node-llama-cpp 3.0.0-beta.4 → 3.0.0-beta.41

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 (749) hide show
  1. package/README.md +14 -11
  2. package/bins/_linux-x64-cuda.moved.txt +1 -0
  3. package/bins/_win-x64-cuda.moved.txt +1 -0
  4. package/bins/linux-arm64/_nlcBuildMetadata.json +1 -0
  5. package/bins/linux-arm64/libggml.so +0 -0
  6. package/bins/linux-arm64/libllama.so +0 -0
  7. package/bins/linux-arm64/llama-addon.node +0 -0
  8. package/bins/linux-armv7l/_nlcBuildMetadata.json +1 -0
  9. package/bins/linux-armv7l/libggml.so +0 -0
  10. package/bins/linux-armv7l/libllama.so +0 -0
  11. package/bins/linux-armv7l/llama-addon.node +0 -0
  12. package/bins/linux-x64/_nlcBuildMetadata.json +1 -0
  13. package/bins/linux-x64/libggml.so +0 -0
  14. package/bins/linux-x64/libllama.so +0 -0
  15. package/bins/linux-x64/llama-addon.node +0 -0
  16. package/bins/linux-x64-vulkan/_nlcBuildMetadata.json +1 -0
  17. package/bins/linux-x64-vulkan/libggml.so +0 -0
  18. package/bins/linux-x64-vulkan/libllama.so +0 -0
  19. package/bins/linux-x64-vulkan/llama-addon.node +0 -0
  20. package/bins/linux-x64-vulkan/vulkan-shaders-gen +0 -0
  21. package/bins/mac-arm64-metal/_nlcBuildMetadata.json +1 -0
  22. package/bins/mac-arm64-metal/ggml-common.h +1833 -0
  23. package/bins/mac-arm64-metal/ggml-metal.metal +6168 -0
  24. package/bins/mac-arm64-metal/libggml.dylib +0 -0
  25. package/bins/mac-arm64-metal/libllama.dylib +0 -0
  26. package/bins/mac-arm64-metal/llama-addon.node +0 -0
  27. package/bins/mac-x64/_nlcBuildMetadata.json +1 -0
  28. package/bins/mac-x64/libggml.dylib +0 -0
  29. package/bins/mac-x64/libllama.dylib +0 -0
  30. package/bins/mac-x64/llama-addon.node +0 -0
  31. package/bins/win-arm64/_nlcBuildMetadata.json +1 -0
  32. package/bins/win-arm64/ggml.dll +0 -0
  33. package/bins/win-arm64/llama-addon.exp +0 -0
  34. package/bins/win-arm64/llama-addon.lib +0 -0
  35. package/bins/win-arm64/llama-addon.node +0 -0
  36. package/bins/win-arm64/llama.dll +0 -0
  37. package/bins/win-x64/_nlcBuildMetadata.json +1 -0
  38. package/bins/win-x64/ggml.dll +0 -0
  39. package/bins/win-x64/llama-addon.exp +0 -0
  40. package/bins/win-x64/llama-addon.lib +0 -0
  41. package/bins/win-x64/llama-addon.node +0 -0
  42. package/bins/win-x64/llama.dll +0 -0
  43. package/bins/win-x64-vulkan/_nlcBuildMetadata.json +1 -0
  44. package/bins/win-x64-vulkan/ggml.dll +0 -0
  45. package/bins/win-x64-vulkan/llama-addon.exp +0 -0
  46. package/bins/win-x64-vulkan/llama-addon.lib +0 -0
  47. package/bins/win-x64-vulkan/llama-addon.node +0 -0
  48. package/bins/win-x64-vulkan/llama.dll +0 -0
  49. package/bins/win-x64-vulkan/vulkan-shaders-gen.exe +0 -0
  50. package/dist/ChatWrapper.d.ts +9 -39
  51. package/dist/ChatWrapper.js +129 -72
  52. package/dist/ChatWrapper.js.map +1 -1
  53. package/dist/apiDocsIndex.d.ts +1 -0
  54. package/dist/apiDocsIndex.js +7 -0
  55. package/dist/apiDocsIndex.js.map +1 -0
  56. package/dist/{utils/getBin.d.ts → bindings/AddonTypes.d.ts} +66 -9
  57. package/dist/bindings/AddonTypes.js +2 -0
  58. package/dist/bindings/AddonTypes.js.map +1 -0
  59. package/dist/bindings/Llama.d.ts +47 -0
  60. package/dist/bindings/Llama.js +356 -0
  61. package/dist/bindings/Llama.js.map +1 -0
  62. package/dist/bindings/consts.d.ts +2 -0
  63. package/dist/bindings/consts.js +11 -0
  64. package/dist/bindings/consts.js.map +1 -0
  65. package/dist/bindings/getLlama.d.ts +152 -0
  66. package/dist/bindings/getLlama.js +403 -0
  67. package/dist/bindings/getLlama.js.map +1 -0
  68. package/dist/bindings/types.d.ts +57 -0
  69. package/dist/bindings/types.js +77 -0
  70. package/dist/bindings/types.js.map +1 -0
  71. package/dist/bindings/utils/MemoryOrchestrator.d.ts +21 -0
  72. package/dist/bindings/utils/MemoryOrchestrator.js +49 -0
  73. package/dist/bindings/utils/MemoryOrchestrator.js.map +1 -0
  74. package/dist/bindings/utils/NoBinaryFoundError.d.ts +2 -0
  75. package/dist/bindings/utils/NoBinaryFoundError.js +7 -0
  76. package/dist/bindings/utils/NoBinaryFoundError.js.map +1 -0
  77. package/dist/bindings/utils/asyncEvery.d.ts +5 -0
  78. package/dist/bindings/utils/asyncEvery.js +15 -0
  79. package/dist/bindings/utils/asyncEvery.js.map +1 -0
  80. package/dist/bindings/utils/asyncSome.d.ts +5 -0
  81. package/dist/bindings/utils/asyncSome.js +27 -0
  82. package/dist/bindings/utils/asyncSome.js.map +1 -0
  83. package/dist/{utils → bindings/utils}/binariesGithubRelease.js +1 -1
  84. package/dist/bindings/utils/binariesGithubRelease.js.map +1 -0
  85. package/dist/bindings/utils/clearAllLocalBuilds.d.ts +1 -0
  86. package/dist/bindings/utils/clearAllLocalBuilds.js +47 -0
  87. package/dist/bindings/utils/clearAllLocalBuilds.js.map +1 -0
  88. package/dist/bindings/utils/cloneLlamaCppRepo.d.ts +11 -0
  89. package/dist/bindings/utils/cloneLlamaCppRepo.js +166 -0
  90. package/dist/bindings/utils/cloneLlamaCppRepo.js.map +1 -0
  91. package/dist/bindings/utils/compileLLamaCpp.d.ts +21 -0
  92. package/dist/bindings/utils/compileLLamaCpp.js +288 -0
  93. package/dist/bindings/utils/compileLLamaCpp.js.map +1 -0
  94. package/dist/bindings/utils/detectAvailableComputeLayers.d.ts +14 -0
  95. package/dist/bindings/utils/detectAvailableComputeLayers.js +305 -0
  96. package/dist/bindings/utils/detectAvailableComputeLayers.js.map +1 -0
  97. package/dist/bindings/utils/detectGlibc.d.ts +4 -0
  98. package/dist/bindings/utils/detectGlibc.js +46 -0
  99. package/dist/bindings/utils/detectGlibc.js.map +1 -0
  100. package/dist/bindings/utils/getBestComputeLayersAvailable.d.ts +10 -0
  101. package/dist/bindings/utils/getBestComputeLayersAvailable.js +29 -0
  102. package/dist/bindings/utils/getBestComputeLayersAvailable.js.map +1 -0
  103. package/dist/bindings/utils/getBuildFolderNameForBuildOptions.d.ts +5 -0
  104. package/dist/bindings/utils/getBuildFolderNameForBuildOptions.js +93 -0
  105. package/dist/bindings/utils/getBuildFolderNameForBuildOptions.js.map +1 -0
  106. package/dist/bindings/utils/getCanUsePrebuiltBinaries.d.ts +1 -0
  107. package/dist/bindings/utils/getCanUsePrebuiltBinaries.js +8 -0
  108. package/dist/bindings/utils/getCanUsePrebuiltBinaries.js.map +1 -0
  109. package/dist/bindings/utils/getExampleUsageCodeOfGetLlama.d.ts +2 -0
  110. package/dist/bindings/utils/getExampleUsageCodeOfGetLlama.js +21 -0
  111. package/dist/bindings/utils/getExampleUsageCodeOfGetLlama.js.map +1 -0
  112. package/dist/bindings/utils/getGpuTypesToUseForOption.d.ts +13 -0
  113. package/dist/bindings/utils/getGpuTypesToUseForOption.js +39 -0
  114. package/dist/bindings/utils/getGpuTypesToUseForOption.js.map +1 -0
  115. package/dist/bindings/utils/getLinuxDistroInfo.d.ts +9 -0
  116. package/dist/bindings/utils/getLinuxDistroInfo.js +46 -0
  117. package/dist/bindings/utils/getLinuxDistroInfo.js.map +1 -0
  118. package/dist/bindings/utils/getLlamaWithoutBackend.d.ts +5 -0
  119. package/dist/bindings/utils/getLlamaWithoutBackend.js +27 -0
  120. package/dist/bindings/utils/getLlamaWithoutBackend.js.map +1 -0
  121. package/dist/bindings/utils/getPlatform.d.ts +2 -0
  122. package/dist/bindings/utils/getPlatform.js +15 -0
  123. package/dist/bindings/utils/getPlatform.js.map +1 -0
  124. package/dist/bindings/utils/getPlatformInfo.d.ts +5 -0
  125. package/dist/bindings/utils/getPlatformInfo.js +28 -0
  126. package/dist/bindings/utils/getPlatformInfo.js.map +1 -0
  127. package/dist/bindings/utils/hasBuildingFromSourceDependenciesInstalled.d.ts +3 -0
  128. package/dist/bindings/utils/hasBuildingFromSourceDependenciesInstalled.js +27 -0
  129. package/dist/bindings/utils/hasBuildingFromSourceDependenciesInstalled.js.map +1 -0
  130. package/dist/bindings/utils/hasFileInPath.d.ts +2 -0
  131. package/dist/bindings/utils/hasFileInPath.js +34 -0
  132. package/dist/bindings/utils/hasFileInPath.js.map +1 -0
  133. package/dist/bindings/utils/lastBuildInfo.d.ts +6 -0
  134. package/dist/bindings/utils/lastBuildInfo.js +17 -0
  135. package/dist/bindings/utils/lastBuildInfo.js.map +1 -0
  136. package/dist/bindings/utils/logBinaryUsageExampleToConsole.d.ts +2 -0
  137. package/dist/bindings/utils/logBinaryUsageExampleToConsole.js +22 -0
  138. package/dist/bindings/utils/logBinaryUsageExampleToConsole.js.map +1 -0
  139. package/dist/bindings/utils/logDistroInstallInstruction.d.ts +14 -0
  140. package/dist/bindings/utils/logDistroInstallInstruction.js +48 -0
  141. package/dist/bindings/utils/logDistroInstallInstruction.js.map +1 -0
  142. package/dist/bindings/utils/resolveCustomCmakeOptions.d.ts +1 -0
  143. package/dist/bindings/utils/resolveCustomCmakeOptions.js +43 -0
  144. package/dist/bindings/utils/resolveCustomCmakeOptions.js.map +1 -0
  145. package/dist/bindings/utils/testBindingBinary.d.ts +1 -0
  146. package/dist/bindings/utils/testBindingBinary.js +100 -0
  147. package/dist/bindings/utils/testBindingBinary.js.map +1 -0
  148. package/dist/bindings/utils/testCmakeBinary.d.ts +6 -0
  149. package/dist/bindings/utils/testCmakeBinary.js +32 -0
  150. package/dist/bindings/utils/testCmakeBinary.js.map +1 -0
  151. package/dist/chatWrappers/AlpacaChatWrapper.d.ts +2 -1
  152. package/dist/chatWrappers/AlpacaChatWrapper.js +9 -2
  153. package/dist/chatWrappers/AlpacaChatWrapper.js.map +1 -1
  154. package/dist/chatWrappers/ChatMLChatWrapper.d.ts +2 -9
  155. package/dist/chatWrappers/ChatMLChatWrapper.js +23 -21
  156. package/dist/chatWrappers/ChatMLChatWrapper.js.map +1 -1
  157. package/dist/chatWrappers/FalconChatWrapper.d.ts +4 -10
  158. package/dist/chatWrappers/FalconChatWrapper.js +38 -21
  159. package/dist/chatWrappers/FalconChatWrapper.js.map +1 -1
  160. package/dist/chatWrappers/FunctionaryChatWrapper.d.ts +8 -32
  161. package/dist/chatWrappers/FunctionaryChatWrapper.js +326 -118
  162. package/dist/chatWrappers/FunctionaryChatWrapper.js.map +1 -1
  163. package/dist/chatWrappers/GemmaChatWrapper.d.ts +7 -0
  164. package/dist/chatWrappers/GemmaChatWrapper.js +96 -0
  165. package/dist/chatWrappers/GemmaChatWrapper.js.map +1 -0
  166. package/dist/chatWrappers/GeneralChatWrapper.d.ts +4 -10
  167. package/dist/chatWrappers/GeneralChatWrapper.js +45 -22
  168. package/dist/chatWrappers/GeneralChatWrapper.js.map +1 -1
  169. package/dist/chatWrappers/Llama2ChatWrapper.d.ts +12 -0
  170. package/dist/chatWrappers/{LlamaChatWrapper.js → Llama2ChatWrapper.js} +38 -20
  171. package/dist/chatWrappers/Llama2ChatWrapper.js.map +1 -0
  172. package/dist/chatWrappers/Llama3ChatWrapper.d.ts +16 -0
  173. package/dist/chatWrappers/Llama3ChatWrapper.js +173 -0
  174. package/dist/chatWrappers/Llama3ChatWrapper.js.map +1 -0
  175. package/dist/chatWrappers/Llama3_1ChatWrapper.d.ts +31 -0
  176. package/dist/chatWrappers/Llama3_1ChatWrapper.js +223 -0
  177. package/dist/chatWrappers/Llama3_1ChatWrapper.js.map +1 -0
  178. package/dist/chatWrappers/generic/JinjaTemplateChatWrapper.d.ts +76 -0
  179. package/dist/chatWrappers/generic/JinjaTemplateChatWrapper.js +371 -0
  180. package/dist/chatWrappers/generic/JinjaTemplateChatWrapper.js.map +1 -0
  181. package/dist/chatWrappers/generic/TemplateChatWrapper.d.ts +54 -0
  182. package/dist/chatWrappers/generic/TemplateChatWrapper.js +200 -0
  183. package/dist/chatWrappers/generic/TemplateChatWrapper.js.map +1 -0
  184. package/dist/chatWrappers/generic/utils/chatHistoryFunctionCallMessageTemplate.d.ts +23 -0
  185. package/dist/chatWrappers/generic/utils/chatHistoryFunctionCallMessageTemplate.js +45 -0
  186. package/dist/chatWrappers/generic/utils/chatHistoryFunctionCallMessageTemplate.js.map +1 -0
  187. package/dist/chatWrappers/utils/ChatModelFunctionsDocumentationGenerator.d.ts +57 -0
  188. package/dist/chatWrappers/utils/ChatModelFunctionsDocumentationGenerator.js +119 -0
  189. package/dist/chatWrappers/utils/ChatModelFunctionsDocumentationGenerator.js.map +1 -0
  190. package/dist/chatWrappers/utils/isJinjaTemplateEquivalentToSpecializedChatWrapper.d.ts +4 -0
  191. package/dist/chatWrappers/utils/isJinjaTemplateEquivalentToSpecializedChatWrapper.js +210 -0
  192. package/dist/chatWrappers/utils/isJinjaTemplateEquivalentToSpecializedChatWrapper.js.map +1 -0
  193. package/dist/chatWrappers/utils/jsonDumps.d.ts +7 -0
  194. package/dist/chatWrappers/utils/jsonDumps.js +18 -0
  195. package/dist/chatWrappers/utils/jsonDumps.js.map +1 -0
  196. package/dist/chatWrappers/utils/resolveChatWrapper.d.ts +71 -0
  197. package/dist/chatWrappers/utils/resolveChatWrapper.js +289 -0
  198. package/dist/chatWrappers/utils/resolveChatWrapper.js.map +1 -0
  199. package/dist/cli/cli.js +21 -7
  200. package/dist/cli/cli.js.map +1 -1
  201. package/dist/cli/commands/BuildCommand.d.ts +11 -4
  202. package/dist/cli/commands/BuildCommand.js +114 -41
  203. package/dist/cli/commands/BuildCommand.js.map +1 -1
  204. package/dist/cli/commands/ChatCommand.d.ts +19 -7
  205. package/dist/cli/commands/ChatCommand.js +306 -150
  206. package/dist/cli/commands/ChatCommand.js.map +1 -1
  207. package/dist/cli/commands/ClearCommand.d.ts +1 -1
  208. package/dist/cli/commands/ClearCommand.js +11 -12
  209. package/dist/cli/commands/ClearCommand.js.map +1 -1
  210. package/dist/cli/commands/CompleteCommand.d.ts +30 -0
  211. package/dist/cli/commands/CompleteCommand.js +374 -0
  212. package/dist/cli/commands/CompleteCommand.js.map +1 -0
  213. package/dist/cli/commands/DebugCommand.d.ts +7 -0
  214. package/dist/cli/commands/DebugCommand.js +54 -0
  215. package/dist/cli/commands/DebugCommand.js.map +1 -0
  216. package/dist/cli/commands/DownloadCommand.d.ts +7 -4
  217. package/dist/cli/commands/DownloadCommand.js +121 -70
  218. package/dist/cli/commands/DownloadCommand.js.map +1 -1
  219. package/dist/cli/commands/InfillCommand.d.ts +32 -0
  220. package/dist/cli/commands/InfillCommand.js +410 -0
  221. package/dist/cli/commands/InfillCommand.js.map +1 -0
  222. package/dist/cli/commands/InitCommand.d.ts +11 -0
  223. package/dist/cli/commands/InitCommand.js +195 -0
  224. package/dist/cli/commands/InitCommand.js.map +1 -0
  225. package/dist/cli/commands/OnPostInstallCommand.js +9 -10
  226. package/dist/cli/commands/OnPostInstallCommand.js.map +1 -1
  227. package/dist/cli/commands/PullCommand.d.ts +12 -0
  228. package/dist/cli/commands/PullCommand.js +117 -0
  229. package/dist/cli/commands/PullCommand.js.map +1 -0
  230. package/dist/cli/commands/inspect/InspectCommand.d.ts +4 -0
  231. package/dist/cli/commands/inspect/InspectCommand.js +19 -0
  232. package/dist/cli/commands/inspect/InspectCommand.js.map +1 -0
  233. package/dist/cli/commands/inspect/commands/InspectGgufCommand.d.ts +12 -0
  234. package/dist/cli/commands/inspect/commands/InspectGgufCommand.js +136 -0
  235. package/dist/cli/commands/inspect/commands/InspectGgufCommand.js.map +1 -0
  236. package/dist/cli/commands/inspect/commands/InspectGpuCommand.d.ts +4 -0
  237. package/dist/cli/commands/inspect/commands/InspectGpuCommand.js +180 -0
  238. package/dist/cli/commands/inspect/commands/InspectGpuCommand.js.map +1 -0
  239. package/dist/cli/commands/inspect/commands/InspectMeasureCommand.d.ts +18 -0
  240. package/dist/cli/commands/inspect/commands/InspectMeasureCommand.js +626 -0
  241. package/dist/cli/commands/inspect/commands/InspectMeasureCommand.js.map +1 -0
  242. package/dist/cli/projectTemplates.d.ts +7 -0
  243. package/dist/cli/projectTemplates.js +10 -0
  244. package/dist/cli/projectTemplates.js.map +1 -0
  245. package/dist/cli/recommendedModels.d.ts +2 -0
  246. package/dist/cli/recommendedModels.js +376 -0
  247. package/dist/cli/recommendedModels.js.map +1 -0
  248. package/dist/cli/startCreateCli.d.ts +2 -0
  249. package/dist/cli/startCreateCli.js +26 -0
  250. package/dist/cli/startCreateCli.js.map +1 -0
  251. package/dist/cli/utils/ConsoleInteraction.d.ts +23 -0
  252. package/dist/cli/utils/ConsoleInteraction.js +122 -0
  253. package/dist/cli/utils/ConsoleInteraction.js.map +1 -0
  254. package/dist/cli/utils/ConsoleTable.d.ts +23 -0
  255. package/dist/cli/utils/ConsoleTable.js +86 -0
  256. package/dist/cli/utils/ConsoleTable.js.map +1 -0
  257. package/dist/cli/utils/basicChooseFromListConsoleInteraction.d.ts +13 -0
  258. package/dist/cli/utils/basicChooseFromListConsoleInteraction.js +111 -0
  259. package/dist/cli/utils/basicChooseFromListConsoleInteraction.js.map +1 -0
  260. package/dist/cli/utils/consolePromptQuestion.d.ts +6 -0
  261. package/dist/cli/utils/consolePromptQuestion.js +82 -0
  262. package/dist/cli/utils/consolePromptQuestion.js.map +1 -0
  263. package/dist/cli/utils/getReadablePath.d.ts +1 -0
  264. package/dist/cli/utils/getReadablePath.js +14 -0
  265. package/dist/cli/utils/getReadablePath.js.map +1 -0
  266. package/dist/cli/utils/interactivelyAskForModel.d.ts +8 -0
  267. package/dist/cli/utils/interactivelyAskForModel.js +461 -0
  268. package/dist/cli/utils/interactivelyAskForModel.js.map +1 -0
  269. package/dist/cli/utils/logUsedGpuTypeOption.d.ts +2 -0
  270. package/dist/cli/utils/logUsedGpuTypeOption.js +9 -0
  271. package/dist/cli/utils/logUsedGpuTypeOption.js.map +1 -0
  272. package/dist/cli/utils/printCommonInfoLines.d.ts +9 -0
  273. package/dist/cli/utils/printCommonInfoLines.js +79 -0
  274. package/dist/cli/utils/printCommonInfoLines.js.map +1 -0
  275. package/dist/cli/utils/printInfoLine.d.ts +12 -0
  276. package/dist/cli/utils/printInfoLine.js +54 -0
  277. package/dist/cli/utils/printInfoLine.js.map +1 -0
  278. package/dist/cli/utils/projectTemplates.d.ts +19 -0
  279. package/dist/cli/utils/projectTemplates.js +47 -0
  280. package/dist/cli/utils/projectTemplates.js.map +1 -0
  281. package/dist/cli/utils/resolveCommandGgufPath.d.ts +5 -0
  282. package/dist/cli/utils/resolveCommandGgufPath.js +72 -0
  283. package/dist/cli/utils/resolveCommandGgufPath.js.map +1 -0
  284. package/dist/cli/utils/resolveHeaderFlag.d.ts +1 -0
  285. package/dist/cli/utils/resolveHeaderFlag.js +21 -0
  286. package/dist/cli/utils/resolveHeaderFlag.js.map +1 -0
  287. package/dist/cli/utils/resolveModelRecommendationFileOptions.d.ts +19 -0
  288. package/dist/cli/utils/resolveModelRecommendationFileOptions.js +7 -0
  289. package/dist/cli/utils/resolveModelRecommendationFileOptions.js.map +1 -0
  290. package/dist/cli/utils/splitAnsiToLines.d.ts +1 -0
  291. package/dist/cli/utils/splitAnsiToLines.js +32 -0
  292. package/dist/cli/utils/splitAnsiToLines.js.map +1 -0
  293. package/dist/cli/utils/withCliCommandDescriptionDocsUrl.d.ts +2 -0
  294. package/dist/cli/utils/withCliCommandDescriptionDocsUrl.js +23 -0
  295. package/dist/cli/utils/withCliCommandDescriptionDocsUrl.js.map +1 -0
  296. package/dist/commands.d.ts +1 -0
  297. package/dist/commands.js +3 -0
  298. package/dist/commands.js.map +1 -1
  299. package/dist/config.d.ts +38 -5
  300. package/dist/config.js +61 -16
  301. package/dist/config.js.map +1 -1
  302. package/dist/consts.d.ts +4 -0
  303. package/dist/consts.js +11 -0
  304. package/dist/consts.js.map +1 -0
  305. package/dist/evaluator/LlamaChat/LlamaChat.d.ts +270 -0
  306. package/dist/evaluator/LlamaChat/LlamaChat.js +1544 -0
  307. package/dist/evaluator/LlamaChat/LlamaChat.js.map +1 -0
  308. package/dist/evaluator/LlamaChat/utils/FunctionCallNameGrammar.d.ts +11 -0
  309. package/dist/evaluator/LlamaChat/utils/FunctionCallNameGrammar.js +55 -0
  310. package/dist/evaluator/LlamaChat/utils/FunctionCallNameGrammar.js.map +1 -0
  311. package/dist/evaluator/LlamaChat/utils/FunctionCallParamsGrammar.d.ts +16 -0
  312. package/dist/evaluator/LlamaChat/utils/FunctionCallParamsGrammar.js +45 -0
  313. package/dist/evaluator/LlamaChat/utils/FunctionCallParamsGrammar.js.map +1 -0
  314. package/dist/evaluator/LlamaChat/utils/LlamaFunctionCallValidationError.d.ts +8 -0
  315. package/dist/evaluator/LlamaChat/utils/LlamaFunctionCallValidationError.js +12 -0
  316. package/dist/evaluator/LlamaChat/utils/LlamaFunctionCallValidationError.js.map +1 -0
  317. package/dist/{llamaEvaluator → evaluator}/LlamaChat/utils/contextShiftStrategies/eraseFirstResponseAndKeepFirstSystemChatContextShiftStrategy.js +42 -16
  318. package/dist/evaluator/LlamaChat/utils/contextShiftStrategies/eraseFirstResponseAndKeepFirstSystemChatContextShiftStrategy.js.map +1 -0
  319. package/dist/evaluator/LlamaChatSession/LlamaChatSession.d.ts +310 -0
  320. package/dist/evaluator/LlamaChatSession/LlamaChatSession.js +425 -0
  321. package/dist/evaluator/LlamaChatSession/LlamaChatSession.js.map +1 -0
  322. package/dist/evaluator/LlamaChatSession/utils/LlamaChatSessionPromptCompletionEngine.d.ts +39 -0
  323. package/dist/evaluator/LlamaChatSession/utils/LlamaChatSessionPromptCompletionEngine.js +186 -0
  324. package/dist/evaluator/LlamaChatSession/utils/LlamaChatSessionPromptCompletionEngine.js.map +1 -0
  325. package/dist/{llamaEvaluator → evaluator}/LlamaChatSession/utils/defineChatSessionFunction.d.ts +3 -0
  326. package/dist/{llamaEvaluator → evaluator}/LlamaChatSession/utils/defineChatSessionFunction.js +3 -0
  327. package/dist/evaluator/LlamaChatSession/utils/defineChatSessionFunction.js.map +1 -0
  328. package/dist/evaluator/LlamaCompletion.d.ts +154 -0
  329. package/dist/evaluator/LlamaCompletion.js +424 -0
  330. package/dist/evaluator/LlamaCompletion.js.map +1 -0
  331. package/dist/{llamaEvaluator → evaluator}/LlamaContext/LlamaContext.d.ts +42 -22
  332. package/dist/{llamaEvaluator → evaluator}/LlamaContext/LlamaContext.js +338 -81
  333. package/dist/evaluator/LlamaContext/LlamaContext.js.map +1 -0
  334. package/dist/evaluator/LlamaContext/types.d.ts +175 -0
  335. package/dist/evaluator/LlamaContext/types.js.map +1 -0
  336. package/dist/evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies/firstInFirstOutStrategy.js.map +1 -0
  337. package/dist/evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies/maximumParallelismStrategy.js.map +1 -0
  338. package/dist/evaluator/LlamaContext/utils/resolveBatchItemsPrioritizationStrategy.d.ts +2 -0
  339. package/dist/{llamaEvaluator/LlamaContext/utils/resolveBatchItemsPrioritizingStrategy.js → evaluator/LlamaContext/utils/resolveBatchItemsPrioritizationStrategy.js} +4 -4
  340. package/dist/evaluator/LlamaContext/utils/resolveBatchItemsPrioritizationStrategy.js.map +1 -0
  341. package/dist/evaluator/LlamaEmbeddingContext.d.ts +51 -0
  342. package/dist/evaluator/LlamaEmbeddingContext.js +73 -0
  343. package/dist/evaluator/LlamaEmbeddingContext.js.map +1 -0
  344. package/dist/evaluator/LlamaGrammar.d.ts +34 -0
  345. package/dist/{llamaEvaluator → evaluator}/LlamaGrammar.js +15 -12
  346. package/dist/evaluator/LlamaGrammar.js.map +1 -0
  347. package/dist/{llamaEvaluator → evaluator}/LlamaGrammarEvaluationState.js +4 -4
  348. package/dist/evaluator/LlamaGrammarEvaluationState.js.map +1 -0
  349. package/dist/{llamaEvaluator → evaluator}/LlamaJsonSchemaGrammar.d.ts +2 -1
  350. package/dist/{llamaEvaluator → evaluator}/LlamaJsonSchemaGrammar.js +3 -3
  351. package/dist/evaluator/LlamaJsonSchemaGrammar.js.map +1 -0
  352. package/dist/evaluator/LlamaModel/LlamaModel.d.ts +242 -0
  353. package/dist/evaluator/LlamaModel/LlamaModel.js +765 -0
  354. package/dist/evaluator/LlamaModel/LlamaModel.js.map +1 -0
  355. package/dist/evaluator/LlamaModel/utils/TokenAttributes.d.ts +29 -0
  356. package/dist/evaluator/LlamaModel/utils/TokenAttributes.js +65 -0
  357. package/dist/evaluator/LlamaModel/utils/TokenAttributes.js.map +1 -0
  358. package/dist/evaluator/TokenBias.d.ts +22 -0
  359. package/dist/evaluator/TokenBias.js +33 -0
  360. package/dist/evaluator/TokenBias.js.map +1 -0
  361. package/dist/evaluator/TokenMeter.d.ts +54 -0
  362. package/dist/evaluator/TokenMeter.js +86 -0
  363. package/dist/evaluator/TokenMeter.js.map +1 -0
  364. package/dist/gguf/consts.d.ts +4 -0
  365. package/dist/gguf/consts.js +12 -0
  366. package/dist/gguf/consts.js.map +1 -0
  367. package/dist/gguf/errors/InvalidGgufMagicError.d.ts +3 -0
  368. package/dist/gguf/errors/InvalidGgufMagicError.js +6 -0
  369. package/dist/gguf/errors/InvalidGgufMagicError.js.map +1 -0
  370. package/dist/gguf/errors/UnsupportedGgufValueTypeError.d.ts +4 -0
  371. package/dist/gguf/errors/UnsupportedGgufValueTypeError.js +9 -0
  372. package/dist/gguf/errors/UnsupportedGgufValueTypeError.js.map +1 -0
  373. package/dist/gguf/fileReaders/GgufFileReader.d.ts +37 -0
  374. package/dist/gguf/fileReaders/GgufFileReader.js +109 -0
  375. package/dist/gguf/fileReaders/GgufFileReader.js.map +1 -0
  376. package/dist/gguf/fileReaders/GgufFsFileReader.d.ts +18 -0
  377. package/dist/gguf/fileReaders/GgufFsFileReader.js +62 -0
  378. package/dist/gguf/fileReaders/GgufFsFileReader.js.map +1 -0
  379. package/dist/gguf/fileReaders/GgufNetworkFetchFileReader.d.ts +23 -0
  380. package/dist/gguf/fileReaders/GgufNetworkFetchFileReader.js +79 -0
  381. package/dist/gguf/fileReaders/GgufNetworkFetchFileReader.js.map +1 -0
  382. package/dist/gguf/insights/GgufInsights.d.ts +50 -0
  383. package/dist/gguf/insights/GgufInsights.js +401 -0
  384. package/dist/gguf/insights/GgufInsights.js.map +1 -0
  385. package/dist/gguf/insights/GgufInsightsConfigurationResolver.d.ts +90 -0
  386. package/dist/gguf/insights/GgufInsightsConfigurationResolver.js +144 -0
  387. package/dist/gguf/insights/GgufInsightsConfigurationResolver.js.map +1 -0
  388. package/dist/gguf/insights/utils/resolveContextContextSizeOption.d.ts +19 -0
  389. package/dist/gguf/insights/utils/resolveContextContextSizeOption.js +78 -0
  390. package/dist/gguf/insights/utils/resolveContextContextSizeOption.js.map +1 -0
  391. package/dist/gguf/insights/utils/resolveModelGpuLayersOption.d.ts +15 -0
  392. package/dist/gguf/insights/utils/resolveModelGpuLayersOption.js +183 -0
  393. package/dist/gguf/insights/utils/resolveModelGpuLayersOption.js.map +1 -0
  394. package/dist/gguf/insights/utils/scoreLevels.d.ts +5 -0
  395. package/dist/gguf/insights/utils/scoreLevels.js +16 -0
  396. package/dist/gguf/insights/utils/scoreLevels.js.map +1 -0
  397. package/dist/gguf/parser/GgufV2Parser.d.ts +20 -0
  398. package/dist/gguf/parser/GgufV2Parser.js +156 -0
  399. package/dist/gguf/parser/GgufV2Parser.js.map +1 -0
  400. package/dist/gguf/parser/GgufV3Parser.d.ts +3 -0
  401. package/dist/gguf/parser/GgufV3Parser.js +4 -0
  402. package/dist/gguf/parser/GgufV3Parser.js.map +1 -0
  403. package/dist/gguf/parser/parseGguf.d.ts +8 -0
  404. package/dist/gguf/parser/parseGguf.js +61 -0
  405. package/dist/gguf/parser/parseGguf.js.map +1 -0
  406. package/dist/gguf/readGgufFileInfo.d.ts +33 -0
  407. package/dist/gguf/readGgufFileInfo.js +66 -0
  408. package/dist/gguf/readGgufFileInfo.js.map +1 -0
  409. package/dist/gguf/types/GgufFileInfoTypes.d.ts +84 -0
  410. package/dist/gguf/types/GgufFileInfoTypes.js +18 -0
  411. package/dist/gguf/types/GgufFileInfoTypes.js.map +1 -0
  412. package/dist/gguf/types/GgufMetadataTypes.d.ts +356 -0
  413. package/dist/gguf/types/GgufMetadataTypes.js +99 -0
  414. package/dist/gguf/types/GgufMetadataTypes.js.map +1 -0
  415. package/dist/gguf/types/GgufTensorInfoTypes.d.ts +37 -0
  416. package/dist/gguf/types/GgufTensorInfoTypes.js +33 -0
  417. package/dist/gguf/types/GgufTensorInfoTypes.js.map +1 -0
  418. package/dist/gguf/utils/GgufReadOffset.d.ts +6 -0
  419. package/dist/gguf/utils/GgufReadOffset.js +18 -0
  420. package/dist/gguf/utils/GgufReadOffset.js.map +1 -0
  421. package/dist/gguf/utils/convertMetadataKeyValueRecordToNestedObject.d.ts +6 -0
  422. package/dist/gguf/utils/convertMetadataKeyValueRecordToNestedObject.js +74 -0
  423. package/dist/gguf/utils/convertMetadataKeyValueRecordToNestedObject.js.map +1 -0
  424. package/dist/gguf/utils/getGgufFileTypeName.d.ts +4 -0
  425. package/dist/gguf/utils/getGgufFileTypeName.js +13 -0
  426. package/dist/gguf/utils/getGgufFileTypeName.js.map +1 -0
  427. package/dist/gguf/utils/getGgufMetadataArchitectureData.d.ts +3 -0
  428. package/dist/gguf/utils/getGgufMetadataArchitectureData.js +4 -0
  429. package/dist/gguf/utils/getGgufMetadataArchitectureData.js.map +1 -0
  430. package/dist/gguf/utils/normalizeGgufDownloadUrl.d.ts +1 -0
  431. package/dist/gguf/utils/normalizeGgufDownloadUrl.js +16 -0
  432. package/dist/gguf/utils/normalizeGgufDownloadUrl.js.map +1 -0
  433. package/dist/gguf/utils/resolveBinarySplitGgufPartUrls.d.ts +2 -0
  434. package/dist/gguf/utils/resolveBinarySplitGgufPartUrls.js +39 -0
  435. package/dist/gguf/utils/resolveBinarySplitGgufPartUrls.js.map +1 -0
  436. package/dist/gguf/utils/resolveSplitGgufParts.d.ts +7 -0
  437. package/dist/gguf/utils/resolveSplitGgufParts.js +55 -0
  438. package/dist/gguf/utils/resolveSplitGgufParts.js.map +1 -0
  439. package/dist/index.d.ts +43 -18
  440. package/dist/index.js +38 -15
  441. package/dist/index.js.map +1 -1
  442. package/dist/state.d.ts +4 -0
  443. package/dist/state.js +14 -0
  444. package/dist/state.js.map +1 -1
  445. package/dist/types.d.ts +130 -5
  446. package/dist/types.js.map +1 -1
  447. package/dist/utils/DeepPartialObject.d.ts +3 -0
  448. package/dist/utils/DeepPartialObject.js +2 -0
  449. package/dist/utils/DeepPartialObject.js.map +1 -0
  450. package/dist/utils/DisposeGuard.d.ts +13 -0
  451. package/dist/utils/DisposeGuard.js +120 -0
  452. package/dist/utils/DisposeGuard.js.map +1 -0
  453. package/dist/utils/InsufficientMemoryError.d.ts +3 -0
  454. package/dist/utils/InsufficientMemoryError.js +6 -0
  455. package/dist/utils/InsufficientMemoryError.js.map +1 -0
  456. package/dist/utils/LlamaText.d.ts +70 -26
  457. package/dist/utils/LlamaText.js +472 -157
  458. package/dist/utils/LlamaText.js.map +1 -1
  459. package/dist/utils/LruCache.d.ts +12 -0
  460. package/dist/utils/LruCache.js +44 -0
  461. package/dist/utils/LruCache.js.map +1 -0
  462. package/dist/utils/ReplHistory.js +5 -1
  463. package/dist/utils/ReplHistory.js.map +1 -1
  464. package/dist/utils/StopGenerationDetector.d.ts +27 -8
  465. package/dist/utils/StopGenerationDetector.js +108 -22
  466. package/dist/utils/StopGenerationDetector.js.map +1 -1
  467. package/dist/utils/TokenStreamRegulator.d.ts +10 -4
  468. package/dist/utils/TokenStreamRegulator.js +102 -10
  469. package/dist/utils/TokenStreamRegulator.js.map +1 -1
  470. package/dist/utils/UnsupportedError.d.ts +2 -0
  471. package/dist/utils/UnsupportedError.js +7 -0
  472. package/dist/utils/UnsupportedError.js.map +1 -0
  473. package/dist/utils/appendUserMessageToChatHistory.js.map +1 -1
  474. package/dist/utils/clearTempFolder.js.map +1 -1
  475. package/dist/utils/cmake.js +38 -20
  476. package/dist/utils/cmake.js.map +1 -1
  477. package/dist/utils/createModelDownloader.d.ts +111 -0
  478. package/dist/utils/createModelDownloader.js +273 -0
  479. package/dist/utils/createModelDownloader.js.map +1 -0
  480. package/dist/utils/findBestOption.d.ts +4 -0
  481. package/dist/utils/findBestOption.js +15 -0
  482. package/dist/utils/findBestOption.js.map +1 -0
  483. package/dist/utils/findCharacterRemovalCountToFitChatHistoryInContext.d.ts +1 -0
  484. package/dist/utils/findCharacterRemovalCountToFitChatHistoryInContext.js +23 -12
  485. package/dist/utils/findCharacterRemovalCountToFitChatHistoryInContext.js.map +1 -1
  486. package/dist/utils/gbnfJson/GbnfGrammarGenerator.js.map +1 -1
  487. package/dist/utils/gbnfJson/getGbnfGrammarForGbnfJsonSchema.d.ts +5 -0
  488. package/dist/utils/gbnfJson/getGbnfGrammarForGbnfJsonSchema.js +11 -0
  489. package/dist/utils/gbnfJson/getGbnfGrammarForGbnfJsonSchema.js.map +1 -0
  490. package/dist/utils/gbnfJson/terminals/GbnfArray.d.ts +3 -1
  491. package/dist/utils/gbnfJson/terminals/GbnfArray.js +10 -5
  492. package/dist/utils/gbnfJson/terminals/GbnfArray.js.map +1 -1
  493. package/dist/utils/gbnfJson/terminals/GbnfBoolean.d.ts +1 -1
  494. package/dist/utils/gbnfJson/terminals/GbnfBoolean.js.map +1 -1
  495. package/dist/utils/gbnfJson/terminals/GbnfBooleanValue.js.map +1 -1
  496. package/dist/utils/gbnfJson/terminals/GbnfGrammar.js.map +1 -1
  497. package/dist/utils/gbnfJson/terminals/GbnfNull.d.ts +1 -1
  498. package/dist/utils/gbnfJson/terminals/GbnfNull.js.map +1 -1
  499. package/dist/utils/gbnfJson/terminals/GbnfNumber.d.ts +1 -1
  500. package/dist/utils/gbnfJson/terminals/GbnfNumber.js.map +1 -1
  501. package/dist/utils/gbnfJson/terminals/GbnfNumberValue.js.map +1 -1
  502. package/dist/utils/gbnfJson/terminals/GbnfObjectMap.d.ts +3 -1
  503. package/dist/utils/gbnfJson/terminals/GbnfObjectMap.js +9 -4
  504. package/dist/utils/gbnfJson/terminals/GbnfObjectMap.js.map +1 -1
  505. package/dist/utils/gbnfJson/terminals/GbnfOr.js.map +1 -1
  506. package/dist/utils/gbnfJson/terminals/GbnfRepetition.d.ts +9 -0
  507. package/dist/utils/gbnfJson/terminals/GbnfRepetition.js +37 -0
  508. package/dist/utils/gbnfJson/terminals/GbnfRepetition.js.map +1 -0
  509. package/dist/utils/gbnfJson/terminals/GbnfString.d.ts +1 -1
  510. package/dist/utils/gbnfJson/terminals/GbnfString.js +23 -5
  511. package/dist/utils/gbnfJson/terminals/GbnfString.js.map +1 -1
  512. package/dist/utils/gbnfJson/terminals/GbnfStringValue.js.map +1 -1
  513. package/dist/utils/gbnfJson/terminals/GbnfVerbatimText.js.map +1 -1
  514. package/dist/utils/gbnfJson/terminals/GbnfWhitespace.d.ts +7 -4
  515. package/dist/utils/gbnfJson/terminals/GbnfWhitespace.js +37 -9
  516. package/dist/utils/gbnfJson/terminals/GbnfWhitespace.js.map +1 -1
  517. package/dist/utils/gbnfJson/terminals/gbnfConsts.d.ts +5 -4
  518. package/dist/utils/gbnfJson/terminals/gbnfConsts.js +14 -3
  519. package/dist/utils/gbnfJson/terminals/gbnfConsts.js.map +1 -1
  520. package/dist/utils/gbnfJson/types.d.ts +3 -0
  521. package/dist/utils/gbnfJson/types.js.map +1 -1
  522. package/dist/utils/gbnfJson/utils/GbnfJsonScopeState.d.ts +10 -0
  523. package/dist/utils/gbnfJson/utils/GbnfJsonScopeState.js +15 -0
  524. package/dist/utils/gbnfJson/utils/GbnfJsonScopeState.js.map +1 -0
  525. package/dist/utils/gbnfJson/utils/getGbnfJsonTerminalForGbnfJsonSchema.d.ts +2 -1
  526. package/dist/utils/gbnfJson/utils/getGbnfJsonTerminalForGbnfJsonSchema.js +6 -5
  527. package/dist/utils/gbnfJson/utils/getGbnfJsonTerminalForGbnfJsonSchema.js.map +1 -1
  528. package/dist/utils/gbnfJson/utils/validateObjectAgainstGbnfSchema.js +2 -2
  529. package/dist/utils/gbnfJson/utils/validateObjectAgainstGbnfSchema.js.map +1 -1
  530. package/dist/utils/getBuildDefaults.d.ts +1 -2
  531. package/dist/utils/getBuildDefaults.js +2 -3
  532. package/dist/utils/getBuildDefaults.js.map +1 -1
  533. package/dist/utils/getConsoleLogPrefix.d.ts +1 -0
  534. package/dist/utils/getConsoleLogPrefix.js +10 -0
  535. package/dist/utils/getConsoleLogPrefix.js.map +1 -0
  536. package/dist/utils/getGrammarsFolder.d.ts +2 -1
  537. package/dist/utils/getGrammarsFolder.js +8 -7
  538. package/dist/utils/getGrammarsFolder.js.map +1 -1
  539. package/dist/utils/getModuleVersion.d.ts +1 -0
  540. package/dist/utils/getModuleVersion.js +13 -0
  541. package/dist/utils/getModuleVersion.js.map +1 -0
  542. package/dist/utils/getQueuedTokensBeforeStopTrigger.d.ts +6 -0
  543. package/dist/utils/getQueuedTokensBeforeStopTrigger.js +22 -0
  544. package/dist/utils/getQueuedTokensBeforeStopTrigger.js.map +1 -0
  545. package/dist/utils/getReadableContextSize.d.ts +1 -0
  546. package/dist/utils/getReadableContextSize.js +7 -0
  547. package/dist/utils/getReadableContextSize.js.map +1 -0
  548. package/dist/utils/getTypeScriptTypeStringForGbnfJsonSchema.js +15 -11
  549. package/dist/utils/getTypeScriptTypeStringForGbnfJsonSchema.js.map +1 -1
  550. package/dist/utils/gitReleaseBundles.js +73 -5
  551. package/dist/utils/gitReleaseBundles.js.map +1 -1
  552. package/dist/utils/hashString.d.ts +1 -0
  553. package/dist/utils/hashString.js +8 -0
  554. package/dist/utils/hashString.js.map +1 -0
  555. package/dist/utils/isLockfileActive.d.ts +4 -0
  556. package/dist/utils/isLockfileActive.js +12 -0
  557. package/dist/utils/isLockfileActive.js.map +1 -0
  558. package/dist/utils/isToken.d.ts +2 -0
  559. package/dist/utils/isToken.js +4 -0
  560. package/dist/utils/isToken.js.map +1 -0
  561. package/dist/utils/isUrl.d.ts +1 -0
  562. package/dist/utils/isUrl.js +15 -0
  563. package/dist/utils/isUrl.js.map +1 -0
  564. package/dist/utils/mergeUnionTypes.d.ts +10 -0
  565. package/dist/utils/mergeUnionTypes.js +2 -0
  566. package/dist/utils/mergeUnionTypes.js.map +1 -0
  567. package/dist/utils/parseModelFileName.d.ts +1 -0
  568. package/dist/utils/parseModelFileName.js +6 -1
  569. package/dist/utils/parseModelFileName.js.map +1 -1
  570. package/dist/utils/parseTextTemplate.d.ts +66 -0
  571. package/dist/utils/parseTextTemplate.js +116 -0
  572. package/dist/utils/parseTextTemplate.js.map +1 -0
  573. package/dist/utils/prettyPrintObject.d.ts +10 -0
  574. package/dist/utils/prettyPrintObject.js +84 -0
  575. package/dist/utils/prettyPrintObject.js.map +1 -0
  576. package/dist/utils/pushAll.d.ts +6 -0
  577. package/dist/utils/pushAll.js +11 -0
  578. package/dist/utils/pushAll.js.map +1 -0
  579. package/dist/utils/removeNullFields.d.ts +2 -1
  580. package/dist/utils/removeNullFields.js +8 -0
  581. package/dist/utils/removeNullFields.js.map +1 -1
  582. package/dist/utils/resolveGithubRelease.d.ts +2 -0
  583. package/dist/utils/resolveGithubRelease.js +36 -0
  584. package/dist/utils/resolveGithubRelease.js.map +1 -0
  585. package/dist/utils/resolveLastTokens.d.ts +2 -0
  586. package/dist/utils/resolveLastTokens.js +12 -0
  587. package/dist/utils/resolveLastTokens.js.map +1 -0
  588. package/dist/utils/runtime.d.ts +4 -0
  589. package/dist/utils/runtime.js +8 -0
  590. package/dist/utils/runtime.js.map +1 -0
  591. package/dist/utils/safeEventCallback.d.ts +6 -0
  592. package/dist/utils/safeEventCallback.js +29 -0
  593. package/dist/utils/safeEventCallback.js.map +1 -0
  594. package/dist/utils/spawnCommand.d.ts +11 -1
  595. package/dist/utils/spawnCommand.js +56 -6
  596. package/dist/utils/spawnCommand.js.map +1 -1
  597. package/dist/utils/tokenizeInput.d.ts +3 -0
  598. package/dist/utils/tokenizeInput.js +12 -0
  599. package/dist/utils/tokenizeInput.js.map +1 -0
  600. package/dist/utils/transformPromisable.d.ts +40 -0
  601. package/dist/utils/transformPromisable.js +64 -0
  602. package/dist/utils/transformPromisable.js.map +1 -0
  603. package/dist/utils/truncateTextAndRoundToWords.d.ts +2 -0
  604. package/dist/utils/truncateTextAndRoundToWords.js +30 -0
  605. package/dist/utils/truncateTextAndRoundToWords.js.map +1 -1
  606. package/dist/utils/utilTypes.d.ts +3 -0
  607. package/dist/utils/utilTypes.js +2 -0
  608. package/dist/utils/utilTypes.js.map +1 -0
  609. package/dist/utils/waitForLockfileRelease.d.ts +5 -0
  610. package/dist/utils/waitForLockfileRelease.js +20 -0
  611. package/dist/utils/waitForLockfileRelease.js.map +1 -0
  612. package/dist/utils/withLockfile.d.ts +7 -0
  613. package/dist/utils/withLockfile.js +44 -0
  614. package/dist/utils/withLockfile.js.map +1 -0
  615. package/dist/utils/withOra.d.ts +2 -0
  616. package/dist/utils/withOra.js +22 -6
  617. package/dist/utils/withOra.js.map +1 -1
  618. package/dist/utils/withProgressLog.d.ts +23 -0
  619. package/dist/utils/withProgressLog.js +211 -0
  620. package/dist/utils/withProgressLog.js.map +1 -0
  621. package/dist/utils/withStatusLogs.d.ts +2 -1
  622. package/dist/utils/withStatusLogs.js +12 -9
  623. package/dist/utils/withStatusLogs.js.map +1 -1
  624. package/dist/utils/wrapAbortSignal.d.ts +2 -0
  625. package/dist/utils/wrapAbortSignal.js +9 -0
  626. package/dist/utils/wrapAbortSignal.js.map +1 -0
  627. package/llama/.clang-format +1 -2
  628. package/llama/CMakeLists.txt +126 -5
  629. package/llama/addon/AddonContext.cpp +762 -0
  630. package/llama/addon/AddonContext.h +53 -0
  631. package/llama/addon/AddonGrammar.cpp +42 -0
  632. package/llama/addon/AddonGrammar.h +18 -0
  633. package/llama/addon/AddonGrammarEvaluationState.cpp +26 -0
  634. package/llama/addon/AddonGrammarEvaluationState.h +15 -0
  635. package/llama/addon/AddonModel.cpp +671 -0
  636. package/llama/addon/AddonModel.h +61 -0
  637. package/llama/addon/AddonModelData.cpp +25 -0
  638. package/llama/addon/AddonModelData.h +15 -0
  639. package/llama/addon/AddonModelLora.cpp +105 -0
  640. package/llama/addon/AddonModelLora.h +28 -0
  641. package/llama/addon/addon.cpp +216 -0
  642. package/llama/addon/addonGlobals.cpp +22 -0
  643. package/llama/addon/addonGlobals.h +12 -0
  644. package/llama/addon/globals/addonLog.cpp +135 -0
  645. package/llama/addon/globals/addonLog.h +21 -0
  646. package/llama/addon/globals/addonProgress.cpp +15 -0
  647. package/llama/addon/globals/addonProgress.h +15 -0
  648. package/llama/addon/globals/getGpuInfo.cpp +108 -0
  649. package/llama/addon/globals/getGpuInfo.h +6 -0
  650. package/llama/binariesGithubRelease.json +1 -1
  651. package/llama/gitRelease.bundle +0 -0
  652. package/llama/gpuInfo/cuda-gpu-info.cu +120 -0
  653. package/llama/gpuInfo/cuda-gpu-info.h +10 -0
  654. package/llama/gpuInfo/metal-gpu-info.h +8 -0
  655. package/llama/gpuInfo/metal-gpu-info.mm +30 -0
  656. package/llama/gpuInfo/vulkan-gpu-info.cpp +83 -0
  657. package/llama/gpuInfo/vulkan-gpu-info.h +9 -0
  658. package/llama/grammars/README.md +297 -6
  659. package/llama/grammars/json.gbnf +4 -4
  660. package/llama/grammars/json_arr.gbnf +4 -4
  661. package/llama/llama.cpp.info.json +4 -0
  662. package/llama/toolchains/win32.host-x64.target-arm64.cmake +41 -0
  663. package/package.json +85 -54
  664. package/templates/packed/electron-typescript-react.json +1 -0
  665. package/templates/packed/node-typescript.json +1 -0
  666. package/dist/AbortError.d.ts +0 -2
  667. package/dist/AbortError.js +0 -7
  668. package/dist/AbortError.js.map +0 -1
  669. package/dist/chatWrappers/LlamaChatWrapper.d.ts +0 -13
  670. package/dist/chatWrappers/LlamaChatWrapper.js.map +0 -1
  671. package/dist/chatWrappers/resolveChatWrapperBasedOnModel.d.ts +0 -13
  672. package/dist/chatWrappers/resolveChatWrapperBasedOnModel.js +0 -57
  673. package/dist/chatWrappers/resolveChatWrapperBasedOnModel.js.map +0 -1
  674. package/dist/llamaEvaluator/LlamaBins.d.ts +0 -18
  675. package/dist/llamaEvaluator/LlamaBins.js +0 -5
  676. package/dist/llamaEvaluator/LlamaBins.js.map +0 -1
  677. package/dist/llamaEvaluator/LlamaChat/LlamaChat.d.ts +0 -175
  678. package/dist/llamaEvaluator/LlamaChat/LlamaChat.js +0 -704
  679. package/dist/llamaEvaluator/LlamaChat/LlamaChat.js.map +0 -1
  680. package/dist/llamaEvaluator/LlamaChat/utils/FunctionCallGrammar.d.ts +0 -21
  681. package/dist/llamaEvaluator/LlamaChat/utils/FunctionCallGrammar.js +0 -120
  682. package/dist/llamaEvaluator/LlamaChat/utils/FunctionCallGrammar.js.map +0 -1
  683. package/dist/llamaEvaluator/LlamaChat/utils/contextShiftStrategies/eraseFirstResponseAndKeepFirstSystemChatContextShiftStrategy.js.map +0 -1
  684. package/dist/llamaEvaluator/LlamaChatSession/LlamaChatSession.d.ts +0 -146
  685. package/dist/llamaEvaluator/LlamaChatSession/LlamaChatSession.js +0 -211
  686. package/dist/llamaEvaluator/LlamaChatSession/LlamaChatSession.js.map +0 -1
  687. package/dist/llamaEvaluator/LlamaChatSession/utils/defineChatSessionFunction.js.map +0 -1
  688. package/dist/llamaEvaluator/LlamaContext/LlamaContext.js.map +0 -1
  689. package/dist/llamaEvaluator/LlamaContext/types.d.ts +0 -82
  690. package/dist/llamaEvaluator/LlamaContext/types.js.map +0 -1
  691. package/dist/llamaEvaluator/LlamaContext/utils/batchItemsPrioritizingStrategies/firstInFirstOutStrategy.js.map +0 -1
  692. package/dist/llamaEvaluator/LlamaContext/utils/batchItemsPrioritizingStrategies/maximumParallelismStrategy.js.map +0 -1
  693. package/dist/llamaEvaluator/LlamaContext/utils/resolveBatchItemsPrioritizingStrategy.d.ts +0 -2
  694. package/dist/llamaEvaluator/LlamaContext/utils/resolveBatchItemsPrioritizingStrategy.js.map +0 -1
  695. package/dist/llamaEvaluator/LlamaEmbeddingContext.d.ts +0 -35
  696. package/dist/llamaEvaluator/LlamaEmbeddingContext.js +0 -73
  697. package/dist/llamaEvaluator/LlamaEmbeddingContext.js.map +0 -1
  698. package/dist/llamaEvaluator/LlamaGrammar.d.ts +0 -28
  699. package/dist/llamaEvaluator/LlamaGrammar.js.map +0 -1
  700. package/dist/llamaEvaluator/LlamaGrammarEvaluationState.js.map +0 -1
  701. package/dist/llamaEvaluator/LlamaJsonSchemaGrammar.js.map +0 -1
  702. package/dist/llamaEvaluator/LlamaModel.d.ts +0 -119
  703. package/dist/llamaEvaluator/LlamaModel.js +0 -322
  704. package/dist/llamaEvaluator/LlamaModel.js.map +0 -1
  705. package/dist/utils/binariesGithubRelease.js.map +0 -1
  706. package/dist/utils/clearLlamaBuild.d.ts +0 -1
  707. package/dist/utils/clearLlamaBuild.js +0 -12
  708. package/dist/utils/clearLlamaBuild.js.map +0 -1
  709. package/dist/utils/cloneLlamaCppRepo.d.ts +0 -2
  710. package/dist/utils/cloneLlamaCppRepo.js +0 -102
  711. package/dist/utils/cloneLlamaCppRepo.js.map +0 -1
  712. package/dist/utils/compileLLamaCpp.d.ts +0 -8
  713. package/dist/utils/compileLLamaCpp.js +0 -132
  714. package/dist/utils/compileLLamaCpp.js.map +0 -1
  715. package/dist/utils/getBin.js +0 -78
  716. package/dist/utils/getBin.js.map +0 -1
  717. package/dist/utils/getGbnfGrammarForGbnfJsonSchema.d.ts +0 -2
  718. package/dist/utils/getGbnfGrammarForGbnfJsonSchema.js +0 -9
  719. package/dist/utils/getGbnfGrammarForGbnfJsonSchema.js.map +0 -1
  720. package/dist/utils/getReleaseInfo.d.ts +0 -7
  721. package/dist/utils/getReleaseInfo.js +0 -30
  722. package/dist/utils/getReleaseInfo.js.map +0 -1
  723. package/dist/utils/parseModelTypeDescription.d.ts +0 -6
  724. package/dist/utils/parseModelTypeDescription.js +0 -9
  725. package/dist/utils/parseModelTypeDescription.js.map +0 -1
  726. package/dist/utils/resolveChatWrapper.d.ts +0 -4
  727. package/dist/utils/resolveChatWrapper.js +0 -16
  728. package/dist/utils/resolveChatWrapper.js.map +0 -1
  729. package/dist/utils/usedBinFlag.d.ts +0 -6
  730. package/dist/utils/usedBinFlag.js +0 -15
  731. package/dist/utils/usedBinFlag.js.map +0 -1
  732. package/llama/addon.cpp +0 -814
  733. package/llama/usedBin.json +0 -3
  734. package/llamaBins/linux-arm64/llama-addon.node +0 -0
  735. package/llamaBins/linux-armv7l/llama-addon.node +0 -0
  736. package/llamaBins/linux-x64/llama-addon.node +0 -0
  737. package/llamaBins/mac-arm64/llama-addon.node +0 -0
  738. package/llamaBins/mac-x64/llama-addon.node +0 -0
  739. package/llamaBins/win-x64/llama-addon.exp +0 -0
  740. package/llamaBins/win-x64/llama-addon.lib +0 -0
  741. package/llamaBins/win-x64/llama-addon.node +0 -0
  742. /package/dist/{utils → bindings/utils}/binariesGithubRelease.d.ts +0 -0
  743. /package/dist/{llamaEvaluator → evaluator}/LlamaChat/utils/contextShiftStrategies/eraseFirstResponseAndKeepFirstSystemChatContextShiftStrategy.d.ts +0 -0
  744. /package/dist/{llamaEvaluator → evaluator}/LlamaContext/types.js +0 -0
  745. /package/dist/{llamaEvaluator/LlamaContext/utils/batchItemsPrioritizingStrategies → evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies}/firstInFirstOutStrategy.d.ts +0 -0
  746. /package/dist/{llamaEvaluator/LlamaContext/utils/batchItemsPrioritizingStrategies → evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies}/firstInFirstOutStrategy.js +0 -0
  747. /package/dist/{llamaEvaluator/LlamaContext/utils/batchItemsPrioritizingStrategies → evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies}/maximumParallelismStrategy.d.ts +0 -0
  748. /package/dist/{llamaEvaluator/LlamaContext/utils/batchItemsPrioritizingStrategies → evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies}/maximumParallelismStrategy.js +0 -0
  749. /package/dist/{llamaEvaluator → evaluator}/LlamaGrammarEvaluationState.d.ts +0 -0
@@ -0,0 +1 @@
1
+ {"version":3,"file":"GgufInsights.js","sourceRoot":"","sources":["../../../src/gguf/insights/GgufInsights.ts"],"names":[],"mappings":"AACA,OAAO,EAAC,sBAAsB,EAAC,MAAM,gDAAgD,CAAC;AACtF,OAAO,EAAC,0BAA0B,EAAE,0BAA0B,EAAC,MAAM,8CAA8C,CAAC;AAGpH,OAAO,EAAC,oBAAoB,EAAC,MAAM,+BAA+B,CAAC;AACnE,OAAO,EAAC,eAAe,EAAC,MAAM,oCAAoC,CAAC;AACnE,OAAO,EAAC,iCAAiC,EAAC,MAAM,wCAAwC,CAAC;AAOzF,MAAM,OAAO,YAAY;IACrB,gBAAgB,CAAiB,MAAM,CAAQ;IAC/C,gBAAgB,CAAkB,UAAU,CAAS;IACrD,gBAAgB,CAAS,YAAY,GAAkB,IAAI,CAAC;IAC5D,gBAAgB,CAAkB,aAAa,CAAe;IAC9D,gBAAgB,CAAkB,sBAAsB,CAAoC;IAE5F,YAAoB,YAA0B,EAAE,KAAY;QACxD,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC;QACpB,IAAI,CAAC,aAAa,GAAG,YAAY,CAAC;QAElC,IAAI,CAAC,UAAU,GAAG,oBAAoB,CAAC,YAAY,CAAC,cAAc,IAAI,EAAE,EAAE,KAAK,CAAC,CAAC;QACjF,IAAI,CAAC,sBAAsB,GAAG,iCAAiC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;IAClF,CAAC;IAED;;;;OAIG;IACI,WAAW,CAAC,aAAsB;QACrC,MAAM,QAAQ,GAAa,EAAE,CAAC;QAC9B,MAAM,iBAAiB,GAAG,CAAC,aAAa,IAAI,IAAI,IAAI,aAAa,KAAK,EAAE,CAAC;YACrE,CAAC,CAAC,MAAM,eAAe,CAAC,aAAa,CAAC,IAAI;YAC1C,CAAC,CAAC,EAAE,CAAC;QAET,IAAI,IAAI,CAAC,aAAa,EAAE,QAAQ,EAAE,SAAS,EAAE,IAAI,EAAE,KAAK,KAAK,MAAM;YAC/D,IAAI,CAAC,aAAa,EAAE,QAAQ,EAAE,SAAS,EAAE,IAAI,EAAE,KAAK,IAAI,IAAI,EAC9D,CAAC;YACC,kHAAkH;YAClH,QAAQ,CAAC,IAAI,CACT,kBAAkB,iBAAiB,6CAA6C;gBAChF,iFAAiF;gBACjF,mEAAmE,CACtE,CAAC;QACN,CAAC;QAED,OAAO,QAAQ,CAAC;IACpB,CAAC;IAED,IAAW,YAAY;QACnB,OAAO,IAAI,CAAC,aAAa,CAAC;IAC9B,CAAC;IAED,IAAW,qBAAqB;QAC5B,OAAO,IAAI,CAAC,sBAAsB,CAAC;IACvC,CAAC;IAED,gDAAgD;IAChD,IAAW,gBAAgB;QACvB,OAAO,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,cAAc,CAAC;IAClE,CAAC;IAED,4DAA4D;IAC5D,IAAW,mBAAmB;QAC1B,OAAO,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,gBAAgB,CAAC;IACpE,CAAC;IAED,IAAW,WAAW;QAClB,IAAI,IAAI,CAAC,YAAY,IAAI,IAAI;YACzB,OAAO,IAAI,CAAC,YAAY,CAAC;QAE7B,MAAM,YAAY,GAAG,CAAC,CAAC;QACvB,IAAI,CAAC,YAAY,GAAG,IAAI,CAAC,cAAc,EAAE,GAAG,YAAY,CAAC;QAEzD,OAAO,IAAI,CAAC,YAAY,CAAC;IAC7B,CAAC;IAED,IAAW,SAAS;QAChB,OAAO,IAAI,CAAC,UAAU,CAAC;IAC3B,CAAC;IAED,IAAW,uBAAuB;QAC9B,wDAAwD;QAExD,IAAI,IAAI,CAAC,aAAa,CAAC,QAAQ,EAAE,OAAO,EAAE,YAAY,KAAK,oBAAoB,CAAC,IAAI;YAChF,OAAO,KAAK,CAAC;aACZ,IAAI,IAAI,CAAC,aAAa,CAAC,QAAQ,EAAE,OAAO,EAAE,YAAY,KAAK,oBAAoB,CAAC,MAAM;YACvF,OAAO,KAAK,CAAC;aACZ,CAAC;YACF,MAAM,KAAK,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,EAAE,SAAS,EAAE,UAAU,IAAI,CAAC,CAAC;YAClF,MAAM,KAAK,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,EAAE,gBAAgB,IAAI,CAAC,CAAC;YAC7E,MAAM,UAAU,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,EAAE,SAAS,EAAE,UAAU,IAAI,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC;YAC1H,MAAM,UAAU,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,EAAE,SAAS,EAAE,YAAY,IAAI,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC;YAE1H,IAAI,UAAU,KAAK,UAAU;gBACzB,OAAO,KAAK,CAAC;QACrB,CAAC;QAED,OAAO,IAAI,CAAC;IAChB,CAAC;IAEM,iCAAiC,CAAC,EAAC,SAAS,EAAsB;QACrE,MAAM,EAAC,GAAG,EAAE,GAAG,EAAC,GAAG,IAAI,CAAC,uBAAuB,CAAC,SAAS,CAAC,CAAC;QAE3D,OAAO;YACH,MAAM,EAAE,oBAAoB,CAAC,GAAG,EAAE,IAAI,CAAC,MAAM,CAAC;YAC9C,OAAO,EAAE,oBAAoB,CAAC,GAAG,EAAE,IAAI,CAAC,MAAM,CAAC;SAClD,CAAC;IACN,CAAC;IAED;;;;OAIG;IACI,mCAAmC,CAAC,EACvC,WAAW,EAAE,cAAc,EAAE,SAAS,EAAE,SAAS,EAAE,kBAAkB,GAAG,KAAK,EAAE,oBAAoB,GAAG,IAAI,EAAE,cAAc,GAAG,KAAK,EAIrI;QACG,IAAI,SAAS,IAAI,IAAI;YAAE,SAAS,GAAG,0BAA0B,EAAE,CAAC;QAChE,IAAI,SAAS,IAAI,IAAI;YAAE,SAAS,GAAG,0BAA0B,CAAC,EAAC,WAAW,EAAE,SAAS,EAAC,CAAC,CAAC;QAExF,MAAM,iBAAiB,GAAG,WAAW,GAAG,SAAS,CAAC;QAElD,MAAM,WAAW,GAAG,IAAI,CAAC,WAAW,CAAC;QACrC,MAAM,cAAc,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,cAAc,IAAI,WAAW,EAAE,WAAW,CAAC,CAAC,CAAC;QACzF,MAAM,cAAc,GAAG,WAAW,GAAG,cAAc,CAAC;QACpD,MAAM,OAAO,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC;QAExD,MAAM,cAAc,GAAG,OAAO,CAAC,UAAU,IAAI,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,SAAS,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,IAAI,CAAC,CAAC;QAC9G,MAAM,UAAU,GAAG,cAAc,GAAG,SAAS,CAAC;QAC9C,MAAM,SAAS,GAAG,kBAAkB;YAChC,CAAC,CAAC,CAAC,OAAO,CAAC,gBAAgB,IAAI,CAAC,CAAC,GAAG,SAAS;YAC7C,CAAC,CAAC,CAAC,CAAC;QAER,MAAM,UAAU,GAAG,CAAC,CAAC,CAAC,iBAAiB;QACvC,MAAM,UAAU,GAAG,CAAC,CAAC,CAAC,gBAAgB;QACtC,MAAM,YAAY,GAAG,CAAC,CAAC,CAAC,mBAAmB;QAC3C,MAAM,WAAW,GAAG,CAAC,CAAC,CAAC,kBAAkB;QAEzC,gDAAgD;QAChD,MAAM,QAAQ,GAAG,UAAU,CAAC;QAC5B,MAAM,IAAI,GAAG,EAAE,GAAG,IAAI,CAAC,CAAC,sBAAsB;QAC9C,MAAM,SAAS,GAAG,UAAU,CAAC;QAC7B,MAAM,WAAW,GAAG,SAAS,GAAG,WAAW,CAAC;QAC5C,MAAM,WAAW,GAAG,UAAU,CAAC;QAC/B,MAAM,OAAO,GAAG,UAAU,GAAG,UAAU,CAAC;QACxC,MAAM,cAAc,GAAG,UAAU,CAAC;QAClC,MAAM,UAAU,GAAG,SAAS,GAAG,UAAU,CAAC;QAC1C,MAAM,UAAU,GAAG,UAAU,CAAC;QAC9B,MAAM,OAAO,GAAG,YAAY,CAAC;QAC7B,MAAM,OAAO,GAAG,YAAY,CAAC;QAC7B,MAAM,OAAO,GAAG,YAAY,CAAC;QAC7B,MAAM,GAAG,GAAG,CAAC,GAAG,WAAW,GAAG,cAAc,GAAG,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,kBAAkB,CAAC;QACtF,MAAM,OAAO,GAAG,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,YAAY,GAAG,UAAU,GAAG,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,cAAc,CAAC;QACnG,MAAM,YAAY,GAAG,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,YAAY,KAAK,oBAAoB,CAAC,KAAK;YACjG,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,SAAS,CAAC;YACxB,CAAC,CAAC,iBAAiB,CAAC;QACxB,MAAM,QAAQ,GAAG,YAAY,GAAG,OAAO,CAAC;QAExC,MAAM,cAAc,GAAG,CACnB,QAAQ;YACR,IAAI;YACJ,SAAS;YACT,WAAW;YACX,WAAW;YACX,OAAO;YACP,cAAc;YACd,UAAU;YACV,UAAU;YACV,OAAO;YACP,OAAO;YACP,OAAO;YACP,GAAG;YACH,QAAQ,CACX,CAAC;QAEF,mHAAmH;QACnH,mHAAmH;QACnH,MAAM,2BAA2B,GAAG,GAAG,EAAE;YACrC,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;YAC/B,MAAM,UAAU,GAAG,IAAI,CAAC,aAAa,CAAC,cAAc,IAAI,EAAE,CAAC;YAE3D,IAAI,4BAA4B,GAAG,CAAC,CAAC;YAErC,IAAI,SAAS,IAAI,IAAI;gBACjB,OAAO,CAAC,CAAC;YAEb,IAAI,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,YAAY,KAAK,oBAAoB,CAAC,KAAK,EAAE,CAAC;gBACnF,MAAM,WAAW,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,YAAY,IAAI,CAAC,CAAC;gBAC9E,MAAM,SAAS,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,SAAS,EAAE,UAAU,IAAI,CAAC,CAAC;gBACrF,MAAM,eAAe,GAAG,OAAO,CAAC,gBAAgB,IAAI,CAAC,CAAC;gBAEtD,IAAI,WAAW,GAAG,CAAC,EAAE,CAAC;oBAClB,MAAM,gBAAgB,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,iBAAiB,IAAI,CAAC,CAAC;oBAExF,OAAO,WAAW,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,gBAAgB,GAAG,CAAC,CAAC,GAAG,eAAe,CAAC,GAAG,CAAC,iBAAiB,GAAG,SAAS,CAAC,CAAC,CAAC;gBACpH,CAAC;gBAED,OAAO,WAAW,GAAG,SAAS,GAAG,CAAC,eAAe,GAAG,CAAC,iBAAiB,GAAG,SAAS,CAAC,CAAC,CAAC;YACzF,CAAC;iBAAM,IAAI,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,YAAY,KAAK,oBAAoB,CAAC,KAAK,EAAE,CAAC;gBAC1F,IAAI,cAAc,KAAK,IAAI,CAAC,WAAW,EAAE,CAAC;oBACtC,4BAA4B,IAAI,CAAC,IAAI,GAAG,GAAG,CAAC,GAAG,CAC3C,IAAI,CAAC,gBAAgB,IAAI,IAAI;wBACzB,CAAC,CAAC,CAAC;wBACH,CAAC,CAAC,iBAAiB,GAAG,IAAI,CAAC,gBAAgB,CAClD,CAAC;gBACN,CAAC;qBAAM,CAAC;oBACJ,4BAA4B,IAAI,CAAC,IAAI,GAAG,GAAG,CAAC,GAAG,CAC3C,CAAC,IAAI,GAAG,EAAE,CAAC,GAAG,CACV,IAAI,CAAC,gBAAgB,IAAI,IAAI;wBACzB,CAAC,CAAC,CAAC;wBACH,CAAC,CAAC,iBAAiB,GAAG,IAAI,CAAC,gBAAgB,CAClD,CACJ,CAAC;gBACN,CAAC;YACL,CAAC;iBAAM,IAAI,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,YAAY,KAAK,oBAAoB,CAAC,KAAK,EAAE,CAAC;gBAC1F,uFAAuF;gBACvF,sEAAsE;gBAEtE,IAAI,cAAc,KAAK,IAAI,CAAC,WAAW,EAAE,CAAC;oBACtC,4BAA4B,IAAI,CAAC,IAAI,GAAG,EAAE,CAAC,GAAG,CAC1C,CAAC,IAAI,GAAG,GAAG,CAAC,GAAG,CACX,IAAI,CAAC,gBAAgB,IAAI,IAAI;wBACzB,CAAC,CAAC,CAAC;wBACH,CAAC,CAAC,iBAAiB,GAAG,IAAI,CAAC,gBAAgB,CAClD,CACJ,CAAC;gBACN,CAAC;qBAAM,CAAC;oBACJ,4BAA4B,IAAI,CAAC,CAAC,IAAI,GAAG,GAAG,CAAC,GAAG,CAC5C,CAAC,IAAI,GAAG,GAAG,CAAC,GAAG,CACX,IAAI,CAAC,gBAAgB,IAAI,IAAI;wBACzB,CAAC,CAAC,CAAC;wBACH,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,GAAG,CAAC,iBAAiB,GAAG,IAAI,CAAC,gBAAgB,CAAC,CAAC,CAAC,CACvE,CACJ,CAAC;gBACN,CAAC;YACL,CAAC;iBAAM,IAAI,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,YAAY,KAAK,oBAAoB,CAAC,QAAQ,EAAE,CAAC;gBAC7F,MAAM,SAAS,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,SAAS,EAAE,UAAU,IAAI,CAAC,CAAC;gBAErF,OAAO,CAAC,WAAW,GAAG,SAAS,GAAG,iBAAiB,GAAG,SAAS,CAAC,GAAG,CAAC,EAAE,GAAG,IAAI,CAAC,CAAC;gBAE/E,6CAA6C;gBAC7C,uDAAuD;gBACvD,2BAA2B;gBAC3B,4CAA4C;gBAC5C,sBAAsB;gBACtB,8DAA8D;gBAC9D,YAAY;gBACZ,SAAS;gBACT,WAAW;gBACX,uDAAuD;gBACvD,2BAA2B;gBAC3B,4CAA4C;gBAC5C,sBAAsB;gBACtB,8DAA8D;gBAC9D,YAAY;gBACZ,SAAS;gBACT,IAAI;YACR,CAAC;YAED,MAAM,aAAa,GAAG,UAAU,CAAC,MAAM,KAAK,CAAC;gBACzC,CAAC,CAAC,IAAI,CAAC,WAAW,GAAG,CACjB,CACI,CAAC,OAAO,CAAC,gBAAgB,IAAI,CAAC,CAAC;oBAC/B,CAAC,OAAO,CAAC,mBAAmB,IAAI,CAAC,CAAC,CACrC,GAAG,CAAC,CACR;gBACD,CAAC,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,MAAM,EAAE,EAAE;oBAChC,OAAO,GAAG,GAAG,MAAM,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,GAAW,EAAE,GAAG,EAAE,EAAE,CAAC,GAAG,GAAG,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC;gBACtF,CAAC,EAAE,CAAC,CAAC,CAAC;YAEV,IAAI,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,YAAY,KAAK,oBAAoB,CAAC,IAAI,EAAE,CAAC;gBAClF,+DAA+D;gBAC/D,OAAO,CAAC,aAAa,GAAG,GAAG,GAAG,CAAC,iBAAiB,GAAG,IAAI,CAAC,CAAC,GAAG,4BAA4B,CAAC;YAC7F,CAAC;YAED,+DAA+D;YAC/D,OAAO,CAAC,aAAa,GAAG,MAAM,GAAG,CAAC,iBAAiB,GAAG,IAAI,CAAC,CAAC,GAAG,4BAA4B,CAAC;QAChG,CAAC,CAAC;QAEF,MAAM,mBAAmB,GAAG,CAAC,cAAc,IAAI,CAAC,oBAAoB,CAAC;YACjE,CAAC,CAAC,CAAC;YACH,CAAC,CAAC,2BAA2B,EAAE,CAAC;QAEpC,MAAM,QAAQ,GAAG,cAAc,KAAK,CAAC,CAAC;QAEtC,MAAM,MAAM,GAAG,CACX,CAAC,QAAQ;YACL,CAAC,CAAC,CAAC,cAAc,GAAG,mBAAmB,CAAC;YACxC,CAAC,CAAC,CAAC,CACV;YACG,IAAI,CAAC,4BAA4B,CAAC,iBAAiB,EAAE,cAAc,CAAC,CAAC;QACzE,MAAM,OAAO,GAAG,QAAQ;YACpB,CAAC,CAAC,CACE,cAAc;gBACd,mBAAmB;gBACnB,IAAI,CAAC,4BAA4B,CAC7B,iBAAiB,EACjB,cAAc,GAAG,WAAW;oBACxB,CAAC,CAAC,CAAC,cAAc,GAAG,CAAC,CAAC;oBACtB,CAAC,CAAC,cAAc,CACvB,CACJ;YACD,CAAC,CAAC,CAAC,CAAC;QAER,OAAO;YACH,MAAM;YACN,OAAO;SACV,CAAC;IACN,CAAC;IAED;;;OAGG;IACI,uBAAuB,CAAC,SAAiB;QAI5C,MAAM,UAAU,GAAG,IAAI,CAAC,aAAa,CAAC,cAAc,IAAI,EAAE,CAAC;QAE3D,IAAI,SAAS,KAAK,CAAC,EAAE,CAAC;YAClB,OAAO;gBACH,GAAG,EAAE,UAAU;gBACf,GAAG,EAAE,EAAE;aACV,CAAC;QACN,CAAC;QAED,MAAM,UAAU,GAAG,IAAI,CAAC,cAAc,EAAE,CAAC;QACzC,MAAM,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,UAAU,GAAG,SAAS,CAAC,CAAC;QAE1D,MAAM,UAAU,GAAqB,EAAE,CAAC;QACxC,MAAM,UAAU,GAAqB,EAAE,CAAC;QAExC,KAAK,MAAM,gBAAgB,IAAI,UAAU,EAAE,CAAC;YACxC,MAAM,EAAC,WAAW,EAAC,GAAG,eAAe,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC;YAE7D,IAAI,SAAS,KAAK,IAAI,CAAC,WAAW,EAAE,CAAC;gBACjC,MAAM,YAAY,GAAG,IAAI,CAAC,aAAa,CAAC,QAAQ,EAAE,OAAO,EAAE,YAAY,CAAC;gBAExE,IAAI,YAAY,KAAK,oBAAoB,CAAC,KAAK,IAAI,YAAY,KAAK,oBAAoB,CAAC,KAAK,EAAE,CAAC;oBAC7F,IAAI,WAAW,IAAI,IAAI,IAAI,WAAW,IAAI,aAAa;wBACnD,UAAU,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC;;wBAElC,UAAU,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC;oBAEtC,SAAS;gBACb,CAAC;YACL,CAAC;YAED,IAAI,WAAW,IAAI,IAAI,IAAI,WAAW,IAAI,aAAa;gBACnD,UAAU,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC;;gBAElC,UAAU,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC;QAC1C,CAAC;QAED,OAAO;YACH,GAAG,EAAE,UAAU;YACf,GAAG,EAAE,UAAU;SAClB,CAAC;IACN,CAAC;IAED,gBAAgB;IACT,sCAAsC;QACzC,MAAM,YAAY,GAAG,IAAI,GAAG,EAAU,CAAC;QAEvC,KAAK,MAAM,gBAAgB,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,cAAc,IAAI,EAAE,CAAC,EAAE,CAAC;YACvE,MAAM,EAAC,WAAW,EAAC,GAAG,eAAe,CAAC,gBAAgB,CAAC,IAAI,CAAC,CAAC;YAE7D,IAAI,WAAW,IAAI,IAAI;gBACnB,YAAY,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;QACtC,CAAC;QAED,OAAO,YAAY,CAAC,IAAI,CAAC;IAC7B,CAAC;IAED,gBAAgB;IACT,cAAc;QACjB,OAAO,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,WAAW,IAAI,IAAI,CAAC,sCAAsC,EAAE,CAAC;IAChH,CAAC;IAED,gBAAgB;IACT,4BAA4B,CAAC,WAAmB,EAAE,MAAc;QACnE,+CAA+C;QAC/C,MAAM,KAAK,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,SAAS,EAAE,UAAU,IAAI,CAAC,CAAC;QACjF,MAAM,KAAK,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,gBAAgB,IAAI,CAAC,CAAC;QAC5E,MAAM,UAAU,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,SAAS,EAAE,UAAU,IAAI,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC;QACzH,MAAM,OAAO,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,SAAS,EAAE,aAAa,IAAI,KAAK,CAAC;QAC1F,MAAM,cAAc,GAAG,UAAU,GAAG,OAAO,CAAC;QAE5C,MAAM,QAAQ,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,GAAG,EAAE,WAAW,IAAI,CAAC,CAAC;QAC/E,MAAM,SAAS,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,GAAG,EAAE,UAAU,IAAI,CAAC,CAAC;QAC/E,MAAM,YAAY,GAAG,CAAC,QAAQ,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,SAAS,CAAC;QAErE,MAAM,UAAU,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,SAAS,EAAE,YAAY,IAAI,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC;QACzH,MAAM,cAAc,GAAG,UAAU,GAAG,OAAO,CAAC;QAE5C,MAAM,SAAS,GAAG,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,GAAG,EAAE,UAAU,IAAI,CAAC,CAAC;QAC/E,MAAM,YAAY,GAAG,SAAS,GAAG,SAAS,CAAC;QAE3C,MAAM,cAAc,GAAG,cAAc,GAAG,YAAY,CAAC;QACrD,MAAM,cAAc,GAAG,cAAc,GAAG,YAAY,CAAC;QAErD,MAAM,WAAW,GAAG,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,YAAY,KAAK,oBAAoB,CAAC,KAAK;YAChG,sFAAsF;YACtF,0CAA0C;YAC1C,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,eAAe;YACrC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,eAAe,CAAC;QAC1C,MAAM,aAAa,GAAG,IAAI,CAAC,aAAa,CAAC,QAAQ,CAAC,OAAO,EAAE,YAAY,KAAK,oBAAoB,CAAC,KAAK;YAClG,sFAAsF;YACtF,0CAA0C;YAC1C,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,eAAe;YACrC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,eAAe,CAAC;QAE1C,MAAM,cAAc,GAAG,MAAM,GAAG,cAAc,GAAG,WAAW,GAAG,WAAW,CAAC;QAC3E,MAAM,gBAAgB,GAAG,MAAM,GAAG,cAAc,GAAG,WAAW,GAAG,aAAa,CAAC;QAE/E,OAAO,cAAc,GAAG,gBAAgB,CAAC;IAC7C,CAAC;IAED;;;;;;OAMG;IACI,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,YAA0B,EAAE,KAAa;QAC9D,IAAI,aAAa,GAAG,KAAK,CAAC;QAC1B,IAAI,aAAa,IAAI,IAAI;YACrB,aAAa,GAAG,MAAM,sBAAsB,EAAE,CAAC;QAEnD,OAAO,IAAI,YAAY,CAAC,YAAY,EAAE,aAAa,CAAC,CAAC;IACzD,CAAC;CACJ;AAED,SAAS,eAAe,CAAC,UAAmB;IAGxC,IAAI,UAAU,IAAI,IAAI;QAClB,OAAO,EAAC,WAAW,EAAE,SAAS,EAAC,CAAC;IAEpC,MAAM,iBAAiB,GAAG,MAAM,CAAC;IACjC,IAAI,CAAC,UAAU,CAAC,UAAU,CAAC,iBAAiB,CAAC;QACzC,OAAO,EAAC,WAAW,EAAE,SAAS,EAAC,CAAC;IAEpC,MAAM,QAAQ,GAAG,UAAU,CAAC,OAAO,CAAC,GAAG,EAAE,iBAAiB,CAAC,MAAM,CAAC,CAAC;IACnE,MAAM,iBAAiB,GAAG,UAAU,CAAC,KAAK,CACtC,iBAAiB,CAAC,MAAM,EACxB,QAAQ,GAAG,CAAC;QACR,CAAC,CAAC,UAAU,CAAC,MAAM;QACnB,CAAC,CAAC,QAAQ,CACjB,CAAC;IAEF,MAAM,WAAW,GAAG,QAAQ,CAAC,iBAAiB,CAAC,CAAC;IAChD,IAAI,MAAM,CAAC,QAAQ,CAAC,WAAW,CAAC;QAC5B,OAAO,EAAC,WAAW,EAAC,CAAC;IAEzB,OAAO,EAAC,WAAW,EAAE,SAAS,EAAC,CAAC;AACpC,CAAC;AAED,SAAS,oBAAoB,CAAC,WAA6B,EAAE,KAAY;IACrE,IAAI,IAAI,GAAG,CAAC,CAAC;IACb,KAAK,MAAM,UAAU,IAAI,WAAW;QAChC,IAAI,IAAI,mBAAmB,CAAC,UAAU,EAAE,KAAK,CAAC,CAAC;IAEnD,OAAO,IAAI,CAAC;AAChB,CAAC;AAED,SAAS,mBAAmB,CAAC,MAAsB,EAAE,KAAY;IAC7D,MAAM,QAAQ,GAAG,KAAK,CAAC,SAAS,CAAC,sBAAsB,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;IACzE,MAAM,SAAS,GAAG,KAAK,CAAC,SAAS,CAAC,uBAAuB,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;IAC3E,MAAM,WAAW,GAAG,KAAK,CAAC,OAAO,CAAC,WAAW,CAAC;IAE9C,IAAI,QAAQ,IAAI,IAAI,IAAI,SAAS,IAAI,IAAI;QACrC,MAAM,IAAI,KAAK,CAAC,4BAA4B,CAAC,CAAC;IAElD,MAAM,EAAC,EAAE,EAAE,EAAE,EAAC,GAAG,gBAAgB,CAAC,MAAM,EAAE,EAAC,QAAQ,EAAE,SAAS,EAAE,WAAW,EAAC,CAAC,CAAC;IAE9E,IAAI,SAAS,KAAK,CAAC,EAAE,CAAC;QAClB,IAAI,UAAU,GAAG,QAAQ,CAAC;QAC1B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,WAAW,EAAE,CAAC,EAAE,EAAE,CAAC;YACnC,UAAU,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;QACtC,CAAC;QAED,OAAO,UAAU,CAAC;IACtB,CAAC;SAAM,CAAC;QACJ,IAAI,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,SAAS,CAAC,CAAC;QACzD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,WAAW,EAAE,CAAC,EAAE,EAAE,CAAC;YACnC,UAAU,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;QACtC,CAAC;QAED,OAAO,UAAU,CAAC;IACtB,CAAC;AACL,CAAC;AAED,SAAS,gBAAgB,CAAC,MAAsB,EAAE,EAC9C,QAAQ,EAAE,SAAS,EAAE,WAAW,EAGnC;IACG,qBAAqB;IACrB,6CAA6C;IAC7C,MAAM,EAAE,GAAG;QACP,GAAG,MAAM,CAAC,UAAU;QACpB,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,WAAW,GAAG,MAAM,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;KAC1E,CAAC,KAAK,CAAC,CAAC,EAAE,WAAW,CAAC,CAAC;IAExB,kBAAkB;IAClB,6CAA6C;IAC7C,MAAM,EAAE,GAAG;QACP,QAAQ;QACR,IAAI,CAAC,KAAK,CAAC,QAAQ,GAAG,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,SAAS,CAAC,CAAC;QAC1C,GAAG,KAAK,CAAC,WAAW,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;KACpC,CAAC;IACF,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,WAAW,EAAE,CAAC,EAAE,EAAE,CAAC;QACnC,EAAE,CAAC,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAClC,CAAC;IAED,OAAO;QACH,EAAE;QACF,EAAE;KACL,CAAC;AACN,CAAC"}
@@ -0,0 +1,90 @@
1
+ import { BuildGpu } from "../../bindings/types.js";
2
+ import { LlamaModelOptions } from "../../evaluator/LlamaModel/LlamaModel.js";
3
+ import { LlamaContextOptions } from "../../evaluator/LlamaContext/types.js";
4
+ import type { GgufInsights } from "./GgufInsights.js";
5
+ export declare class GgufInsightsConfigurationResolver {
6
+ private constructor();
7
+ get ggufInsights(): GgufInsights;
8
+ /**
9
+ * Score the compatibility of the model configuration with the current GPU and VRAM state.
10
+ * Assumes a model is loaded with the default `"auto"` configurations.
11
+ * Scored based on the following criteria:
12
+ * - The number of GPU layers that can be offloaded to the GPU (only if there's a GPU. If there's no GPU then by how small the model is)
13
+ * - Whether all layers can be offloaded to the GPU (gives additional points)
14
+ * - Whether the resolved context size is at least as large as the specified `contextSize`
15
+ *
16
+ * IF the resolved context size is larger than the specified context size, for each multiplier of the specified `contextSize`
17
+ * that the resolved context size is larger by, 1 bonus point is given in the `bonusScore`.
18
+ *
19
+ * `contextSize` defaults to `4096` (if the model train context size is lower than this, the model train context size is used instead).
20
+ */
21
+ scoreModelConfigurationCompatibility({ contextSize, embeddingContext, flashAttention }?: {
22
+ contextSize?: number;
23
+ embeddingContext?: boolean;
24
+ flashAttention?: boolean;
25
+ }, { getVramState, getRamState, llamaVramPaddingSize, llamaGpu, llamaSupportsGpuOffloading }?: {
26
+ getVramState?(): Promise<{
27
+ total: number;
28
+ free: number;
29
+ }>;
30
+ getRamState?(): Promise<{
31
+ total: number;
32
+ free: number;
33
+ }>;
34
+ llamaVramPaddingSize?: number;
35
+ llamaGpu?: BuildGpu;
36
+ llamaSupportsGpuOffloading?: boolean;
37
+ }): Promise<{
38
+ /**
39
+ * A number between `0` (inclusive) and `1` (inclusive) representing the compatibility score.
40
+ */
41
+ compatibilityScore: number;
42
+ /**
43
+ * A number starting at `0` with no upper limit representing the bonus score.
44
+ * For each multiplier of the specified `contextSize` that the resolved context size is larger by, 1 bonus point is given.
45
+ */
46
+ bonusScore: number;
47
+ /**
48
+ * The total score, which is the sum of the compatibility and bonus scores.
49
+ */
50
+ totalScore: number;
51
+ /**
52
+ * The resolved values used to calculate the scores.
53
+ */
54
+ resolvedValues: {
55
+ gpuLayers: number;
56
+ contextSize: number;
57
+ modelRamUsage: number;
58
+ contextRamUsage: number;
59
+ totalRamUsage: number;
60
+ modelVramUsage: number;
61
+ contextVramUsage: number;
62
+ totalVramUsage: number;
63
+ };
64
+ }>;
65
+ resolveModelGpuLayers(gpuLayers: LlamaModelOptions["gpuLayers"], { ignoreMemorySafetyChecks, getVramState, llamaVramPaddingSize, llamaGpu, llamaSupportsGpuOffloading, defaultContextFlashAttention }?: {
66
+ ignoreMemorySafetyChecks?: boolean;
67
+ getVramState?(): Promise<{
68
+ total: number;
69
+ free: number;
70
+ }>;
71
+ llamaVramPaddingSize?: number;
72
+ llamaGpu?: BuildGpu;
73
+ llamaSupportsGpuOffloading?: boolean;
74
+ defaultContextFlashAttention?: boolean;
75
+ }): Promise<number>;
76
+ resolveContextContextSize(contextSize: LlamaContextOptions["contextSize"], { modelGpuLayers, batchSize, modelTrainContextSize, flashAttention, getVramState, llamaGpu, ignoreMemorySafetyChecks, isEmbeddingContext, sequences }: {
77
+ modelGpuLayers: number;
78
+ modelTrainContextSize: number;
79
+ flashAttention?: boolean;
80
+ batchSize?: LlamaContextOptions["batchSize"];
81
+ sequences?: number;
82
+ getVramState?(): Promise<{
83
+ total: number;
84
+ free: number;
85
+ }>;
86
+ llamaGpu?: BuildGpu;
87
+ ignoreMemorySafetyChecks?: boolean;
88
+ isEmbeddingContext?: boolean;
89
+ }): Promise<number>;
90
+ }
@@ -0,0 +1,144 @@
1
+ import os from "os";
2
+ import { getDefaultContextSequences } from "../../evaluator/LlamaContext/LlamaContext.js";
3
+ import { resolveModelGpuLayersOption } from "./utils/resolveModelGpuLayersOption.js";
4
+ import { resolveContextContextSizeOption } from "./utils/resolveContextContextSizeOption.js";
5
+ import { scoreLevels } from "./utils/scoreLevels.js";
6
+ export class GgufInsightsConfigurationResolver {
7
+ /** @internal */ _ggufInsights;
8
+ constructor(ggufInsights) {
9
+ this._ggufInsights = ggufInsights;
10
+ }
11
+ get ggufInsights() {
12
+ return this._ggufInsights;
13
+ }
14
+ /**
15
+ * Score the compatibility of the model configuration with the current GPU and VRAM state.
16
+ * Assumes a model is loaded with the default `"auto"` configurations.
17
+ * Scored based on the following criteria:
18
+ * - The number of GPU layers that can be offloaded to the GPU (only if there's a GPU. If there's no GPU then by how small the model is)
19
+ * - Whether all layers can be offloaded to the GPU (gives additional points)
20
+ * - Whether the resolved context size is at least as large as the specified `contextSize`
21
+ *
22
+ * IF the resolved context size is larger than the specified context size, for each multiplier of the specified `contextSize`
23
+ * that the resolved context size is larger by, 1 bonus point is given in the `bonusScore`.
24
+ *
25
+ * `contextSize` defaults to `4096` (if the model train context size is lower than this, the model train context size is used instead).
26
+ */
27
+ async scoreModelConfigurationCompatibility({ contextSize = Math.min(4096, this._ggufInsights.trainContextSize ?? 4096), embeddingContext = false, flashAttention = false } = {}, { getVramState = (() => this._ggufInsights._llama._vramOrchestrator.getMemoryState()), getRamState = (async () => ({ total: os.totalmem(), free: os.freemem() })), llamaVramPaddingSize = this._ggufInsights._llama.vramPaddingSize, llamaGpu = this._ggufInsights._llama.gpu, llamaSupportsGpuOffloading = this._ggufInsights._llama.supportsGpuOffloading } = {}) {
28
+ const defaultTrainContextSize = 4096;
29
+ const [vramState, ramState] = await Promise.all([
30
+ getVramState(),
31
+ getRamState()
32
+ ]);
33
+ const resolvedGpuLayers = await this.resolveModelGpuLayers("auto", {
34
+ getVramState: async () => vramState,
35
+ llamaVramPaddingSize,
36
+ llamaGpu,
37
+ llamaSupportsGpuOffloading
38
+ });
39
+ const canUseGpu = llamaSupportsGpuOffloading && llamaGpu !== false;
40
+ const estimatedModelResourceUsage = this._ggufInsights.estimateModelResourceRequirements({
41
+ gpuLayers: resolvedGpuLayers
42
+ });
43
+ const resolvedContextSize = await this.resolveContextContextSize("auto", {
44
+ getVramState: async () => ({
45
+ total: vramState.total,
46
+ free: Math.max(0, vramState.free - estimatedModelResourceUsage.gpuVram)
47
+ }),
48
+ llamaGpu,
49
+ isEmbeddingContext: embeddingContext,
50
+ modelGpuLayers: resolvedGpuLayers,
51
+ modelTrainContextSize: this._ggufInsights.trainContextSize ?? defaultTrainContextSize
52
+ });
53
+ const estimatedContextResourceUsage = this._ggufInsights.estimateContextResourceRequirements({
54
+ contextSize: resolvedContextSize,
55
+ isEmbeddingContext: embeddingContext,
56
+ modelGpuLayers: resolvedGpuLayers,
57
+ flashAttention
58
+ });
59
+ const rankPoints = {
60
+ gpuLayers: 60,
61
+ allLayersAreOffloaded: 10,
62
+ contextSize: 30,
63
+ ramUsageFitsInRam: 10,
64
+ cpuOnlySmallModelSize: 60, // also defined inside `scoreModelSizeForCpuOnlyUsage`
65
+ bonusContextSize: 10
66
+ };
67
+ const gpuLayersPoints = rankPoints.gpuLayers * Math.min(1, resolvedGpuLayers / this._ggufInsights.totalLayers);
68
+ const allLayersAreOffloadedPoints = rankPoints.allLayersAreOffloaded * (resolvedGpuLayers === this._ggufInsights.totalLayers ? 1 : 0);
69
+ const contextSizePoints = rankPoints.contextSize * Math.min(1, resolvedContextSize / contextSize);
70
+ const ramUsageFitsInRamPoints = rankPoints.ramUsageFitsInRam * (estimatedModelResourceUsage.cpuRam <= ramState.free
71
+ ? 1
72
+ : estimatedModelResourceUsage.cpuRam <= ramState.total
73
+ ? 0.5
74
+ : (0.5 - Math.min(0.5, 0.5 * ((estimatedModelResourceUsage.cpuRam - ramState.total) / ramState.total))));
75
+ const bonusContextSizePoints = 10 * Math.min(1, Math.max(0, resolvedContextSize - contextSize) / contextSize);
76
+ const compatibilityScore = canUseGpu
77
+ ? ((gpuLayersPoints + allLayersAreOffloadedPoints + contextSizePoints + ramUsageFitsInRamPoints) /
78
+ (rankPoints.gpuLayers + rankPoints.allLayersAreOffloaded + rankPoints.contextSize + rankPoints.ramUsageFitsInRam))
79
+ : ((contextSizePoints + ramUsageFitsInRamPoints + scoreModelSizeForCpuOnlyUsage(this._ggufInsights.modelSize)) /
80
+ (rankPoints.contextSize + rankPoints.ramUsageFitsInRam + rankPoints.cpuOnlySmallModelSize));
81
+ const bonusScore = bonusContextSizePoints / rankPoints.bonusContextSize;
82
+ return {
83
+ compatibilityScore,
84
+ bonusScore,
85
+ totalScore: compatibilityScore + bonusScore,
86
+ resolvedValues: {
87
+ gpuLayers: resolvedGpuLayers,
88
+ contextSize: resolvedContextSize,
89
+ modelRamUsage: estimatedModelResourceUsage.cpuRam,
90
+ contextRamUsage: estimatedContextResourceUsage.cpuRam,
91
+ totalRamUsage: estimatedModelResourceUsage.cpuRam + estimatedContextResourceUsage.cpuRam,
92
+ modelVramUsage: estimatedModelResourceUsage.gpuVram,
93
+ contextVramUsage: estimatedContextResourceUsage.gpuVram,
94
+ totalVramUsage: estimatedModelResourceUsage.gpuVram + estimatedContextResourceUsage.gpuVram
95
+ }
96
+ };
97
+ }
98
+ async resolveModelGpuLayers(gpuLayers, { ignoreMemorySafetyChecks = false, getVramState = (() => this._ggufInsights._llama._vramOrchestrator.getMemoryState()), llamaVramPaddingSize = this._ggufInsights._llama.vramPaddingSize, llamaGpu = this._ggufInsights._llama.gpu, llamaSupportsGpuOffloading = this._ggufInsights._llama.supportsGpuOffloading, defaultContextFlashAttention = false } = {}) {
99
+ return resolveModelGpuLayersOption(gpuLayers, {
100
+ ggufInsights: this._ggufInsights,
101
+ ignoreMemorySafetyChecks,
102
+ getVramState,
103
+ llamaVramPaddingSize,
104
+ llamaGpu,
105
+ llamaSupportsGpuOffloading,
106
+ defaultContextFlashAttention
107
+ });
108
+ }
109
+ async resolveContextContextSize(contextSize, { modelGpuLayers, batchSize, modelTrainContextSize, flashAttention = false, getVramState = (() => this._ggufInsights._llama._vramOrchestrator.getMemoryState()), llamaGpu = this._ggufInsights._llama.gpu, ignoreMemorySafetyChecks = false, isEmbeddingContext = false, sequences = getDefaultContextSequences() }) {
110
+ return await resolveContextContextSizeOption({
111
+ contextSize,
112
+ batchSize,
113
+ sequences,
114
+ modelFileInsights: this._ggufInsights,
115
+ modelGpuLayers,
116
+ modelTrainContextSize,
117
+ flashAttention,
118
+ getVramState,
119
+ llamaGpu,
120
+ ignoreMemorySafetyChecks,
121
+ isEmbeddingContext
122
+ });
123
+ }
124
+ /** @internal */
125
+ static _create(ggufInsights) {
126
+ return new GgufInsightsConfigurationResolver(ggufInsights);
127
+ }
128
+ }
129
+ function scoreModelSizeForCpuOnlyUsage(modelSize) {
130
+ const s1GB = Math.pow(1024, 3);
131
+ return 60 - scoreLevels(modelSize, [{
132
+ start: s1GB,
133
+ end: s1GB * 2.5,
134
+ points: 40
135
+ }, {
136
+ start: s1GB * 2.5,
137
+ end: s1GB * 4,
138
+ points: 15
139
+ }, {
140
+ start: s1GB * 4,
141
+ points: 5
142
+ }]);
143
+ }
144
+ //# sourceMappingURL=GgufInsightsConfigurationResolver.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"GgufInsightsConfigurationResolver.js","sourceRoot":"","sources":["../../../src/gguf/insights/GgufInsightsConfigurationResolver.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,IAAI,CAAC;AAIpB,OAAO,EAAC,0BAA0B,EAAC,MAAM,8CAA8C,CAAC;AACxF,OAAO,EAAC,2BAA2B,EAAC,MAAM,wCAAwC,CAAC;AACnF,OAAO,EAAC,+BAA+B,EAAC,MAAM,4CAA4C,CAAC;AAC3F,OAAO,EAAC,WAAW,EAAC,MAAM,wBAAwB,CAAC;AAInD,MAAM,OAAO,iCAAiC;IAC1C,gBAAgB,CAAkB,aAAa,CAAe;IAE9D,YAAoB,YAA0B;QAC1C,IAAI,CAAC,aAAa,GAAG,YAAY,CAAC;IACtC,CAAC;IAED,IAAW,YAAY;QACnB,OAAO,IAAI,CAAC,aAAa,CAAC;IAC9B,CAAC;IAED;;;;;;;;;;;;OAYG;IACI,KAAK,CAAC,oCAAoC,CAAC,EAC9C,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,aAAa,CAAC,gBAAgB,IAAI,IAAI,CAAC,EACzE,gBAAgB,GAAG,KAAK,EACxB,cAAc,GAAG,KAAK,KAKtB,EAAE,EAAE,EACJ,YAAY,GAAG,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,iBAAiB,CAAC,cAAc,EAAE,CAAC,EACnF,WAAW,GAAG,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC,EAAC,KAAK,EAAE,EAAE,CAAC,QAAQ,EAAE,EAAE,IAAI,EAAE,EAAE,CAAC,OAAO,EAAE,EAAC,CAAC,CAAC,EACxE,oBAAoB,GAAG,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,eAAe,EAChE,QAAQ,GAAG,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,GAAG,EACxC,0BAA0B,GAAG,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,qBAAqB,KAO5E,EAAE;QAiCF,MAAM,uBAAuB,GAAG,IAAI,CAAC;QAErC,MAAM,CACF,SAAS,EACT,QAAQ,CACX,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;YAClB,YAAY,EAAE;YACd,WAAW,EAAE;SAChB,CAAC,CAAC;QACH,MAAM,iBAAiB,GAAG,MAAM,IAAI,CAAC,qBAAqB,CAAC,MAAM,EAAE;YAC/D,YAAY,EAAE,KAAK,IAAI,EAAE,CAAC,SAAS;YACnC,oBAAoB;YACpB,QAAQ;YACR,0BAA0B;SAC7B,CAAC,CAAC;QACH,MAAM,SAAS,GAAG,0BAA0B,IAAI,QAAQ,KAAK,KAAK,CAAC;QACnE,MAAM,2BAA2B,GAAG,IAAI,CAAC,aAAa,CAAC,iCAAiC,CAAC;YACrF,SAAS,EAAE,iBAAiB;SAC/B,CAAC,CAAC;QAEH,MAAM,mBAAmB,GAAG,MAAM,IAAI,CAAC,yBAAyB,CAAC,MAAM,EAAE;YACrE,YAAY,EAAE,KAAK,IAAI,EAAE,CAAC,CAAC;gBACvB,KAAK,EAAE,SAAS,CAAC,KAAK;gBACtB,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,SAAS,CAAC,IAAI,GAAG,2BAA2B,CAAC,OAAO,CAAC;aAC1E,CAAC;YACF,QAAQ;YACR,kBAAkB,EAAE,gBAAgB;YACpC,cAAc,EAAE,iBAAiB;YACjC,qBAAqB,EAAE,IAAI,CAAC,aAAa,CAAC,gBAAgB,IAAI,uBAAuB;SACxF,CAAC,CAAC;QACH,MAAM,6BAA6B,GAAG,IAAI,CAAC,aAAa,CAAC,mCAAmC,CAAC;YACzF,WAAW,EAAE,mBAAmB;YAChC,kBAAkB,EAAE,gBAAgB;YACpC,cAAc,EAAE,iBAAiB;YACjC,cAAc;SACjB,CAAC,CAAC;QAEH,MAAM,UAAU,GAAG;YACf,SAAS,EAAE,EAAE;YACb,qBAAqB,EAAE,EAAE;YACzB,WAAW,EAAE,EAAE;YACf,iBAAiB,EAAE,EAAE;YACrB,qBAAqB,EAAE,EAAE,EAAE,sDAAsD;YACjF,gBAAgB,EAAE,EAAE;SACd,CAAC;QAEX,MAAM,eAAe,GAAG,UAAU,CAAC,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,iBAAiB,GAAG,IAAI,CAAC,aAAa,CAAC,WAAW,CAAC,CAAC;QAC/G,MAAM,2BAA2B,GAAG,UAAU,CAAC,qBAAqB,GAAG,CACnE,iBAAiB,KAAK,IAAI,CAAC,aAAa,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAC/D,CAAC;QACF,MAAM,iBAAiB,GAAG,UAAU,CAAC,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,mBAAmB,GAAG,WAAW,CAAC,CAAC;QAClG,MAAM,uBAAuB,GAAG,UAAU,CAAC,iBAAiB,GAAG,CAC3D,2BAA2B,CAAC,MAAM,IAAI,QAAQ,CAAC,IAAI;YAC/C,CAAC,CAAC,CAAC;YACH,CAAC,CAAC,2BAA2B,CAAC,MAAM,IAAI,QAAQ,CAAC,KAAK;gBAClD,CAAC,CAAC,GAAG;gBACL,CAAC,CAAC,CACE,GAAG,GAAG,IAAI,CAAC,GAAG,CACV,GAAG,EACH,GAAG,GAAG,CACF,CAAC,2BAA2B,CAAC,MAAM,GAAG,QAAQ,CAAC,KAAK,CAAC,GAAG,QAAQ,CAAC,KAAK,CACzE,CACJ,CACJ,CACZ,CAAC;QACF,MAAM,sBAAsB,GAAG,EAAE,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,mBAAmB,GAAG,WAAW,CAAC,GAAG,WAAW,CAAC,CAAC;QAE9G,MAAM,kBAAkB,GAAG,SAAS;YAChC,CAAC,CAAC,CACE,CAAC,eAAe,GAAG,2BAA2B,GAAG,iBAAiB,GAAG,uBAAuB,CAAC;gBAC7F,CAAC,UAAU,CAAC,SAAS,GAAG,UAAU,CAAC,qBAAqB,GAAG,UAAU,CAAC,WAAW,GAAG,UAAU,CAAC,iBAAiB,CAAC,CACpH;YACD,CAAC,CAAC,CACE,CAAC,iBAAiB,GAAG,uBAAuB,GAAG,6BAA6B,CAAC,IAAI,CAAC,aAAa,CAAC,SAAS,CAAC,CAAC;gBAC3G,CAAC,UAAU,CAAC,WAAW,GAAG,UAAU,CAAC,iBAAiB,GAAG,UAAU,CAAC,qBAAqB,CAAC,CAAC,CAAC;QACpG,MAAM,UAAU,GAAG,sBAAsB,GAAG,UAAU,CAAC,gBAAgB,CAAC;QAExE,OAAO;YACH,kBAAkB;YAClB,UAAU;YACV,UAAU,EAAE,kBAAkB,GAAG,UAAU;YAE3C,cAAc,EAAE;gBACZ,SAAS,EAAE,iBAAiB;gBAC5B,WAAW,EAAE,mBAAmB;gBAEhC,aAAa,EAAE,2BAA2B,CAAC,MAAM;gBACjD,eAAe,EAAE,6BAA6B,CAAC,MAAM;gBACrD,aAAa,EAAE,2BAA2B,CAAC,MAAM,GAAG,6BAA6B,CAAC,MAAM;gBAExF,cAAc,EAAE,2BAA2B,CAAC,OAAO;gBACnD,gBAAgB,EAAE,6BAA6B,CAAC,OAAO;gBACvD,cAAc,EAAE,2BAA2B,CAAC,OAAO,GAAG,6BAA6B,CAAC,OAAO;aAC9F;SACJ,CAAC;IACN,CAAC;IAEM,KAAK,CAAC,qBAAqB,CAAC,SAAyC,EAAE,EAC1E,wBAAwB,GAAG,KAAK,EAChC,YAAY,GAAG,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,iBAAiB,CAAC,cAAc,EAAE,CAAC,EACnF,oBAAoB,GAAG,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,eAAe,EAAE,QAAQ,GAAG,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,GAAG,EAC1G,0BAA0B,GAAG,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,qBAAqB,EAC5E,4BAA4B,GAAG,KAAK,KAIpC,EAAE;QACF,OAAO,2BAA2B,CAAC,SAAS,EAAE;YAC1C,YAAY,EAAE,IAAI,CAAC,aAAa;YAChC,wBAAwB;YACxB,YAAY;YACZ,oBAAoB;YACpB,QAAQ;YACR,0BAA0B;YAC1B,4BAA4B;SAC/B,CAAC,CAAC;IACP,CAAC;IAEM,KAAK,CAAC,yBAAyB,CAAC,WAA+C,EAAE,EACpF,cAAc,EACd,SAAS,EACT,qBAAqB,EACrB,cAAc,GAAG,KAAK,EACtB,YAAY,GAAG,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,iBAAiB,CAAC,cAAc,EAAE,CAAC,EACnF,QAAQ,GAAG,IAAI,CAAC,aAAa,CAAC,MAAM,CAAC,GAAG,EACxC,wBAAwB,GAAG,KAAK,EAChC,kBAAkB,GAAG,KAAK,EAC1B,SAAS,GAAG,0BAA0B,EAAE,EAW3C;QACG,OAAO,MAAM,+BAA+B,CAAC;YACzC,WAAW;YACX,SAAS;YACT,SAAS;YACT,iBAAiB,EAAE,IAAI,CAAC,aAAa;YACrC,cAAc;YACd,qBAAqB;YACrB,cAAc;YACd,YAAY;YACZ,QAAQ;YACR,wBAAwB;YACxB,kBAAkB;SACrB,CAAC,CAAC;IACP,CAAC;IAED,gBAAgB;IACT,MAAM,CAAC,OAAO,CAAC,YAA0B;QAC5C,OAAO,IAAI,iCAAiC,CAAC,YAAY,CAAC,CAAC;IAC/D,CAAC;CACJ;AAED,SAAS,6BAA6B,CAAC,SAAiB;IACpD,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IAC/B,OAAO,EAAE,GAAG,WAAW,CAAC,SAAS,EAAE,CAAC;YAChC,KAAK,EAAE,IAAI;YACX,GAAG,EAAE,IAAI,GAAG,GAAG;YACf,MAAM,EAAE,EAAE;SACb,EAAE;YACC,KAAK,EAAE,IAAI,GAAG,GAAG;YACjB,GAAG,EAAE,IAAI,GAAG,CAAC;YACb,MAAM,EAAE,EAAE;SACb,EAAE;YACC,KAAK,EAAE,IAAI,GAAG,CAAC;YACf,MAAM,EAAE,CAAC;SACZ,CAAC,CAAC,CAAC;AACR,CAAC"}
@@ -0,0 +1,19 @@
1
+ import { LlamaContextOptions } from "../../../evaluator/LlamaContext/types.js";
2
+ import { GgufInsights } from "../GgufInsights.js";
3
+ import { BuildGpu } from "../../../bindings/types.js";
4
+ export declare function resolveContextContextSizeOption({ contextSize, batchSize, sequences, modelFileInsights, modelGpuLayers, modelTrainContextSize, flashAttention, getVramState, llamaGpu, ignoreMemorySafetyChecks, isEmbeddingContext }: {
5
+ contextSize?: LlamaContextOptions["contextSize"];
6
+ batchSize?: LlamaContextOptions["batchSize"];
7
+ sequences: number;
8
+ modelFileInsights: GgufInsights;
9
+ modelGpuLayers: number;
10
+ modelTrainContextSize: number;
11
+ flashAttention: boolean;
12
+ getVramState(): Promise<{
13
+ total: number;
14
+ free: number;
15
+ }>;
16
+ llamaGpu: BuildGpu;
17
+ ignoreMemorySafetyChecks?: boolean;
18
+ isEmbeddingContext?: boolean;
19
+ }): Promise<number>;
@@ -0,0 +1,78 @@
1
+ import { minAllowedContextSizeInCalculations } from "../../../config.js";
2
+ import { getDefaultContextBatchSize, getDefaultModelContextSize } from "../../../evaluator/LlamaContext/LlamaContext.js";
3
+ export async function resolveContextContextSizeOption({ contextSize, batchSize, sequences, modelFileInsights, modelGpuLayers, modelTrainContextSize, flashAttention, getVramState, llamaGpu, ignoreMemorySafetyChecks = false, isEmbeddingContext = false }) {
4
+ if (contextSize == null)
5
+ contextSize = "auto";
6
+ if (typeof contextSize === "number") {
7
+ const resolvedContextSize = Math.max(1, Math.floor(contextSize));
8
+ if (ignoreMemorySafetyChecks)
9
+ return resolvedContextSize;
10
+ const vramState = await getVramState();
11
+ const contextVram = modelFileInsights.estimateContextResourceRequirements({
12
+ contextSize: resolvedContextSize,
13
+ batchSize: batchSize ?? getDefaultContextBatchSize({ contextSize: resolvedContextSize, sequences }),
14
+ modelGpuLayers: modelGpuLayers,
15
+ sequences,
16
+ flashAttention,
17
+ isEmbeddingContext
18
+ }).gpuVram;
19
+ if (contextVram > vramState.free)
20
+ throw new Error(`The context size of ${resolvedContextSize}${sequences > 1 ? ` with ${sequences} sequences` : ""} is too large for the available VRAM`);
21
+ return resolvedContextSize;
22
+ }
23
+ else if (contextSize === "auto" || typeof contextSize === "object") {
24
+ if (llamaGpu === false)
25
+ return modelTrainContextSize;
26
+ const vramState = await getVramState();
27
+ if (vramState.total === 0)
28
+ return modelTrainContextSize;
29
+ const freeVram = vramState.free;
30
+ const maxContextSize = contextSize === "auto"
31
+ ? getDefaultModelContextSize({ trainContextSize: modelTrainContextSize })
32
+ : Math.min(contextSize.max ?? getDefaultModelContextSize({ trainContextSize: modelTrainContextSize }), getDefaultModelContextSize({ trainContextSize: modelTrainContextSize }));
33
+ const minContextSize = contextSize === "auto"
34
+ ? minAllowedContextSizeInCalculations
35
+ : Math.max(contextSize.min ?? minAllowedContextSizeInCalculations, minAllowedContextSizeInCalculations);
36
+ let highestCompatibleContextSize = null;
37
+ let step = -Math.max(1, Math.floor((maxContextSize - minContextSize) / 4));
38
+ for (let testContextSize = maxContextSize; testContextSize >= minContextSize && testContextSize <= maxContextSize;) {
39
+ const contextVram = modelFileInsights.estimateContextResourceRequirements({
40
+ contextSize: testContextSize,
41
+ batchSize: batchSize ?? getDefaultContextBatchSize({ contextSize: testContextSize, sequences }),
42
+ modelGpuLayers: modelGpuLayers,
43
+ sequences,
44
+ flashAttention,
45
+ isEmbeddingContext
46
+ }).gpuVram;
47
+ if (contextVram <= freeVram) {
48
+ if (highestCompatibleContextSize == null || testContextSize > highestCompatibleContextSize) {
49
+ highestCompatibleContextSize = testContextSize;
50
+ if (step === -1)
51
+ break;
52
+ else if (step < 0)
53
+ step = Math.max(1, Math.floor(-step / 2));
54
+ }
55
+ }
56
+ else if (step > 0)
57
+ step = -Math.max(1, Math.floor(step / 2));
58
+ if (testContextSize == minContextSize && step === -1)
59
+ break;
60
+ testContextSize += step;
61
+ if (testContextSize < minContextSize) {
62
+ testContextSize = minContextSize;
63
+ step = Math.max(1, Math.floor(Math.abs(step) / 2));
64
+ }
65
+ else if (testContextSize > maxContextSize) {
66
+ testContextSize = maxContextSize;
67
+ step = -Math.max(1, Math.floor(Math.abs(step) / 2));
68
+ }
69
+ }
70
+ if (highestCompatibleContextSize != null)
71
+ return highestCompatibleContextSize;
72
+ if (ignoreMemorySafetyChecks)
73
+ return minContextSize;
74
+ throw new Error(`The available VRAM is too small to fit the context size of ${maxContextSize}${sequences > 1 ? ` with ${sequences} sequences` : ""}`);
75
+ }
76
+ throw new Error(`Invalid context size: "${contextSize}"`);
77
+ }
78
+ //# sourceMappingURL=resolveContextContextSizeOption.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"resolveContextContextSizeOption.js","sourceRoot":"","sources":["../../../../src/gguf/insights/utils/resolveContextContextSizeOption.ts"],"names":[],"mappings":"AAGA,OAAO,EAAC,mCAAmC,EAAC,MAAM,oBAAoB,CAAC;AACvE,OAAO,EAAC,0BAA0B,EAAE,0BAA0B,EAAC,MAAM,iDAAiD,CAAC;AAEvH,MAAM,CAAC,KAAK,UAAU,+BAA+B,CAAC,EAClD,WAAW,EAAE,SAAS,EAAE,SAAS,EAAE,iBAAiB,EAAE,cAAc,EAAE,qBAAqB,EAAE,cAAc,EAAE,YAAY,EAAE,QAAQ,EACnI,wBAAwB,GAAG,KAAK,EAAE,kBAAkB,GAAG,KAAK,EAa/D;IACG,IAAI,WAAW,IAAI,IAAI;QACnB,WAAW,GAAG,MAAM,CAAC;IAEzB,IAAI,OAAO,WAAW,KAAK,QAAQ,EAAE,CAAC;QAClC,MAAM,mBAAmB,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,WAAW,CAAC,CAAC,CAAC;QAEjE,IAAI,wBAAwB;YACxB,OAAO,mBAAmB,CAAC;QAE/B,MAAM,SAAS,GAAG,MAAM,YAAY,EAAE,CAAC;QACvC,MAAM,WAAW,GAAG,iBAAiB,CAAC,mCAAmC,CAAC;YACtE,WAAW,EAAE,mBAAmB;YAChC,SAAS,EAAE,SAAS,IAAI,0BAA0B,CAAC,EAAC,WAAW,EAAE,mBAAmB,EAAE,SAAS,EAAC,CAAC;YACjG,cAAc,EAAE,cAAc;YAC9B,SAAS;YACT,cAAc;YACd,kBAAkB;SACrB,CAAC,CAAC,OAAO,CAAC;QAEX,IAAI,WAAW,GAAG,SAAS,CAAC,IAAI;YAC5B,MAAM,IAAI,KAAK,CAAC,uBAAuB,mBAAmB,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,SAAS,YAAY,CAAC,CAAC,CAAC,EAAE,sCAAsC,CAAC,CAAC;QAE5J,OAAO,mBAAmB,CAAC;IAC/B,CAAC;SAAM,IAAI,WAAW,KAAK,MAAM,IAAI,OAAO,WAAW,KAAK,QAAQ,EAAE,CAAC;QACnE,IAAI,QAAQ,KAAK,KAAK;YAClB,OAAO,qBAAqB,CAAC;QAEjC,MAAM,SAAS,GAAG,MAAM,YAAY,EAAE,CAAC;QAEvC,IAAI,SAAS,CAAC,KAAK,KAAK,CAAC;YACrB,OAAO,qBAAqB,CAAC;QAEjC,MAAM,QAAQ,GAAG,SAAS,CAAC,IAAI,CAAC;QAEhC,MAAM,cAAc,GAAG,WAAW,KAAK,MAAM;YACzC,CAAC,CAAC,0BAA0B,CAAC,EAAC,gBAAgB,EAAE,qBAAqB,EAAC,CAAC;YACvE,CAAC,CAAC,IAAI,CAAC,GAAG,CACN,WAAW,CAAC,GAAG,IAAI,0BAA0B,CAAC,EAAC,gBAAgB,EAAE,qBAAqB,EAAC,CAAC,EACxF,0BAA0B,CAAC,EAAC,gBAAgB,EAAE,qBAAqB,EAAC,CAAC,CACxE,CAAC;QAEN,MAAM,cAAc,GAAG,WAAW,KAAK,MAAM;YACzC,CAAC,CAAC,mCAAmC;YACrC,CAAC,CAAC,IAAI,CAAC,GAAG,CACN,WAAW,CAAC,GAAG,IAAI,mCAAmC,EACtD,mCAAmC,CACtC,CAAC;QAEN,IAAI,4BAA4B,GAAkB,IAAI,CAAC;QACvD,IAAI,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,cAAc,GAAG,cAAc,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;QAC3E,KAAK,IAAI,eAAe,GAAG,cAAc,EAAE,eAAe,IAAI,cAAc,IAAI,eAAe,IAAI,cAAc,GAAG,CAAC;YACjH,MAAM,WAAW,GAAG,iBAAiB,CAAC,mCAAmC,CAAC;gBACtE,WAAW,EAAE,eAAe;gBAC5B,SAAS,EAAE,SAAS,IAAI,0BAA0B,CAAC,EAAC,WAAW,EAAE,eAAe,EAAE,SAAS,EAAC,CAAC;gBAC7F,cAAc,EAAE,cAAc;gBAC9B,SAAS;gBACT,cAAc;gBACd,kBAAkB;aACrB,CAAC,CAAC,OAAO,CAAC;YAEX,IAAI,WAAW,IAAI,QAAQ,EAAE,CAAC;gBAC1B,IAAI,4BAA4B,IAAI,IAAI,IAAI,eAAe,GAAG,4BAA4B,EAAE,CAAC;oBACzF,4BAA4B,GAAG,eAAe,CAAC;oBAE/C,IAAI,IAAI,KAAK,CAAC,CAAC;wBACX,MAAM;yBACL,IAAI,IAAI,GAAG,CAAC;wBACb,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC;gBAClD,CAAC;YACL,CAAC;iBAAM,IAAI,IAAI,GAAG,CAAC;gBACf,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC;YAE9C,IAAI,eAAe,IAAI,cAAc,IAAI,IAAI,KAAK,CAAC,CAAC;gBAChD,MAAM;YAEV,eAAe,IAAI,IAAI,CAAC;YACxB,IAAI,eAAe,GAAG,cAAc,EAAE,CAAC;gBACnC,eAAe,GAAG,cAAc,CAAC;gBACjC,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;YACvD,CAAC;iBAAM,IAAI,eAAe,GAAG,cAAc,EAAE,CAAC;gBAC1C,eAAe,GAAG,cAAc,CAAC;gBACjC,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;YACxD,CAAC;QACL,CAAC;QAED,IAAI,4BAA4B,IAAI,IAAI;YACpC,OAAO,4BAA4B,CAAC;QAExC,IAAI,wBAAwB;YACxB,OAAO,cAAc,CAAC;QAE1B,MAAM,IAAI,KAAK,CAAC,8DAA8D,cAAc,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,SAAS,YAAY,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;IAC1J,CAAC;IAED,MAAM,IAAI,KAAK,CAAC,0BAA0B,WAAW,GAAG,CAAC,CAAC;AAC9D,CAAC"}
@@ -0,0 +1,15 @@
1
+ import { LlamaModelOptions } from "../../../evaluator/LlamaModel/LlamaModel.js";
2
+ import { BuildGpu } from "../../../bindings/types.js";
3
+ import type { GgufInsights } from "../GgufInsights.js";
4
+ export declare function resolveModelGpuLayersOption(gpuLayers: LlamaModelOptions["gpuLayers"], { ggufInsights, ignoreMemorySafetyChecks, getVramState, llamaVramPaddingSize, llamaGpu, llamaSupportsGpuOffloading, defaultContextFlashAttention }: {
5
+ ggufInsights: GgufInsights;
6
+ ignoreMemorySafetyChecks?: boolean;
7
+ getVramState(): Promise<{
8
+ total: number;
9
+ free: number;
10
+ }>;
11
+ llamaVramPaddingSize: number;
12
+ llamaGpu: BuildGpu;
13
+ llamaSupportsGpuOffloading: boolean;
14
+ defaultContextFlashAttention: boolean;
15
+ }): Promise<number>;
@@ -0,0 +1,183 @@
1
+ import { InsufficientMemoryError } from "../../../utils/InsufficientMemoryError.js";
2
+ import { findBestOption } from "../../../utils/findBestOption.js";
3
+ import { getDefaultContextBatchSize, getDefaultModelContextSize } from "../../../evaluator/LlamaContext/LlamaContext.js";
4
+ import { minAllowedContextSizeInCalculations } from "../../../config.js";
5
+ import { scoreLevels } from "./scoreLevels.js";
6
+ const fitContextExtraMemoryPaddingPercentage = 0.5;
7
+ export async function resolveModelGpuLayersOption(gpuLayers, { ggufInsights, ignoreMemorySafetyChecks = false, getVramState, llamaVramPaddingSize, llamaGpu, llamaSupportsGpuOffloading, defaultContextFlashAttention }) {
8
+ if (gpuLayers == null)
9
+ gpuLayers = "auto";
10
+ if (!llamaSupportsGpuOffloading)
11
+ return 0;
12
+ if (gpuLayers === "max" || typeof gpuLayers === "number") {
13
+ const resolvedGpuLayers = typeof gpuLayers === "number"
14
+ ? Math.max(0, Math.min(ggufInsights.totalLayers, gpuLayers))
15
+ : ggufInsights.totalLayers;
16
+ if (ignoreMemorySafetyChecks)
17
+ return resolvedGpuLayers;
18
+ const vramState = await getVramState();
19
+ const maxLayersRequirements = getVramRequiredForGpuLayers({
20
+ gpuLayers: resolvedGpuLayers,
21
+ ggufInsights,
22
+ currentVram: vramState.free,
23
+ defaultContextFlashAttention
24
+ });
25
+ if (maxLayersRequirements == null)
26
+ throw new InsufficientMemoryError("Not enough VRAM to fit the model with the specified settings");
27
+ return resolvedGpuLayers;
28
+ }
29
+ else if (gpuLayers === "auto" || typeof gpuLayers === "object") {
30
+ if (llamaGpu === false)
31
+ return 0;
32
+ const vramState = await getVramState();
33
+ if (vramState.total === 0)
34
+ return 0;
35
+ let freeVram = vramState.free;
36
+ if (typeof gpuLayers === "object" && gpuLayers.fitContext?.contextSize != null) {
37
+ freeVram -= llamaVramPaddingSize * fitContextExtraMemoryPaddingPercentage;
38
+ if (freeVram < 0)
39
+ freeVram = 0;
40
+ }
41
+ const bestGpuLayersOption = getBestGpuLayersForFreeVram({
42
+ ggufInsights,
43
+ freeVram,
44
+ fitContext: typeof gpuLayers === "object"
45
+ ? gpuLayers.fitContext
46
+ : undefined,
47
+ minGpuLayers: typeof gpuLayers === "object"
48
+ ? gpuLayers.min
49
+ : undefined,
50
+ maxGpuLayers: typeof gpuLayers === "object"
51
+ ? gpuLayers.max
52
+ : undefined,
53
+ defaultContextFlashAttention
54
+ });
55
+ const hasGpuLayersRequirements = typeof gpuLayers === "object" &&
56
+ (gpuLayers.min != null || gpuLayers.max != null || gpuLayers.fitContext?.contextSize != null);
57
+ if (!ignoreMemorySafetyChecks && bestGpuLayersOption == null && hasGpuLayersRequirements)
58
+ throw new InsufficientMemoryError("Not enough VRAM to fit the model with the specified settings");
59
+ return bestGpuLayersOption ?? 0;
60
+ }
61
+ throw new Error(`Invalid gpuLayers value: ${gpuLayers}`);
62
+ }
63
+ function getBestGpuLayersForFreeVram({ ggufInsights, freeVram, fitContext, minGpuLayers, maxGpuLayers, defaultContextFlashAttention }) {
64
+ return findBestOption({
65
+ *generator() {
66
+ const minLayers = Math.floor(Math.max(0, minGpuLayers ?? 0));
67
+ const maxLayers = Math.floor(Math.min(ggufInsights.totalLayers, maxGpuLayers ?? ggufInsights.totalLayers));
68
+ for (let layers = maxLayers; layers >= minLayers; layers--) {
69
+ yield {
70
+ gpuLayers: layers
71
+ };
72
+ }
73
+ },
74
+ score(option) {
75
+ const layersRequirements = getVramRequiredForGpuLayers({
76
+ gpuLayers: option.gpuLayers,
77
+ ggufInsights,
78
+ currentVram: freeVram,
79
+ fitContext,
80
+ defaultContextFlashAttention
81
+ });
82
+ if (layersRequirements == null)
83
+ return null;
84
+ return scoreGpuLayersAndContextCombination({ gpuLayers: option.gpuLayers, contextSize: layersRequirements.contextSize }, {
85
+ totalGpuLayers: ggufInsights.totalLayers,
86
+ trainContextSize: getDefaultModelContextSize({ trainContextSize: ggufInsights.trainContextSize })
87
+ });
88
+ }
89
+ })?.gpuLayers ?? null;
90
+ }
91
+ function scoreGpuLayersAndContextCombination({ gpuLayers, contextSize }, { totalGpuLayers, trainContextSize }) {
92
+ function scoreGpuLayers() {
93
+ return scoreLevels(gpuLayers, [{
94
+ start: 0,
95
+ points: 4
96
+ }, {
97
+ start: 1,
98
+ points: 26
99
+ }, {
100
+ start: totalGpuLayers,
101
+ points: 14,
102
+ end: totalGpuLayers
103
+ }]);
104
+ }
105
+ function scoreContextSize() {
106
+ const gpuLayersPercentage = gpuLayers / totalGpuLayers;
107
+ return scoreLevels(contextSize, [{
108
+ start: 0,
109
+ points: 2
110
+ }, {
111
+ start: 1024,
112
+ points: 4
113
+ }, {
114
+ start: 2048,
115
+ points: gpuLayersPercentage < 0.1 ? 1 : 8
116
+ }, {
117
+ start: 4096,
118
+ points: gpuLayersPercentage < 0.3 ? 4 : 16
119
+ }, {
120
+ start: 8192,
121
+ points: gpuLayersPercentage < 0.6 ? 1 : 8,
122
+ end: Math.max(trainContextSize, 16384)
123
+ }]);
124
+ }
125
+ return scoreGpuLayers() + scoreContextSize();
126
+ }
127
+ function getVramRequiredForGpuLayers({ gpuLayers, ggufInsights, currentVram, fitContext, defaultContextFlashAttention = false }) {
128
+ const modelVram = ggufInsights.estimateModelResourceRequirements({ gpuLayers }).gpuVram;
129
+ if (modelVram > currentVram)
130
+ return null;
131
+ if (fitContext != null && fitContext.contextSize != null) {
132
+ const contextVram = ggufInsights.estimateContextResourceRequirements({
133
+ contextSize: fitContext.contextSize,
134
+ batchSize: getDefaultContextBatchSize({ contextSize: fitContext.contextSize, sequences: 1 }),
135
+ modelGpuLayers: gpuLayers,
136
+ sequences: 1,
137
+ isEmbeddingContext: fitContext.embeddingContext ?? false,
138
+ flashAttention: defaultContextFlashAttention
139
+ }).gpuVram;
140
+ const totalVram = modelVram + contextVram;
141
+ if (totalVram > currentVram)
142
+ return null;
143
+ return {
144
+ contextSize: fitContext.contextSize,
145
+ contextVram,
146
+ totalVram
147
+ };
148
+ }
149
+ const maxContext = findMaxPossibleContextSizeForVram({
150
+ gpuLayers,
151
+ ggufInsights,
152
+ vram: currentVram - modelVram,
153
+ isEmbeddingContext: fitContext?.embeddingContext ?? false,
154
+ flashAttention: defaultContextFlashAttention
155
+ });
156
+ if (maxContext == null || modelVram + maxContext.vram > currentVram)
157
+ return null;
158
+ return {
159
+ contextSize: maxContext.contextSize,
160
+ contextVram: maxContext.vram,
161
+ totalVram: modelVram + maxContext.vram
162
+ };
163
+ }
164
+ function findMaxPossibleContextSizeForVram({ gpuLayers, ggufInsights, vram, isEmbeddingContext, flashAttention }) {
165
+ const maxContextSize = getDefaultModelContextSize({ trainContextSize: ggufInsights.trainContextSize });
166
+ for (let contextSize = maxContextSize; contextSize >= minAllowedContextSizeInCalculations; contextSize--) {
167
+ const contextVram = ggufInsights.estimateContextResourceRequirements({
168
+ contextSize,
169
+ batchSize: getDefaultContextBatchSize({ contextSize, sequences: 1 }),
170
+ modelGpuLayers: gpuLayers,
171
+ sequences: 1,
172
+ isEmbeddingContext,
173
+ flashAttention
174
+ }).gpuVram;
175
+ if (contextVram <= vram)
176
+ return {
177
+ contextSize,
178
+ vram: contextVram
179
+ };
180
+ }
181
+ return null;
182
+ }
183
+ //# sourceMappingURL=resolveModelGpuLayersOption.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"resolveModelGpuLayersOption.js","sourceRoot":"","sources":["../../../../src/gguf/insights/utils/resolveModelGpuLayersOption.ts"],"names":[],"mappings":"AAEA,OAAO,EAAC,uBAAuB,EAAC,MAAM,2CAA2C,CAAC;AAClF,OAAO,EAAC,cAAc,EAAC,MAAM,kCAAkC,CAAC;AAChE,OAAO,EAAC,0BAA0B,EAAE,0BAA0B,EAAC,MAAM,iDAAiD,CAAC;AACvH,OAAO,EAAC,mCAAmC,EAAC,MAAM,oBAAoB,CAAC;AACvE,OAAO,EAAC,WAAW,EAAC,MAAM,kBAAkB,CAAC;AAG7C,MAAM,sCAAsC,GAAG,GAAG,CAAC;AAEnD,MAAM,CAAC,KAAK,UAAU,2BAA2B,CAAC,SAAyC,EAAE,EACzF,YAAY,EAAE,wBAAwB,GAAG,KAAK,EAAE,YAAY,EAAE,oBAAoB,EAClF,QAAQ,EAAE,0BAA0B,EAAE,4BAA4B,EAKrE;IACG,IAAI,SAAS,IAAI,IAAI;QACjB,SAAS,GAAG,MAAM,CAAC;IAEvB,IAAI,CAAC,0BAA0B;QAC3B,OAAO,CAAC,CAAC;IAEb,IAAI,SAAS,KAAK,KAAK,IAAI,OAAO,SAAS,KAAK,QAAQ,EAAE,CAAC;QACvD,MAAM,iBAAiB,GAAG,OAAO,SAAS,KAAK,QAAQ;YACnD,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,YAAY,CAAC,WAAW,EAAE,SAAS,CAAC,CAAC;YAC5D,CAAC,CAAC,YAAY,CAAC,WAAW,CAAC;QAE/B,IAAI,wBAAwB;YACxB,OAAO,iBAAiB,CAAC;QAE7B,MAAM,SAAS,GAAG,MAAM,YAAY,EAAE,CAAC;QACvC,MAAM,qBAAqB,GAAG,2BAA2B,CAAC;YACtD,SAAS,EAAE,iBAAiB;YAC5B,YAAY;YACZ,WAAW,EAAE,SAAS,CAAC,IAAI;YAC3B,4BAA4B;SAC/B,CAAC,CAAC;QAEH,IAAI,qBAAqB,IAAI,IAAI;YAC7B,MAAM,IAAI,uBAAuB,CAAC,8DAA8D,CAAC,CAAC;QAEtG,OAAO,iBAAiB,CAAC;IAC7B,CAAC;SAAM,IAAI,SAAS,KAAK,MAAM,IAAI,OAAO,SAAS,KAAK,QAAQ,EAAE,CAAC;QAC/D,IAAI,QAAQ,KAAK,KAAK;YAClB,OAAO,CAAC,CAAC;QAEb,MAAM,SAAS,GAAG,MAAM,YAAY,EAAE,CAAC;QACvC,IAAI,SAAS,CAAC,KAAK,KAAK,CAAC;YACrB,OAAO,CAAC,CAAC;QAEb,IAAI,QAAQ,GAAG,SAAS,CAAC,IAAI,CAAC;QAC9B,IAAI,OAAO,SAAS,KAAK,QAAQ,IAAI,SAAS,CAAC,UAAU,EAAE,WAAW,IAAI,IAAI,EAAE,CAAC;YAC7E,QAAQ,IAAI,oBAAoB,GAAG,sCAAsC,CAAC;YAE1E,IAAI,QAAQ,GAAG,CAAC;gBACZ,QAAQ,GAAG,CAAC,CAAC;QACrB,CAAC;QAED,MAAM,mBAAmB,GAAG,2BAA2B,CAAC;YACpD,YAAY;YACZ,QAAQ;YACR,UAAU,EAAE,OAAO,SAAS,KAAK,QAAQ;gBACrC,CAAC,CAAC,SAAS,CAAC,UAAU;gBACtB,CAAC,CAAC,SAAS;YACf,YAAY,EAAE,OAAO,SAAS,KAAK,QAAQ;gBACvC,CAAC,CAAC,SAAS,CAAC,GAAG;gBACf,CAAC,CAAC,SAAS;YACf,YAAY,EAAE,OAAO,SAAS,KAAK,QAAQ;gBACvC,CAAC,CAAC,SAAS,CAAC,GAAG;gBACf,CAAC,CAAC,SAAS;YACf,4BAA4B;SAC/B,CAAC,CAAC;QAEH,MAAM,wBAAwB,GAAG,OAAO,SAAS,KAAK,QAAQ;YAC1D,CAAC,SAAS,CAAC,GAAG,IAAI,IAAI,IAAI,SAAS,CAAC,GAAG,IAAI,IAAI,IAAI,SAAS,CAAC,UAAU,EAAE,WAAW,IAAI,IAAI,CAAC,CAAC;QAElG,IAAI,CAAC,wBAAwB,IAAI,mBAAmB,IAAI,IAAI,IAAI,wBAAwB;YACpF,MAAM,IAAI,uBAAuB,CAAC,8DAA8D,CAAC,CAAC;QAEtG,OAAO,mBAAmB,IAAI,CAAC,CAAC;IACpC,CAAC;IAED,MAAM,IAAI,KAAK,CAAC,4BAA4B,SAAS,EAAE,CAAC,CAAC;AAC7D,CAAC;AAED,SAAS,2BAA2B,CAAC,EACjC,YAAY,EACZ,QAAQ,EACR,UAAU,EACV,YAAY,EACZ,YAAY,EACZ,4BAA4B,EAQ/B;IACG,OAAO,cAAc,CAAC;QAClB,CAAC,SAAS;YACN,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,YAAY,IAAI,CAAC,CAAC,CAAC,CAAC;YAC7D,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,YAAY,CAAC,WAAW,EAAE,YAAY,IAAI,YAAY,CAAC,WAAW,CAAC,CAAC,CAAC;YAE3G,KAAK,IAAI,MAAM,GAAG,SAAS,EAAE,MAAM,IAAI,SAAS,EAAE,MAAM,EAAE,EAAE,CAAC;gBACzD,MAAM;oBACF,SAAS,EAAE,MAAM;iBACpB,CAAC;YACN,CAAC;QACL,CAAC;QACD,KAAK,CAAC,MAAM;YACR,MAAM,kBAAkB,GAAG,2BAA2B,CAAC;gBACnD,SAAS,EAAE,MAAM,CAAC,SAAS;gBAC3B,YAAY;gBACZ,WAAW,EAAE,QAAQ;gBACrB,UAAU;gBACV,4BAA4B;aAC/B,CAAC,CAAC;YAEH,IAAI,kBAAkB,IAAI,IAAI;gBAC1B,OAAO,IAAI,CAAC;YAEhB,OAAO,mCAAmC,CAAC,EAAC,SAAS,EAAE,MAAM,CAAC,SAAS,EAAE,WAAW,EAAE,kBAAkB,CAAC,WAAW,EAAC,EAAE;gBACnH,cAAc,EAAE,YAAY,CAAC,WAAW;gBACxC,gBAAgB,EAAE,0BAA0B,CAAC,EAAC,gBAAgB,EAAE,YAAY,CAAC,gBAAgB,EAAC,CAAC;aAClG,CAAC,CAAC;QACP,CAAC;KACJ,CAAC,EAAE,SAAS,IAAI,IAAI,CAAC;AAC1B,CAAC;AAED,SAAS,mCAAmC,CAAC,EAAC,SAAS,EAAE,WAAW,EAA2C,EAAE,EAC7G,cAAc,EAAE,gBAAgB,EAGnC;IACG,SAAS,cAAc;QACnB,OAAO,WAAW,CAAC,SAAS,EAAE,CAAC;gBAC3B,KAAK,EAAE,CAAC;gBACR,MAAM,EAAE,CAAC;aACZ,EAAE;gBACC,KAAK,EAAE,CAAC;gBACR,MAAM,EAAE,EAAE;aACb,EAAE;gBACC,KAAK,EAAE,cAAc;gBACrB,MAAM,EAAE,EAAE;gBACV,GAAG,EAAE,cAAc;aACtB,CAAC,CAAC,CAAC;IACR,CAAC;IAED,SAAS,gBAAgB;QACrB,MAAM,mBAAmB,GAAG,SAAS,GAAG,cAAc,CAAC;QAEvD,OAAO,WAAW,CAAC,WAAW,EAAE,CAAC;gBAC7B,KAAK,EAAE,CAAC;gBACR,MAAM,EAAE,CAAC;aACZ,EAAE;gBACC,KAAK,EAAE,IAAI;gBACX,MAAM,EAAE,CAAC;aACZ,EAAE;gBACC,KAAK,EAAE,IAAI;gBACX,MAAM,EAAE,mBAAmB,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;aAC5C,EAAE;gBACC,KAAK,EAAE,IAAI;gBACX,MAAM,EAAE,mBAAmB,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE;aAC7C,EAAE;gBACC,KAAK,EAAE,IAAI;gBACX,MAAM,EAAE,mBAAmB,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;gBACzC,GAAG,EAAE,IAAI,CAAC,GAAG,CAAC,gBAAgB,EAAE,KAAK,CAAC;aACzC,CAAC,CAAC,CAAC;IACR,CAAC;IAED,OAAO,cAAc,EAAE,GAAG,gBAAgB,EAAE,CAAC;AACjD,CAAC;AAED,SAAS,2BAA2B,CAAC,EACjC,SAAS,EAAE,YAAY,EAAE,WAAW,EAAE,UAAU,EAAE,4BAA4B,GAAG,KAAK,EAIzF;IACG,MAAM,SAAS,GAAG,YAAY,CAAC,iCAAiC,CAAC,EAAC,SAAS,EAAC,CAAC,CAAC,OAAO,CAAC;IAEtF,IAAI,SAAS,GAAG,WAAW;QACvB,OAAO,IAAI,CAAC;IAEhB,IAAI,UAAU,IAAI,IAAI,IAAI,UAAU,CAAC,WAAW,IAAI,IAAI,EAAE,CAAC;QACvD,MAAM,WAAW,GAAG,YAAY,CAAC,mCAAmC,CAAC;YACjE,WAAW,EAAE,UAAU,CAAC,WAAW;YACnC,SAAS,EAAE,0BAA0B,CAAC,EAAC,WAAW,EAAE,UAAU,CAAC,WAAW,EAAE,SAAS,EAAE,CAAC,EAAC,CAAC;YAC1F,cAAc,EAAE,SAAS;YACzB,SAAS,EAAE,CAAC;YACZ,kBAAkB,EAAE,UAAU,CAAC,gBAAgB,IAAI,KAAK;YACxD,cAAc,EAAE,4BAA4B;SAC/C,CAAC,CAAC,OAAO,CAAC;QAEX,MAAM,SAAS,GAAG,SAAS,GAAG,WAAW,CAAC;QAC1C,IAAI,SAAS,GAAG,WAAW;YACvB,OAAO,IAAI,CAAC;QAEhB,OAAO;YACH,WAAW,EAAE,UAAU,CAAC,WAAW;YACnC,WAAW;YACX,SAAS;SACZ,CAAC;IACN,CAAC;IAED,MAAM,UAAU,GAAG,iCAAiC,CAAC;QACjD,SAAS;QACT,YAAY;QACZ,IAAI,EAAE,WAAW,GAAG,SAAS;QAC7B,kBAAkB,EAAE,UAAU,EAAE,gBAAgB,IAAI,KAAK;QACzD,cAAc,EAAE,4BAA4B;KAC/C,CAAC,CAAC;IAEH,IAAI,UAAU,IAAI,IAAI,IAAI,SAAS,GAAG,UAAU,CAAC,IAAI,GAAG,WAAW;QAC/D,OAAO,IAAI,CAAC;IAEhB,OAAO;QACH,WAAW,EAAE,UAAU,CAAC,WAAW;QACnC,WAAW,EAAE,UAAU,CAAC,IAAI;QAC5B,SAAS,EAAE,SAAS,GAAG,UAAU,CAAC,IAAI;KACzC,CAAC;AACN,CAAC;AAED,SAAS,iCAAiC,CAAC,EAAC,SAAS,EAAE,YAAY,EAAE,IAAI,EAAE,kBAAkB,EAAE,cAAc,EAE5G;IACG,MAAM,cAAc,GAAG,0BAA0B,CAAC,EAAC,gBAAgB,EAAE,YAAY,CAAC,gBAAgB,EAAC,CAAC,CAAC;IAErG,KAAK,IAAI,WAAW,GAAG,cAAc,EAAE,WAAW,IAAI,mCAAmC,EAAE,WAAW,EAAE,EAAE,CAAC;QACvG,MAAM,WAAW,GAAG,YAAY,CAAC,mCAAmC,CAAC;YACjE,WAAW;YACX,SAAS,EAAE,0BAA0B,CAAC,EAAC,WAAW,EAAE,SAAS,EAAE,CAAC,EAAC,CAAC;YAClE,cAAc,EAAE,SAAS;YACzB,SAAS,EAAE,CAAC;YACZ,kBAAkB;YAClB,cAAc;SACjB,CAAC,CAAC,OAAO,CAAC;QAEX,IAAI,WAAW,IAAI,IAAI;YACnB,OAAO;gBACH,WAAW;gBACX,IAAI,EAAE,WAAW;aACpB,CAAC;IACV,CAAC;IAED,OAAO,IAAI,CAAC;AAChB,CAAC"}
@@ -0,0 +1,5 @@
1
+ export declare function scoreLevels(num: number, levels: {
2
+ start: number;
3
+ end?: number;
4
+ points: number;
5
+ }[]): number;
@@ -0,0 +1,16 @@
1
+ export function scoreLevels(num, levels) {
2
+ let res = 0;
3
+ for (let i = 0; i < levels.length; i++) {
4
+ const level = levels[i];
5
+ const start = level.start;
6
+ const end = level.end ?? levels[i + 1]?.start ?? Math.max(start, num);
7
+ if (num < start)
8
+ break;
9
+ else if (num >= end)
10
+ res += level.points;
11
+ else
12
+ res += level.points * ((num - start) / (end - start));
13
+ }
14
+ return res;
15
+ }
16
+ //# sourceMappingURL=scoreLevels.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"scoreLevels.js","sourceRoot":"","sources":["../../../../src/gguf/insights/utils/scoreLevels.ts"],"names":[],"mappings":"AAAA,MAAM,UAAU,WAAW,CAAC,GAAW,EAAE,MAAyD;IAC9F,IAAI,GAAG,GAAG,CAAC,CAAC;IAEZ,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;QACxB,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC;QAC1B,MAAM,GAAG,GAAG,KAAK,CAAC,GAAG,IAAI,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,IAAI,IAAI,CAAC,GAAG,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;QAEtE,IAAI,GAAG,GAAG,KAAK;YACX,MAAM;aACL,IAAI,GAAG,IAAI,GAAG;YACf,GAAG,IAAI,KAAK,CAAC,MAAM,CAAC;;YAEpB,GAAG,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,GAAG,KAAK,CAAC,GAAG,CAAC,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC;IAC9D,CAAC;IAED,OAAO,GAAG,CAAC;AACf,CAAC"}