{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|role_start|>system<|role_end|>你是Megrez-3B-Instruct,将针对用户的问题给出详细的、积极的回答。<|turn_end|>' }}{% endif %}{{ '<|role_start|>' + message['role'] + '<|role_end|>' + message['content'] + '<|turn_end|>' }}{% endfor %}{% if add_generation_prompt %}{{ '<|role_start|>assistant<|role_end|>' }}{% endif %}"),
264
- /* .expected_output= */ "<|role_start|>system<|role_end|>You are a helpful assistant<|turn_end|><|role_start|>user<|role_end|>Hello<|turn_end|><|role_start|>assistant<|role_end|>Hi there<|turn_end|><|role_start|>user<|role_end|>Who are you<|turn_end|><|role_start|>assistant<|role_end|> I am an assistant <|turn_end|><|role_start|>user<|role_end|>Another question<|turn_end|><|role_start|>assistant<|role_end|>",
265
- /* .expected_output_jinja= */ "",
266
- /* .bos_token= */ "",
267
- /* .eos_token= */ "",
268
- },
269
- {
270
- /* .name= */ "phi-4",
271
- /* .template_str= */ "{% for message in messages %}{% if (message['role'] == 'system') %}{{'<|im_start|>system<|im_sep|>' + message['content'] + '<|im_end|>'}}{% elif (message['role'] == 'user') %}{{'<|im_start|>user<|im_sep|>' + message['content'] + '<|im_end|><|im_start|>assistant<|im_sep|>'}}{% elif (message['role'] == 'assistant') %}{{message['content'] + '<|im_end|>'}}{% endif %}{% endfor %}",
272
- /* .expected_output= */ "<|im_start|>system<|im_sep|>You are a helpful assistant<|im_end|><|im_start|>user<|im_sep|>Hello<|im_end|><|im_start|>assistant<|im_sep|>Hi there<|im_end|><|im_start|>user<|im_sep|>Who are you<|im_end|><|im_start|>assistant<|im_sep|> I am an assistant <|im_end|><|im_start|>user<|im_sep|>Another question<|im_end|><|im_start|>assistant<|im_sep|>",
273
- /* .expected_output_jinja= */ "",
274
- /* .bos_token= */ "",
275
- /* .eos_token= */ "",
276
- },
277
- {
278
- /* .name= */ "yandex/YandexGPT-5-Lite-8B-instruct",
279
- /* .template_str= */ "<s>{%- set names = {'assistant': ' Ассистент:', 'user': ' Пользователь:'} %}\n{%- set tools_prefix = 'Тебе доступны следующие функции:' %}\n{%- macro __render_tool(tool) %}\n {%- set name = tool.function.name %}\n {%- set description = tool.function.description|default('') %}\n {%- set parameters = tool.function.parameters|tojson %}\n {{- '\\n' }}function {{ '{' }}'name':'{{ name }}',\n {%- if tool.function.description %}'description':'{{ description }}',{% endif %}\n'parameters':{{ parameters }}\n {{- '}' }}\n{%- endmacro %}\n{%- macro __render_tools(tools) %}\n {{- tools_prefix }}\n {%- for tool in tools %}\n {{- __render_tool(tool) }}\n {%- endfor %}\n {{- '\\n\\n' }}\n{%- endmacro %}\n{%- macro __render_tool_message(message) %}\n {{- '\\n\\nРезультат вызова' }} {{ message.name }}: {{ message.content }} {{ '\\n\\n' }}\n{%- endmacro %}\n{%- if tools -%}\n {{- __render_tools(tools) }}\n{%- endif -%}\n{%- macro __render_user_message(message) %}\n{{ names.user }} {{ message.content + '\\n\\n' }}\n{%- endmacro %}\n{%- macro __render_assistant_message(message) %}\n {{- names.assistant }}\n {%- set call = message['function_call'] %}\n {%- if call %}\n {{- '\\n[TOOL_CALL_START]' }}{{ call.name }}{{ '\\n' }}{{ call.arguments|tojson }}\n {%- else %}\n {{- ' ' + message.content + '\\n\\n' }}\n {%- endif %}\n{%- endmacro %}\n{%- if not add_generation_prompt is defined %}\n{%- set add_generation_prompt = false %}\n{%- endif %}\n{%- for message in messages %}\n {%- if message['role'] == 'user' %}\n {{- __render_user_message(message) }}\n {%- endif %}\n {%- if message.role == 'assistant' and not loop.last %}\n {{- __render_assistant_message(message) }}\n {%- endif %}\n {%- if message.role == 'tool' %}\n {{- __render_tool_message(message) }}\n {%- endif %}\n {%- if loop.last %}\n {{- ' Ассистент:[SEP]' }}\n {%- endif %}\n{%- endfor %}\n",
280
- /* .expected_output= */ "<s> Пользователь: Hello\n\n Ассистент: Hi there\n\n Пользователь: Who are you\n\n Ассистент: I am an assistant \n\n Пользователь: Another question\n\n Ассистент:[SEP]",
281
- /* .expected_output_jinja= */ "<s> Пользователь: You are a helpful assistant\nHello\n\n Ассистент: Hi there\n\n Пользователь: Who are you\n\n Ассистент: I am an assistant \n\n Пользователь: Another question\n\n Ассистент:[SEP]",
282
- /* .bos_token= */ "",
283
- /* .eos_token= */ "",
284
- },
285
- {
286
- /* .name= */ "inclusionAI/Ling-lite",
287
- /* .template_str */ "{% for message in messages %}{% set role = message['role'] | lower %}{% if role == 'user' %}{% set role = 'HUMAN' %}{% endif %}{% set role = role | upper %}{{ '<role>' + role + '</role>' + message['content'] }}{% endfor %}{% if add_generation_prompt %}{{ '<role>ASSISTANT</role>' }}{% endif %}",
288
- /* .expected_output= */ "<role>SYSTEM</role>You are a helpful assistant<role>HUMAN</role>Hello<role>ASSISTANT</role>Hi there<role>HUMAN</role>Who are you<role>ASSISTANT</role> I am an assistant <role>HUMAN</role>Another question<role>ASSISTANT</role>",
289
- /* .expected_output_jinja= */ "",
290
- /* .bos_token= */ "",
291
- /* .eos_token= */ "",
292
- },
293
- };
294
- std::vector<char> formatted_chat(1024);
295
- int32_t res;
296
-
297
- // list all supported templates
298
- std::vector<const char *> supported_tmpl;
299
- res = llama_chat_builtin_templates(nullptr, 0);
300
- assert(res > 0);
301
- supported_tmpl.resize(res);
302
- res = llama_chat_builtin_templates(supported_tmpl.data(), supported_tmpl.size());
303
- printf("Built-in chat templates:\n");
304
- for (auto tmpl : supported_tmpl) {
305
- printf(" %s\n", tmpl);
306
- }
307
-
308
- // test invalid chat template
309
- res = llama_chat_apply_template("INVALID TEMPLATE", conversation.data(), conversation.size(), true, formatted_chat.data(), formatted_chat.size());
310
- assert(res < 0);
311
- const auto add_generation_prompt = true;
312
-
313
- for (const auto & test_case : test_cases) {
314
- printf("\n\n=== %s ===\n\n", test_case.name.c_str());
315
- formatted_chat.resize(1024);
316
- res = llama_chat_apply_template(
317
- test_case.template_str.c_str(),
318
- conversation.data(),
319
- conversation.size(),
320
- add_generation_prompt,
321
- formatted_chat.data(),
322
- formatted_chat.size()
323
- );
324
- formatted_chat.resize(res);
325
- std::string output(formatted_chat.data(), formatted_chat.size());
326
- if (output != test_case.expected_output) {
327
- printf("Expected:\n%s\n", test_case.expected_output.c_str());
328
- printf("-------------------------\n");
329
- printf("Actual:\n%s\n", output.c_str());
330
- fflush(stdout);
331
- assert(output == test_case.expected_output);
332
- }
333
- }
334
-
335
- std::vector<common_chat_msg> messages;
336
- for (const auto & msg : conversation) {
337
- messages.push_back(simple_msg(msg.role, msg.content));
338
- }
339
- for (const auto & test_case : test_cases) {
340
- if (!test_case.supported_with_jinja) {
341
- continue;
342
- }
343
- printf("\n\n=== %s (jinja) ===\n\n", test_case.name.c_str());
344
- try {
345
- auto tmpls = common_chat_templates_init(/* model= */ nullptr, test_case.template_str.c_str(), test_case.bos_token, test_case.eos_token);
346
- common_chat_templates_inputs inputs;
347
- inputs.use_jinja = true;
348
- inputs.messages = messages;
349
- inputs.add_generation_prompt = add_generation_prompt;
350
- auto output = common_chat_templates_apply(tmpls.get(), inputs).prompt;
351
- output = normalize_newlines(output);
352
- auto expected_output = normalize_newlines(test_case.expected_output_jinja.empty() ? test_case.expected_output : test_case.expected_output_jinja);
353
- if (output != expected_output) {
354
- printf("Expected:\n%s\n", expected_output.c_str());
355
- printf("-------------------------\n");
356
- printf("Actual:\n%s\n", output.c_str());
357
- fflush(stdout);
358
- assert(output == expected_output);
359
- }
360
- } catch (const std::exception & e) {
361
- printf("ERROR: %s\n", e.what());
362
- assert(false);
363
- }
364
- }
365
-
366
- // test llama_chat_format_single for system message
367
- printf("\n\n=== llama_chat_format_single (system message) ===\n\n");
368
- std::vector<common_chat_msg> chat2;
369
- auto sys_msg = simple_msg("system", "You are a helpful assistant");
370
-
371
- auto fmt_sys = [&](std::string tmpl_str) {
372
- auto tmpls = common_chat_templates_init(/* model= */ nullptr, tmpl_str);
373
- auto output = common_chat_format_single(tmpls.get(), chat2, sys_msg, false, /* use_jinja= */ false);
374
- printf("fmt_sys(%s) : %s\n", tmpl_str.c_str(), output.c_str());
375
- printf("-------------------------\n");
376
- return output;
377
- };
378
- assert(fmt_sys("chatml") == "<|im_start|>system\nYou are a helpful assistant<|im_end|>\n");
379
- assert(fmt_sys("mistral-v1") == " [INST] You are a helpful assistant\n\n");
380
- assert(fmt_sys("mistral-v3") == "[INST] You are a helpful assistant\n\n");
381
- assert(fmt_sys("mistral-v3-tekken") == "[INST]You are a helpful assistant\n\n");
382
- assert(fmt_sys("mistral-v7") == "[SYSTEM_PROMPT] You are a helpful assistant[/SYSTEM_PROMPT]");
383
- assert(fmt_sys("llama2") == "[INST] You are a helpful assistant\n");
384
- assert(fmt_sys("llama2-sys") == "[INST] <<SYS>>\nYou are a helpful assistant\n<</SYS>>\n\n");
385
- assert(fmt_sys("mistral") == "[INST] You are a helpful assistant\n"); // for old pre-v1 templates
386
- assert(fmt_sys("gemma") == ""); // for gemma, system message is merged with user message
387
- assert(fmt_sys("llama3") == "<|start_header_id|>system<|end_header_id|>\n\nYou are a helpful assistant<|eot_id|>");
388
- assert(fmt_sys("gigachat") == "<s>You are a helpful assistant<|message_sep|>");
389
-
390
-
391
- // test llama_chat_format_single for user message
392
- printf("\n\n=== llama_chat_format_single (user message) ===\n\n");
393
- chat2.push_back(simple_msg("system", "You are a helpful assistant"));
394
- chat2.push_back(simple_msg("user", "Hello"));
395
- chat2.push_back(simple_msg("assistant", "I am assistant"));
396
- auto new_msg = simple_msg("user", "How are you");
397
-
398
- auto fmt_single = [&](const std::string & tmpl_str) {
399
- auto tmpls = common_chat_templates_init(/* model= */ nullptr, tmpl_str.c_str());
400
- auto output = common_chat_format_single(tmpls.get(), chat2, new_msg, true, /* use_jinja= */ false);
401
- printf("fmt_single(%s) : %s\n", tmpl_str.c_str(), output.c_str());
402
- printf("-------------------------\n");
403
- return output;
404
- };
405
- assert(fmt_single("chatml") == "\n<|im_start|>user\nHow are you<|im_end|>\n<|im_start|>assistant\n");
406
- assert(fmt_single("mistral-v1") == " [INST] How are you [/INST]");
407
- assert(fmt_single("mistral-v3") == "[INST] How are you[/INST]");
408
- assert(fmt_single("mistral-v3-tekken") == "[INST]How are you[/INST]");
409
- assert(fmt_single("mistral-v7") == "[INST] How are you[/INST]");
410
- assert(fmt_single("llama2") == "[INST] How are you [/INST]");
411
- assert(fmt_single("mistral") == "[INST] How are you [/INST]"); // for old pre-v1 templates
412
- assert(fmt_single("gemma") == "\n<start_of_turn>user\nHow are you<end_of_turn>\n<start_of_turn>model\n");
413
- assert(fmt_single("llama3") == "<|start_header_id|>user<|end_header_id|>\n\nHow are you<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n");
414
- assert(fmt_single("gigachat") == "user<|role_sep|>How are you<|message_sep|>available functions<|role_sep|>[]<|message_sep|>assistant<|role_sep|>");
415
-
416
- return 0;
417
- }