@realtimex/node-llama-cpp 0.1.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (876) hide show
  1. package/LICENSE +21 -0
  2. package/README.md +108 -0
  3. package/dist/ChatWrapper.d.ts +27 -0
  4. package/dist/ChatWrapper.js +233 -0
  5. package/dist/ChatWrapper.js.map +1 -0
  6. package/dist/apiDocsIndex.d.ts +1 -0
  7. package/dist/apiDocsIndex.js +7 -0
  8. package/dist/apiDocsIndex.js.map +1 -0
  9. package/dist/bindings/AddonTypes.d.ts +203 -0
  10. package/dist/bindings/AddonTypes.js +2 -0
  11. package/dist/bindings/AddonTypes.js.map +1 -0
  12. package/dist/bindings/Llama.d.ts +104 -0
  13. package/dist/bindings/Llama.js +570 -0
  14. package/dist/bindings/Llama.js.map +1 -0
  15. package/dist/bindings/consts.d.ts +2 -0
  16. package/dist/bindings/consts.js +13 -0
  17. package/dist/bindings/consts.js.map +1 -0
  18. package/dist/bindings/getLlama.d.ts +297 -0
  19. package/dist/bindings/getLlama.js +569 -0
  20. package/dist/bindings/getLlama.js.map +1 -0
  21. package/dist/bindings/types.d.ts +72 -0
  22. package/dist/bindings/types.js +105 -0
  23. package/dist/bindings/types.js.map +1 -0
  24. package/dist/bindings/utils/MemoryOrchestrator.d.ts +23 -0
  25. package/dist/bindings/utils/MemoryOrchestrator.js +50 -0
  26. package/dist/bindings/utils/MemoryOrchestrator.js.map +1 -0
  27. package/dist/bindings/utils/NoBinaryFoundError.d.ts +2 -0
  28. package/dist/bindings/utils/NoBinaryFoundError.js +7 -0
  29. package/dist/bindings/utils/NoBinaryFoundError.js.map +1 -0
  30. package/dist/bindings/utils/asyncEvery.d.ts +5 -0
  31. package/dist/bindings/utils/asyncEvery.js +15 -0
  32. package/dist/bindings/utils/asyncEvery.js.map +1 -0
  33. package/dist/bindings/utils/asyncSome.d.ts +5 -0
  34. package/dist/bindings/utils/asyncSome.js +29 -0
  35. package/dist/bindings/utils/asyncSome.js.map +1 -0
  36. package/dist/bindings/utils/binariesGithubRelease.d.ts +6 -0
  37. package/dist/bindings/utils/binariesGithubRelease.js +15 -0
  38. package/dist/bindings/utils/binariesGithubRelease.js.map +1 -0
  39. package/dist/bindings/utils/clearAllLocalBuilds.d.ts +1 -0
  40. package/dist/bindings/utils/clearAllLocalBuilds.js +47 -0
  41. package/dist/bindings/utils/clearAllLocalBuilds.js.map +1 -0
  42. package/dist/bindings/utils/cloneLlamaCppRepo.d.ts +11 -0
  43. package/dist/bindings/utils/cloneLlamaCppRepo.js +166 -0
  44. package/dist/bindings/utils/cloneLlamaCppRepo.js.map +1 -0
  45. package/dist/bindings/utils/compileLLamaCpp.d.ts +22 -0
  46. package/dist/bindings/utils/compileLLamaCpp.js +526 -0
  47. package/dist/bindings/utils/compileLLamaCpp.js.map +1 -0
  48. package/dist/bindings/utils/detectAvailableComputeLayers.d.ts +18 -0
  49. package/dist/bindings/utils/detectAvailableComputeLayers.js +311 -0
  50. package/dist/bindings/utils/detectAvailableComputeLayers.js.map +1 -0
  51. package/dist/bindings/utils/detectBuildTools.d.ts +14 -0
  52. package/dist/bindings/utils/detectBuildTools.js +149 -0
  53. package/dist/bindings/utils/detectBuildTools.js.map +1 -0
  54. package/dist/bindings/utils/detectGlibc.d.ts +4 -0
  55. package/dist/bindings/utils/detectGlibc.js +74 -0
  56. package/dist/bindings/utils/detectGlibc.js.map +1 -0
  57. package/dist/bindings/utils/disposeBeforeExit.d.ts +8 -0
  58. package/dist/bindings/utils/disposeBeforeExit.js +36 -0
  59. package/dist/bindings/utils/disposeBeforeExit.js.map +1 -0
  60. package/dist/bindings/utils/getBestComputeLayersAvailable.d.ts +9 -0
  61. package/dist/bindings/utils/getBestComputeLayersAvailable.js +29 -0
  62. package/dist/bindings/utils/getBestComputeLayersAvailable.js.map +1 -0
  63. package/dist/bindings/utils/getBuildFolderNameForBuildOptions.d.ts +6 -0
  64. package/dist/bindings/utils/getBuildFolderNameForBuildOptions.js +105 -0
  65. package/dist/bindings/utils/getBuildFolderNameForBuildOptions.js.map +1 -0
  66. package/dist/bindings/utils/getCanUsePrebuiltBinaries.d.ts +1 -0
  67. package/dist/bindings/utils/getCanUsePrebuiltBinaries.js +8 -0
  68. package/dist/bindings/utils/getCanUsePrebuiltBinaries.js.map +1 -0
  69. package/dist/bindings/utils/getExampleUsageCodeOfGetLlama.d.ts +2 -0
  70. package/dist/bindings/utils/getExampleUsageCodeOfGetLlama.js +21 -0
  71. package/dist/bindings/utils/getExampleUsageCodeOfGetLlama.js.map +1 -0
  72. package/dist/bindings/utils/getGpuTypesToUseForOption.d.ts +12 -0
  73. package/dist/bindings/utils/getGpuTypesToUseForOption.js +39 -0
  74. package/dist/bindings/utils/getGpuTypesToUseForOption.js.map +1 -0
  75. package/dist/bindings/utils/getLinuxDistroInfo.d.ts +9 -0
  76. package/dist/bindings/utils/getLinuxDistroInfo.js +46 -0
  77. package/dist/bindings/utils/getLinuxDistroInfo.js.map +1 -0
  78. package/dist/bindings/utils/getLlamaGpuTypes.d.ts +13 -0
  79. package/dist/bindings/utils/getLlamaGpuTypes.js +36 -0
  80. package/dist/bindings/utils/getLlamaGpuTypes.js.map +1 -0
  81. package/dist/bindings/utils/getLlamaWithoutBackend.d.ts +5 -0
  82. package/dist/bindings/utils/getLlamaWithoutBackend.js +40 -0
  83. package/dist/bindings/utils/getLlamaWithoutBackend.js.map +1 -0
  84. package/dist/bindings/utils/getPlatform.d.ts +2 -0
  85. package/dist/bindings/utils/getPlatform.js +15 -0
  86. package/dist/bindings/utils/getPlatform.js.map +1 -0
  87. package/dist/bindings/utils/getPlatformInfo.d.ts +5 -0
  88. package/dist/bindings/utils/getPlatformInfo.js +28 -0
  89. package/dist/bindings/utils/getPlatformInfo.js.map +1 -0
  90. package/dist/bindings/utils/hasBuildingFromSourceDependenciesInstalled.d.ts +3 -0
  91. package/dist/bindings/utils/hasBuildingFromSourceDependenciesInstalled.js +27 -0
  92. package/dist/bindings/utils/hasBuildingFromSourceDependenciesInstalled.js.map +1 -0
  93. package/dist/bindings/utils/hasFileInPath.d.ts +2 -0
  94. package/dist/bindings/utils/hasFileInPath.js +34 -0
  95. package/dist/bindings/utils/hasFileInPath.js.map +1 -0
  96. package/dist/bindings/utils/lastBuildInfo.d.ts +6 -0
  97. package/dist/bindings/utils/lastBuildInfo.js +17 -0
  98. package/dist/bindings/utils/lastBuildInfo.js.map +1 -0
  99. package/dist/bindings/utils/logBinaryUsageExampleToConsole.d.ts +2 -0
  100. package/dist/bindings/utils/logBinaryUsageExampleToConsole.js +22 -0
  101. package/dist/bindings/utils/logBinaryUsageExampleToConsole.js.map +1 -0
  102. package/dist/bindings/utils/logDistroInstallInstruction.d.ts +14 -0
  103. package/dist/bindings/utils/logDistroInstallInstruction.js +48 -0
  104. package/dist/bindings/utils/logDistroInstallInstruction.js.map +1 -0
  105. package/dist/bindings/utils/resolveActualBindingBinaryPath.d.ts +1 -0
  106. package/dist/bindings/utils/resolveActualBindingBinaryPath.js +18 -0
  107. package/dist/bindings/utils/resolveActualBindingBinaryPath.js.map +1 -0
  108. package/dist/bindings/utils/resolveCustomCmakeOptions.d.ts +1 -0
  109. package/dist/bindings/utils/resolveCustomCmakeOptions.js +43 -0
  110. package/dist/bindings/utils/resolveCustomCmakeOptions.js.map +1 -0
  111. package/dist/bindings/utils/testBindingBinary.d.ts +2 -0
  112. package/dist/bindings/utils/testBindingBinary.js +269 -0
  113. package/dist/bindings/utils/testBindingBinary.js.map +1 -0
  114. package/dist/bindings/utils/testCmakeBinary.d.ts +5 -0
  115. package/dist/bindings/utils/testCmakeBinary.js +32 -0
  116. package/dist/bindings/utils/testCmakeBinary.js.map +1 -0
  117. package/dist/chatWrappers/AlpacaChatWrapper.d.ts +17 -0
  118. package/dist/chatWrappers/AlpacaChatWrapper.js +33 -0
  119. package/dist/chatWrappers/AlpacaChatWrapper.js.map +1 -0
  120. package/dist/chatWrappers/ChatMLChatWrapper.d.ts +6 -0
  121. package/dist/chatWrappers/ChatMLChatWrapper.js +85 -0
  122. package/dist/chatWrappers/ChatMLChatWrapper.js.map +1 -0
  123. package/dist/chatWrappers/DeepSeekChatWrapper.d.ts +37 -0
  124. package/dist/chatWrappers/DeepSeekChatWrapper.js +294 -0
  125. package/dist/chatWrappers/DeepSeekChatWrapper.js.map +1 -0
  126. package/dist/chatWrappers/EmptyChatWrapper.d.ts +4 -0
  127. package/dist/chatWrappers/EmptyChatWrapper.js +5 -0
  128. package/dist/chatWrappers/EmptyChatWrapper.js.map +1 -0
  129. package/dist/chatWrappers/FalconChatWrapper.d.ts +19 -0
  130. package/dist/chatWrappers/FalconChatWrapper.js +126 -0
  131. package/dist/chatWrappers/FalconChatWrapper.js.map +1 -0
  132. package/dist/chatWrappers/FunctionaryChatWrapper.d.ts +17 -0
  133. package/dist/chatWrappers/FunctionaryChatWrapper.js +622 -0
  134. package/dist/chatWrappers/FunctionaryChatWrapper.js.map +1 -0
  135. package/dist/chatWrappers/GemmaChatWrapper.d.ts +7 -0
  136. package/dist/chatWrappers/GemmaChatWrapper.js +96 -0
  137. package/dist/chatWrappers/GemmaChatWrapper.js.map +1 -0
  138. package/dist/chatWrappers/GeneralChatWrapper.d.ts +19 -0
  139. package/dist/chatWrappers/GeneralChatWrapper.js +140 -0
  140. package/dist/chatWrappers/GeneralChatWrapper.js.map +1 -0
  141. package/dist/chatWrappers/HarmonyChatWrapper.d.ts +78 -0
  142. package/dist/chatWrappers/HarmonyChatWrapper.js +539 -0
  143. package/dist/chatWrappers/HarmonyChatWrapper.js.map +1 -0
  144. package/dist/chatWrappers/Llama2ChatWrapper.d.ts +12 -0
  145. package/dist/chatWrappers/Llama2ChatWrapper.js +95 -0
  146. package/dist/chatWrappers/Llama2ChatWrapper.js.map +1 -0
  147. package/dist/chatWrappers/Llama3ChatWrapper.d.ts +16 -0
  148. package/dist/chatWrappers/Llama3ChatWrapper.js +173 -0
  149. package/dist/chatWrappers/Llama3ChatWrapper.js.map +1 -0
  150. package/dist/chatWrappers/Llama3_1ChatWrapper.d.ts +32 -0
  151. package/dist/chatWrappers/Llama3_1ChatWrapper.js +290 -0
  152. package/dist/chatWrappers/Llama3_1ChatWrapper.js.map +1 -0
  153. package/dist/chatWrappers/Llama3_2LightweightChatWrapper.d.ts +35 -0
  154. package/dist/chatWrappers/Llama3_2LightweightChatWrapper.js +264 -0
  155. package/dist/chatWrappers/Llama3_2LightweightChatWrapper.js.map +1 -0
  156. package/dist/chatWrappers/MistralChatWrapper.d.ts +16 -0
  157. package/dist/chatWrappers/MistralChatWrapper.js +180 -0
  158. package/dist/chatWrappers/MistralChatWrapper.js.map +1 -0
  159. package/dist/chatWrappers/QwenChatWrapper.d.ts +36 -0
  160. package/dist/chatWrappers/QwenChatWrapper.js +344 -0
  161. package/dist/chatWrappers/QwenChatWrapper.js.map +1 -0
  162. package/dist/chatWrappers/SeedChatWrapper.d.ts +25 -0
  163. package/dist/chatWrappers/SeedChatWrapper.js +183 -0
  164. package/dist/chatWrappers/SeedChatWrapper.js.map +1 -0
  165. package/dist/chatWrappers/generic/JinjaTemplateChatWrapper.d.ts +138 -0
  166. package/dist/chatWrappers/generic/JinjaTemplateChatWrapper.js +665 -0
  167. package/dist/chatWrappers/generic/JinjaTemplateChatWrapper.js.map +1 -0
  168. package/dist/chatWrappers/generic/TemplateChatWrapper.d.ts +76 -0
  169. package/dist/chatWrappers/generic/TemplateChatWrapper.js +212 -0
  170. package/dist/chatWrappers/generic/TemplateChatWrapper.js.map +1 -0
  171. package/dist/chatWrappers/generic/utils/UniqueIdGenerator.d.ts +7 -0
  172. package/dist/chatWrappers/generic/utils/UniqueIdGenerator.js +30 -0
  173. package/dist/chatWrappers/generic/utils/UniqueIdGenerator.js.map +1 -0
  174. package/dist/chatWrappers/generic/utils/chatHistoryFunctionCallMessageTemplate.d.ts +24 -0
  175. package/dist/chatWrappers/generic/utils/chatHistoryFunctionCallMessageTemplate.js +45 -0
  176. package/dist/chatWrappers/generic/utils/chatHistoryFunctionCallMessageTemplate.js.map +1 -0
  177. package/dist/chatWrappers/generic/utils/extractFunctionCallSettingsFromJinjaTemplate.d.ts +25 -0
  178. package/dist/chatWrappers/generic/utils/extractFunctionCallSettingsFromJinjaTemplate.js +690 -0
  179. package/dist/chatWrappers/generic/utils/extractFunctionCallSettingsFromJinjaTemplate.js.map +1 -0
  180. package/dist/chatWrappers/generic/utils/extractSegmentSettingsFromTokenizerAndChatTemplate.d.ts +2 -0
  181. package/dist/chatWrappers/generic/utils/extractSegmentSettingsFromTokenizerAndChatTemplate.js +66 -0
  182. package/dist/chatWrappers/generic/utils/extractSegmentSettingsFromTokenizerAndChatTemplate.js.map +1 -0
  183. package/dist/chatWrappers/generic/utils/getFirstValidResult.d.ts +6 -0
  184. package/dist/chatWrappers/generic/utils/getFirstValidResult.js +19 -0
  185. package/dist/chatWrappers/generic/utils/getFirstValidResult.js.map +1 -0
  186. package/dist/chatWrappers/generic/utils/squashChatHistoryItems.d.ts +2 -0
  187. package/dist/chatWrappers/generic/utils/squashChatHistoryItems.js +35 -0
  188. package/dist/chatWrappers/generic/utils/squashChatHistoryItems.js.map +1 -0
  189. package/dist/chatWrappers/generic/utils/templateSegmentOptionsToChatWrapperSettings.d.ts +22 -0
  190. package/dist/chatWrappers/generic/utils/templateSegmentOptionsToChatWrapperSettings.js +28 -0
  191. package/dist/chatWrappers/generic/utils/templateSegmentOptionsToChatWrapperSettings.js.map +1 -0
  192. package/dist/chatWrappers/utils/ChatModelFunctionsDocumentationGenerator.d.ts +76 -0
  193. package/dist/chatWrappers/utils/ChatModelFunctionsDocumentationGenerator.js +177 -0
  194. package/dist/chatWrappers/utils/ChatModelFunctionsDocumentationGenerator.js.map +1 -0
  195. package/dist/chatWrappers/utils/chunkChatItems.d.ts +10 -0
  196. package/dist/chatWrappers/utils/chunkChatItems.js +44 -0
  197. package/dist/chatWrappers/utils/chunkChatItems.js.map +1 -0
  198. package/dist/chatWrappers/utils/getModelLinageNames.d.ts +2 -0
  199. package/dist/chatWrappers/utils/getModelLinageNames.js +18 -0
  200. package/dist/chatWrappers/utils/getModelLinageNames.js.map +1 -0
  201. package/dist/chatWrappers/utils/isJinjaTemplateEquivalentToSpecializedChatWrapper.d.ts +4 -0
  202. package/dist/chatWrappers/utils/isJinjaTemplateEquivalentToSpecializedChatWrapper.js +394 -0
  203. package/dist/chatWrappers/utils/isJinjaTemplateEquivalentToSpecializedChatWrapper.js.map +1 -0
  204. package/dist/chatWrappers/utils/isLlama3_2LightweightModel.d.ts +2 -0
  205. package/dist/chatWrappers/utils/isLlama3_2LightweightModel.js +9 -0
  206. package/dist/chatWrappers/utils/isLlama3_2LightweightModel.js.map +1 -0
  207. package/dist/chatWrappers/utils/jsonDumps.d.ts +7 -0
  208. package/dist/chatWrappers/utils/jsonDumps.js +18 -0
  209. package/dist/chatWrappers/utils/jsonDumps.js.map +1 -0
  210. package/dist/chatWrappers/utils/resolveChatWrapper.d.ts +148 -0
  211. package/dist/chatWrappers/utils/resolveChatWrapper.js +325 -0
  212. package/dist/chatWrappers/utils/resolveChatWrapper.js.map +1 -0
  213. package/dist/cli/cli.d.ts +2 -0
  214. package/dist/cli/cli.js +45 -0
  215. package/dist/cli/cli.js.map +1 -0
  216. package/dist/cli/commands/ChatCommand.d.ts +59 -0
  217. package/dist/cli/commands/ChatCommand.js +856 -0
  218. package/dist/cli/commands/ChatCommand.js.map +1 -0
  219. package/dist/cli/commands/CompleteCommand.d.ts +47 -0
  220. package/dist/cli/commands/CompleteCommand.js +658 -0
  221. package/dist/cli/commands/CompleteCommand.js.map +1 -0
  222. package/dist/cli/commands/DebugCommand.d.ts +7 -0
  223. package/dist/cli/commands/DebugCommand.js +55 -0
  224. package/dist/cli/commands/DebugCommand.js.map +1 -0
  225. package/dist/cli/commands/InfillCommand.d.ts +49 -0
  226. package/dist/cli/commands/InfillCommand.js +693 -0
  227. package/dist/cli/commands/InfillCommand.js.map +1 -0
  228. package/dist/cli/commands/InitCommand.d.ts +12 -0
  229. package/dist/cli/commands/InitCommand.js +230 -0
  230. package/dist/cli/commands/InitCommand.js.map +1 -0
  231. package/dist/cli/commands/OnPostInstallCommand.d.ts +4 -0
  232. package/dist/cli/commands/OnPostInstallCommand.js +62 -0
  233. package/dist/cli/commands/OnPostInstallCommand.js.map +1 -0
  234. package/dist/cli/commands/PullCommand.d.ts +13 -0
  235. package/dist/cli/commands/PullCommand.js +158 -0
  236. package/dist/cli/commands/PullCommand.js.map +1 -0
  237. package/dist/cli/commands/inspect/InspectCommand.d.ts +4 -0
  238. package/dist/cli/commands/inspect/InspectCommand.js +21 -0
  239. package/dist/cli/commands/inspect/InspectCommand.js.map +1 -0
  240. package/dist/cli/commands/inspect/commands/InspectEstimateCommand.d.ts +17 -0
  241. package/dist/cli/commands/inspect/commands/InspectEstimateCommand.js +275 -0
  242. package/dist/cli/commands/inspect/commands/InspectEstimateCommand.js.map +1 -0
  243. package/dist/cli/commands/inspect/commands/InspectGgufCommand.d.ts +13 -0
  244. package/dist/cli/commands/inspect/commands/InspectGgufCommand.js +230 -0
  245. package/dist/cli/commands/inspect/commands/InspectGgufCommand.js.map +1 -0
  246. package/dist/cli/commands/inspect/commands/InspectGpuCommand.d.ts +4 -0
  247. package/dist/cli/commands/inspect/commands/InspectGpuCommand.js +296 -0
  248. package/dist/cli/commands/inspect/commands/InspectGpuCommand.js.map +1 -0
  249. package/dist/cli/commands/inspect/commands/InspectMeasureCommand.d.ts +26 -0
  250. package/dist/cli/commands/inspect/commands/InspectMeasureCommand.js +869 -0
  251. package/dist/cli/commands/inspect/commands/InspectMeasureCommand.js.map +1 -0
  252. package/dist/cli/commands/source/SourceCommand.d.ts +4 -0
  253. package/dist/cli/commands/source/SourceCommand.js +19 -0
  254. package/dist/cli/commands/source/SourceCommand.js.map +1 -0
  255. package/dist/cli/commands/source/commands/BuildCommand.d.ts +16 -0
  256. package/dist/cli/commands/source/commands/BuildCommand.js +148 -0
  257. package/dist/cli/commands/source/commands/BuildCommand.js.map +1 -0
  258. package/dist/cli/commands/source/commands/ClearCommand.d.ts +7 -0
  259. package/dist/cli/commands/source/commands/ClearCommand.js +54 -0
  260. package/dist/cli/commands/source/commands/ClearCommand.js.map +1 -0
  261. package/dist/cli/commands/source/commands/DownloadCommand.d.ts +16 -0
  262. package/dist/cli/commands/source/commands/DownloadCommand.js +219 -0
  263. package/dist/cli/commands/source/commands/DownloadCommand.js.map +1 -0
  264. package/dist/cli/projectTemplates.d.ts +7 -0
  265. package/dist/cli/projectTemplates.js +10 -0
  266. package/dist/cli/projectTemplates.js.map +1 -0
  267. package/dist/cli/recommendedModels.d.ts +2 -0
  268. package/dist/cli/recommendedModels.js +428 -0
  269. package/dist/cli/recommendedModels.js.map +1 -0
  270. package/dist/cli/startCreateCli.d.ts +2 -0
  271. package/dist/cli/startCreateCli.js +26 -0
  272. package/dist/cli/startCreateCli.js.map +1 -0
  273. package/dist/cli/utils/ConsoleInteraction.d.ts +22 -0
  274. package/dist/cli/utils/ConsoleInteraction.js +122 -0
  275. package/dist/cli/utils/ConsoleInteraction.js.map +1 -0
  276. package/dist/cli/utils/ConsoleTable.d.ts +24 -0
  277. package/dist/cli/utils/ConsoleTable.js +90 -0
  278. package/dist/cli/utils/ConsoleTable.js.map +1 -0
  279. package/dist/cli/utils/basicChooseFromListConsoleInteraction.d.ts +13 -0
  280. package/dist/cli/utils/basicChooseFromListConsoleInteraction.js +111 -0
  281. package/dist/cli/utils/basicChooseFromListConsoleInteraction.js.map +1 -0
  282. package/dist/cli/utils/consolePromptQuestion.d.ts +6 -0
  283. package/dist/cli/utils/consolePromptQuestion.js +81 -0
  284. package/dist/cli/utils/consolePromptQuestion.js.map +1 -0
  285. package/dist/cli/utils/getReadablePath.d.ts +1 -0
  286. package/dist/cli/utils/getReadablePath.js +14 -0
  287. package/dist/cli/utils/getReadablePath.js.map +1 -0
  288. package/dist/cli/utils/interactivelyAskForModel.d.ts +13 -0
  289. package/dist/cli/utils/interactivelyAskForModel.js +485 -0
  290. package/dist/cli/utils/interactivelyAskForModel.js.map +1 -0
  291. package/dist/cli/utils/isRunningUnderRosetta.d.ts +1 -0
  292. package/dist/cli/utils/isRunningUnderRosetta.js +20 -0
  293. package/dist/cli/utils/isRunningUnderRosetta.js.map +1 -0
  294. package/dist/cli/utils/logUsedGpuTypeOption.d.ts +2 -0
  295. package/dist/cli/utils/logUsedGpuTypeOption.js +9 -0
  296. package/dist/cli/utils/logUsedGpuTypeOption.js.map +1 -0
  297. package/dist/cli/utils/packageJsonConfig.d.ts +6 -0
  298. package/dist/cli/utils/packageJsonConfig.js +51 -0
  299. package/dist/cli/utils/packageJsonConfig.js.map +1 -0
  300. package/dist/cli/utils/packageManager.d.ts +1 -0
  301. package/dist/cli/utils/packageManager.js +15 -0
  302. package/dist/cli/utils/packageManager.js.map +1 -0
  303. package/dist/cli/utils/parseXtcArg.d.ts +5 -0
  304. package/dist/cli/utils/parseXtcArg.js +16 -0
  305. package/dist/cli/utils/parseXtcArg.js.map +1 -0
  306. package/dist/cli/utils/printCommonInfoLines.d.ts +12 -0
  307. package/dist/cli/utils/printCommonInfoLines.js +163 -0
  308. package/dist/cli/utils/printCommonInfoLines.js.map +1 -0
  309. package/dist/cli/utils/printInfoLine.d.ts +12 -0
  310. package/dist/cli/utils/printInfoLine.js +54 -0
  311. package/dist/cli/utils/printInfoLine.js.map +1 -0
  312. package/dist/cli/utils/printModelDestination.d.ts +2 -0
  313. package/dist/cli/utils/printModelDestination.js +11 -0
  314. package/dist/cli/utils/printModelDestination.js.map +1 -0
  315. package/dist/cli/utils/projectTemplates.d.ts +19 -0
  316. package/dist/cli/utils/projectTemplates.js +47 -0
  317. package/dist/cli/utils/projectTemplates.js.map +1 -0
  318. package/dist/cli/utils/renderModelCompatibilityPercentageWithColors.d.ts +6 -0
  319. package/dist/cli/utils/renderModelCompatibilityPercentageWithColors.js +14 -0
  320. package/dist/cli/utils/renderModelCompatibilityPercentageWithColors.js.map +1 -0
  321. package/dist/cli/utils/resolveCommandGgufPath.d.ts +19 -0
  322. package/dist/cli/utils/resolveCommandGgufPath.js +123 -0
  323. package/dist/cli/utils/resolveCommandGgufPath.js.map +1 -0
  324. package/dist/cli/utils/resolveHeaderFlag.d.ts +1 -0
  325. package/dist/cli/utils/resolveHeaderFlag.js +21 -0
  326. package/dist/cli/utils/resolveHeaderFlag.js.map +1 -0
  327. package/dist/cli/utils/resolveModelRecommendationFileOptions.d.ts +14 -0
  328. package/dist/cli/utils/resolveModelRecommendationFileOptions.js +12 -0
  329. package/dist/cli/utils/resolveModelRecommendationFileOptions.js.map +1 -0
  330. package/dist/cli/utils/resolveNpmrcConfig.d.ts +18 -0
  331. package/dist/cli/utils/resolveNpmrcConfig.js +129 -0
  332. package/dist/cli/utils/resolveNpmrcConfig.js.map +1 -0
  333. package/dist/cli/utils/splitAnsiToLines.d.ts +1 -0
  334. package/dist/cli/utils/splitAnsiToLines.js +32 -0
  335. package/dist/cli/utils/splitAnsiToLines.js.map +1 -0
  336. package/dist/cli/utils/toBytes.d.ts +1 -0
  337. package/dist/cli/utils/toBytes.js +5 -0
  338. package/dist/cli/utils/toBytes.js.map +1 -0
  339. package/dist/cli/utils/withCliCommandDescriptionDocsUrl.d.ts +2 -0
  340. package/dist/cli/utils/withCliCommandDescriptionDocsUrl.js +23 -0
  341. package/dist/cli/utils/withCliCommandDescriptionDocsUrl.js.map +1 -0
  342. package/dist/commands.d.ts +6 -0
  343. package/dist/commands.js +9 -0
  344. package/dist/commands.js.map +1 -0
  345. package/dist/config.d.ts +79 -0
  346. package/dist/config.js +127 -0
  347. package/dist/config.js.map +1 -0
  348. package/dist/consts.d.ts +4 -0
  349. package/dist/consts.js +11 -0
  350. package/dist/consts.js.map +1 -0
  351. package/dist/evaluator/LlamaChat/LlamaChat.d.ts +500 -0
  352. package/dist/evaluator/LlamaChat/LlamaChat.js +2696 -0
  353. package/dist/evaluator/LlamaChat/LlamaChat.js.map +1 -0
  354. package/dist/evaluator/LlamaChat/utils/FunctionCallNameGrammar.d.ts +11 -0
  355. package/dist/evaluator/LlamaChat/utils/FunctionCallNameGrammar.js +55 -0
  356. package/dist/evaluator/LlamaChat/utils/FunctionCallNameGrammar.js.map +1 -0
  357. package/dist/evaluator/LlamaChat/utils/FunctionCallParamsGrammar.d.ts +16 -0
  358. package/dist/evaluator/LlamaChat/utils/FunctionCallParamsGrammar.js +45 -0
  359. package/dist/evaluator/LlamaChat/utils/FunctionCallParamsGrammar.js.map +1 -0
  360. package/dist/evaluator/LlamaChat/utils/LlamaFunctionCallValidationError.d.ts +8 -0
  361. package/dist/evaluator/LlamaChat/utils/LlamaFunctionCallValidationError.js +12 -0
  362. package/dist/evaluator/LlamaChat/utils/LlamaFunctionCallValidationError.js.map +1 -0
  363. package/dist/evaluator/LlamaChat/utils/contextShiftStrategies/eraseFirstResponseAndKeepFirstSystemChatContextShiftStrategy.d.ts +16 -0
  364. package/dist/evaluator/LlamaChat/utils/contextShiftStrategies/eraseFirstResponseAndKeepFirstSystemChatContextShiftStrategy.js +260 -0
  365. package/dist/evaluator/LlamaChat/utils/contextShiftStrategies/eraseFirstResponseAndKeepFirstSystemChatContextShiftStrategy.js.map +1 -0
  366. package/dist/evaluator/LlamaChatSession/LlamaChatSession.d.ts +520 -0
  367. package/dist/evaluator/LlamaChatSession/LlamaChatSession.js +628 -0
  368. package/dist/evaluator/LlamaChatSession/LlamaChatSession.js.map +1 -0
  369. package/dist/evaluator/LlamaChatSession/utils/LlamaChatSessionPromptCompletionEngine.d.ts +45 -0
  370. package/dist/evaluator/LlamaChatSession/utils/LlamaChatSessionPromptCompletionEngine.js +191 -0
  371. package/dist/evaluator/LlamaChatSession/utils/LlamaChatSessionPromptCompletionEngine.js.map +1 -0
  372. package/dist/evaluator/LlamaChatSession/utils/defineChatSessionFunction.d.ts +15 -0
  373. package/dist/evaluator/LlamaChatSession/utils/defineChatSessionFunction.js +16 -0
  374. package/dist/evaluator/LlamaChatSession/utils/defineChatSessionFunction.js.map +1 -0
  375. package/dist/evaluator/LlamaCompletion.d.ts +219 -0
  376. package/dist/evaluator/LlamaCompletion.js +498 -0
  377. package/dist/evaluator/LlamaCompletion.js.map +1 -0
  378. package/dist/evaluator/LlamaContext/LlamaContext.d.ts +336 -0
  379. package/dist/evaluator/LlamaContext/LlamaContext.js +1919 -0
  380. package/dist/evaluator/LlamaContext/LlamaContext.js.map +1 -0
  381. package/dist/evaluator/LlamaContext/LlamaContextSequenceCheckpoints.d.ts +27 -0
  382. package/dist/evaluator/LlamaContext/LlamaContextSequenceCheckpoints.js +130 -0
  383. package/dist/evaluator/LlamaContext/LlamaContextSequenceCheckpoints.js.map +1 -0
  384. package/dist/evaluator/LlamaContext/LlamaSampler.d.ts +1 -0
  385. package/dist/evaluator/LlamaContext/LlamaSampler.js +39 -0
  386. package/dist/evaluator/LlamaContext/LlamaSampler.js.map +1 -0
  387. package/dist/evaluator/LlamaContext/TokenPredictor.d.ts +55 -0
  388. package/dist/evaluator/LlamaContext/TokenPredictor.js +20 -0
  389. package/dist/evaluator/LlamaContext/TokenPredictor.js.map +1 -0
  390. package/dist/evaluator/LlamaContext/tokenPredictors/DraftSequenceTokenPredictor.d.ts +56 -0
  391. package/dist/evaluator/LlamaContext/tokenPredictors/DraftSequenceTokenPredictor.js +266 -0
  392. package/dist/evaluator/LlamaContext/tokenPredictors/DraftSequenceTokenPredictor.js.map +1 -0
  393. package/dist/evaluator/LlamaContext/tokenPredictors/InputLookupTokenPredictor.d.ts +58 -0
  394. package/dist/evaluator/LlamaContext/tokenPredictors/InputLookupTokenPredictor.js +138 -0
  395. package/dist/evaluator/LlamaContext/tokenPredictors/InputLookupTokenPredictor.js.map +1 -0
  396. package/dist/evaluator/LlamaContext/types.d.ts +602 -0
  397. package/dist/evaluator/LlamaContext/types.js +2 -0
  398. package/dist/evaluator/LlamaContext/types.js.map +1 -0
  399. package/dist/evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies/firstInFirstOutStrategy.d.ts +5 -0
  400. package/dist/evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies/firstInFirstOutStrategy.js +16 -0
  401. package/dist/evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies/firstInFirstOutStrategy.js.map +1 -0
  402. package/dist/evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies/maximumParallelismStrategy.d.ts +5 -0
  403. package/dist/evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies/maximumParallelismStrategy.js +42 -0
  404. package/dist/evaluator/LlamaContext/utils/batchItemsPrioritizationStrategies/maximumParallelismStrategy.js.map +1 -0
  405. package/dist/evaluator/LlamaContext/utils/padSafeContextSize.d.ts +1 -0
  406. package/dist/evaluator/LlamaContext/utils/padSafeContextSize.js +18 -0
  407. package/dist/evaluator/LlamaContext/utils/padSafeContextSize.js.map +1 -0
  408. package/dist/evaluator/LlamaContext/utils/resolveBatchItemsPrioritizationStrategy.d.ts +2 -0
  409. package/dist/evaluator/LlamaContext/utils/resolveBatchItemsPrioritizationStrategy.js +13 -0
  410. package/dist/evaluator/LlamaContext/utils/resolveBatchItemsPrioritizationStrategy.js.map +1 -0
  411. package/dist/evaluator/LlamaEmbedding.d.ts +21 -0
  412. package/dist/evaluator/LlamaEmbedding.js +53 -0
  413. package/dist/evaluator/LlamaEmbedding.js.map +1 -0
  414. package/dist/evaluator/LlamaEmbeddingContext.d.ts +52 -0
  415. package/dist/evaluator/LlamaEmbeddingContext.js +86 -0
  416. package/dist/evaluator/LlamaEmbeddingContext.js.map +1 -0
  417. package/dist/evaluator/LlamaGrammar.d.ts +39 -0
  418. package/dist/evaluator/LlamaGrammar.js +72 -0
  419. package/dist/evaluator/LlamaGrammar.js.map +1 -0
  420. package/dist/evaluator/LlamaGrammarEvaluationState.d.ts +19 -0
  421. package/dist/evaluator/LlamaGrammarEvaluationState.js +29 -0
  422. package/dist/evaluator/LlamaGrammarEvaluationState.js.map +1 -0
  423. package/dist/evaluator/LlamaJsonSchemaGrammar.d.ts +17 -0
  424. package/dist/evaluator/LlamaJsonSchemaGrammar.js +35 -0
  425. package/dist/evaluator/LlamaJsonSchemaGrammar.js.map +1 -0
  426. package/dist/evaluator/LlamaModel/LlamaModel.d.ts +344 -0
  427. package/dist/evaluator/LlamaModel/LlamaModel.js +853 -0
  428. package/dist/evaluator/LlamaModel/LlamaModel.js.map +1 -0
  429. package/dist/evaluator/LlamaModel/utils/TokenAttributes.d.ts +29 -0
  430. package/dist/evaluator/LlamaModel/utils/TokenAttributes.js +65 -0
  431. package/dist/evaluator/LlamaModel/utils/TokenAttributes.js.map +1 -0
  432. package/dist/evaluator/LlamaRankingContext.d.ts +91 -0
  433. package/dist/evaluator/LlamaRankingContext.js +178 -0
  434. package/dist/evaluator/LlamaRankingContext.js.map +1 -0
  435. package/dist/evaluator/TokenBias.d.ts +37 -0
  436. package/dist/evaluator/TokenBias.js +68 -0
  437. package/dist/evaluator/TokenBias.js.map +1 -0
  438. package/dist/evaluator/TokenMeter.d.ts +45 -0
  439. package/dist/evaluator/TokenMeter.js +74 -0
  440. package/dist/evaluator/TokenMeter.js.map +1 -0
  441. package/dist/evaluator/utils/chunkDocument.d.ts +86 -0
  442. package/dist/evaluator/utils/chunkDocument.js +212 -0
  443. package/dist/evaluator/utils/chunkDocument.js.map +1 -0
  444. package/dist/gguf/consts.d.ts +4 -0
  445. package/dist/gguf/consts.js +12 -0
  446. package/dist/gguf/consts.js.map +1 -0
  447. package/dist/gguf/errors/InvalidGgufMagicError.d.ts +3 -0
  448. package/dist/gguf/errors/InvalidGgufMagicError.js +6 -0
  449. package/dist/gguf/errors/InvalidGgufMagicError.js.map +1 -0
  450. package/dist/gguf/errors/UnsupportedGgufValueTypeError.d.ts +4 -0
  451. package/dist/gguf/errors/UnsupportedGgufValueTypeError.js +9 -0
  452. package/dist/gguf/errors/UnsupportedGgufValueTypeError.js.map +1 -0
  453. package/dist/gguf/fileReaders/GgufFileReader.d.ts +36 -0
  454. package/dist/gguf/fileReaders/GgufFileReader.js +106 -0
  455. package/dist/gguf/fileReaders/GgufFileReader.js.map +1 -0
  456. package/dist/gguf/fileReaders/GgufFsFileReader.d.ts +16 -0
  457. package/dist/gguf/fileReaders/GgufFsFileReader.js +62 -0
  458. package/dist/gguf/fileReaders/GgufFsFileReader.js.map +1 -0
  459. package/dist/gguf/fileReaders/GgufNetworkFetchFileReader.d.ts +28 -0
  460. package/dist/gguf/fileReaders/GgufNetworkFetchFileReader.js +94 -0
  461. package/dist/gguf/fileReaders/GgufNetworkFetchFileReader.js.map +1 -0
  462. package/dist/gguf/insights/GgufInsights.d.ts +78 -0
  463. package/dist/gguf/insights/GgufInsights.js +854 -0
  464. package/dist/gguf/insights/GgufInsights.js.map +1 -0
  465. package/dist/gguf/insights/GgufInsightsConfigurationResolver.d.ts +203 -0
  466. package/dist/gguf/insights/GgufInsightsConfigurationResolver.js +284 -0
  467. package/dist/gguf/insights/GgufInsightsConfigurationResolver.js.map +1 -0
  468. package/dist/gguf/insights/GgufInsightsTokens.d.ts +5 -0
  469. package/dist/gguf/insights/GgufInsightsTokens.js +40 -0
  470. package/dist/gguf/insights/GgufInsightsTokens.js.map +1 -0
  471. package/dist/gguf/insights/utils/getRamUsageFromUnifiedVram.d.ts +5 -0
  472. package/dist/gguf/insights/utils/getRamUsageFromUnifiedVram.js +7 -0
  473. package/dist/gguf/insights/utils/getRamUsageFromUnifiedVram.js.map +1 -0
  474. package/dist/gguf/insights/utils/resolveContextContextSizeOption.d.ts +33 -0
  475. package/dist/gguf/insights/utils/resolveContextContextSizeOption.js +117 -0
  476. package/dist/gguf/insights/utils/resolveContextContextSizeOption.js.map +1 -0
  477. package/dist/gguf/insights/utils/resolveModelGpuLayersOption.d.ts +20 -0
  478. package/dist/gguf/insights/utils/resolveModelGpuLayersOption.js +251 -0
  479. package/dist/gguf/insights/utils/resolveModelGpuLayersOption.js.map +1 -0
  480. package/dist/gguf/insights/utils/scoreLevels.d.ts +5 -0
  481. package/dist/gguf/insights/utils/scoreLevels.js +16 -0
  482. package/dist/gguf/insights/utils/scoreLevels.js.map +1 -0
  483. package/dist/gguf/parser/GgufV2Parser.d.ts +20 -0
  484. package/dist/gguf/parser/GgufV2Parser.js +184 -0
  485. package/dist/gguf/parser/GgufV2Parser.js.map +1 -0
  486. package/dist/gguf/parser/GgufV3Parser.d.ts +3 -0
  487. package/dist/gguf/parser/GgufV3Parser.js +4 -0
  488. package/dist/gguf/parser/GgufV3Parser.js.map +1 -0
  489. package/dist/gguf/parser/parseGguf.d.ts +8 -0
  490. package/dist/gguf/parser/parseGguf.js +61 -0
  491. package/dist/gguf/parser/parseGguf.js.map +1 -0
  492. package/dist/gguf/readGgufFileInfo.d.ts +54 -0
  493. package/dist/gguf/readGgufFileInfo.js +82 -0
  494. package/dist/gguf/readGgufFileInfo.js.map +1 -0
  495. package/dist/gguf/types/GgufFileInfoTypes.d.ts +85 -0
  496. package/dist/gguf/types/GgufFileInfoTypes.js +18 -0
  497. package/dist/gguf/types/GgufFileInfoTypes.js.map +1 -0
  498. package/dist/gguf/types/GgufMetadataTypes.d.ts +480 -0
  499. package/dist/gguf/types/GgufMetadataTypes.js +194 -0
  500. package/dist/gguf/types/GgufMetadataTypes.js.map +1 -0
  501. package/dist/gguf/types/GgufTensorInfoTypes.d.ts +63 -0
  502. package/dist/gguf/types/GgufTensorInfoTypes.js +54 -0
  503. package/dist/gguf/types/GgufTensorInfoTypes.js.map +1 -0
  504. package/dist/gguf/utils/GgufReadOffset.d.ts +6 -0
  505. package/dist/gguf/utils/GgufReadOffset.js +18 -0
  506. package/dist/gguf/utils/GgufReadOffset.js.map +1 -0
  507. package/dist/gguf/utils/convertMetadataKeyValueRecordToNestedObject.d.ts +6 -0
  508. package/dist/gguf/utils/convertMetadataKeyValueRecordToNestedObject.js +76 -0
  509. package/dist/gguf/utils/convertMetadataKeyValueRecordToNestedObject.js.map +1 -0
  510. package/dist/gguf/utils/getGgufFileTypeName.d.ts +4 -0
  511. package/dist/gguf/utils/getGgufFileTypeName.js +13 -0
  512. package/dist/gguf/utils/getGgufFileTypeName.js.map +1 -0
  513. package/dist/gguf/utils/getGgufMetadataArchitectureData.d.ts +3 -0
  514. package/dist/gguf/utils/getGgufMetadataArchitectureData.js +4 -0
  515. package/dist/gguf/utils/getGgufMetadataArchitectureData.js.map +1 -0
  516. package/dist/gguf/utils/getGgufMetadataKeyValue.d.ts +1 -0
  517. package/dist/gguf/utils/getGgufMetadataKeyValue.js +27 -0
  518. package/dist/gguf/utils/getGgufMetadataKeyValue.js.map +1 -0
  519. package/dist/gguf/utils/ggufQuantNames.d.ts +2 -0
  520. package/dist/gguf/utils/ggufQuantNames.js +42 -0
  521. package/dist/gguf/utils/ggufQuantNames.js.map +1 -0
  522. package/dist/gguf/utils/normalizeGgufDownloadUrl.d.ts +2 -0
  523. package/dist/gguf/utils/normalizeGgufDownloadUrl.js +18 -0
  524. package/dist/gguf/utils/normalizeGgufDownloadUrl.js.map +1 -0
  525. package/dist/gguf/utils/resolveBinarySplitGgufPartUrls.d.ts +2 -0
  526. package/dist/gguf/utils/resolveBinarySplitGgufPartUrls.js +38 -0
  527. package/dist/gguf/utils/resolveBinarySplitGgufPartUrls.js.map +1 -0
  528. package/dist/gguf/utils/resolveSplitGgufParts.d.ts +7 -0
  529. package/dist/gguf/utils/resolveSplitGgufParts.js +64 -0
  530. package/dist/gguf/utils/resolveSplitGgufParts.js.map +1 -0
  531. package/dist/index.d.ts +71 -0
  532. package/dist/index.js +64 -0
  533. package/dist/index.js.map +1 -0
  534. package/dist/state.d.ts +6 -0
  535. package/dist/state.js +22 -0
  536. package/dist/state.js.map +1 -0
  537. package/dist/tsconfig.tsbuildinfo +1 -0
  538. package/dist/types.d.ts +408 -0
  539. package/dist/types.js +13 -0
  540. package/dist/types.js.map +1 -0
  541. package/dist/utils/DisposeGuard.d.ts +13 -0
  542. package/dist/utils/DisposeGuard.js +120 -0
  543. package/dist/utils/DisposeGuard.js.map +1 -0
  544. package/dist/utils/InsufficientMemoryError.d.ts +3 -0
  545. package/dist/utils/InsufficientMemoryError.js +6 -0
  546. package/dist/utils/InsufficientMemoryError.js.map +1 -0
  547. package/dist/utils/LlamaText.d.ts +92 -0
  548. package/dist/utils/LlamaText.js +527 -0
  549. package/dist/utils/LlamaText.js.map +1 -0
  550. package/dist/utils/LruCache.d.ts +12 -0
  551. package/dist/utils/LruCache.js +44 -0
  552. package/dist/utils/LruCache.js.map +1 -0
  553. package/dist/utils/OpenAIFormat.d.ts +177 -0
  554. package/dist/utils/OpenAIFormat.js +488 -0
  555. package/dist/utils/OpenAIFormat.js.map +1 -0
  556. package/dist/utils/OverridesObject.d.ts +7 -0
  557. package/dist/utils/OverridesObject.js +2 -0
  558. package/dist/utils/OverridesObject.js.map +1 -0
  559. package/dist/utils/ReplHistory.d.ts +9 -0
  560. package/dist/utils/ReplHistory.js +72 -0
  561. package/dist/utils/ReplHistory.js.map +1 -0
  562. package/dist/utils/StopGenerationDetector.d.ts +47 -0
  563. package/dist/utils/StopGenerationDetector.js +291 -0
  564. package/dist/utils/StopGenerationDetector.js.map +1 -0
  565. package/dist/utils/ThreadsSplitter.d.ts +32 -0
  566. package/dist/utils/ThreadsSplitter.js +177 -0
  567. package/dist/utils/ThreadsSplitter.js.map +1 -0
  568. package/dist/utils/TokenStreamRegulator.d.ts +38 -0
  569. package/dist/utils/TokenStreamRegulator.js +200 -0
  570. package/dist/utils/TokenStreamRegulator.js.map +1 -0
  571. package/dist/utils/UnsupportedError.d.ts +2 -0
  572. package/dist/utils/UnsupportedError.js +7 -0
  573. package/dist/utils/UnsupportedError.js.map +1 -0
  574. package/dist/utils/appendUserMessageToChatHistory.d.ts +6 -0
  575. package/dist/utils/appendUserMessageToChatHistory.js +22 -0
  576. package/dist/utils/appendUserMessageToChatHistory.js.map +1 -0
  577. package/dist/utils/clearTempFolder.d.ts +1 -0
  578. package/dist/utils/clearTempFolder.js +16 -0
  579. package/dist/utils/clearTempFolder.js.map +1 -0
  580. package/dist/utils/cmake.d.ts +10 -0
  581. package/dist/utils/cmake.js +146 -0
  582. package/dist/utils/cmake.js.map +1 -0
  583. package/dist/utils/compareTokens.d.ts +2 -0
  584. package/dist/utils/compareTokens.js +4 -0
  585. package/dist/utils/compareTokens.js.map +1 -0
  586. package/dist/utils/createModelDownloader.d.ts +262 -0
  587. package/dist/utils/createModelDownloader.js +486 -0
  588. package/dist/utils/createModelDownloader.js.map +1 -0
  589. package/dist/utils/findBestOption.d.ts +4 -0
  590. package/dist/utils/findBestOption.js +15 -0
  591. package/dist/utils/findBestOption.js.map +1 -0
  592. package/dist/utils/findCharacterRemovalCountToFitChatHistoryInContext.d.ts +20 -0
  593. package/dist/utils/findCharacterRemovalCountToFitChatHistoryInContext.js +85 -0
  594. package/dist/utils/findCharacterRemovalCountToFitChatHistoryInContext.js.map +1 -0
  595. package/dist/utils/gbnfJson/GbnfGrammarGenerator.d.ts +19 -0
  596. package/dist/utils/gbnfJson/GbnfGrammarGenerator.js +60 -0
  597. package/dist/utils/gbnfJson/GbnfGrammarGenerator.js.map +1 -0
  598. package/dist/utils/gbnfJson/GbnfTerminal.d.ts +11 -0
  599. package/dist/utils/gbnfJson/GbnfTerminal.js +54 -0
  600. package/dist/utils/gbnfJson/GbnfTerminal.js.map +1 -0
  601. package/dist/utils/gbnfJson/getGbnfGrammarForGbnfJsonSchema.d.ts +5 -0
  602. package/dist/utils/gbnfJson/getGbnfGrammarForGbnfJsonSchema.js +11 -0
  603. package/dist/utils/gbnfJson/getGbnfGrammarForGbnfJsonSchema.js.map +1 -0
  604. package/dist/utils/gbnfJson/terminals/GbnfAnyJson.d.ts +9 -0
  605. package/dist/utils/gbnfJson/terminals/GbnfAnyJson.js +53 -0
  606. package/dist/utils/gbnfJson/terminals/GbnfAnyJson.js.map +1 -0
  607. package/dist/utils/gbnfJson/terminals/GbnfArray.d.ts +18 -0
  608. package/dist/utils/gbnfJson/terminals/GbnfArray.js +83 -0
  609. package/dist/utils/gbnfJson/terminals/GbnfArray.js.map +1 -0
  610. package/dist/utils/gbnfJson/terminals/GbnfBoolean.d.ts +7 -0
  611. package/dist/utils/gbnfJson/terminals/GbnfBoolean.js +22 -0
  612. package/dist/utils/gbnfJson/terminals/GbnfBoolean.js.map +1 -0
  613. package/dist/utils/gbnfJson/terminals/GbnfBooleanValue.d.ts +7 -0
  614. package/dist/utils/gbnfJson/terminals/GbnfBooleanValue.js +17 -0
  615. package/dist/utils/gbnfJson/terminals/GbnfBooleanValue.js.map +1 -0
  616. package/dist/utils/gbnfJson/terminals/GbnfCommaWhitespace.d.ts +11 -0
  617. package/dist/utils/gbnfJson/terminals/GbnfCommaWhitespace.js +28 -0
  618. package/dist/utils/gbnfJson/terminals/GbnfCommaWhitespace.js.map +1 -0
  619. package/dist/utils/gbnfJson/terminals/GbnfFormatString.d.ts +11 -0
  620. package/dist/utils/gbnfJson/terminals/GbnfFormatString.js +90 -0
  621. package/dist/utils/gbnfJson/terminals/GbnfFormatString.js.map +1 -0
  622. package/dist/utils/gbnfJson/terminals/GbnfGrammar.d.ts +9 -0
  623. package/dist/utils/gbnfJson/terminals/GbnfGrammar.js +23 -0
  624. package/dist/utils/gbnfJson/terminals/GbnfGrammar.js.map +1 -0
  625. package/dist/utils/gbnfJson/terminals/GbnfInsideStringChar.d.ts +5 -0
  626. package/dist/utils/gbnfJson/terminals/GbnfInsideStringChar.js +24 -0
  627. package/dist/utils/gbnfJson/terminals/GbnfInsideStringChar.js.map +1 -0
  628. package/dist/utils/gbnfJson/terminals/GbnfNull.d.ts +5 -0
  629. package/dist/utils/gbnfJson/terminals/GbnfNull.js +11 -0
  630. package/dist/utils/gbnfJson/terminals/GbnfNull.js.map +1 -0
  631. package/dist/utils/gbnfJson/terminals/GbnfNumber.d.ts +9 -0
  632. package/dist/utils/gbnfJson/terminals/GbnfNumber.js +22 -0
  633. package/dist/utils/gbnfJson/terminals/GbnfNumber.js.map +1 -0
  634. package/dist/utils/gbnfJson/terminals/GbnfNumberValue.d.ts +9 -0
  635. package/dist/utils/gbnfJson/terminals/GbnfNumberValue.js +21 -0
  636. package/dist/utils/gbnfJson/terminals/GbnfNumberValue.js.map +1 -0
  637. package/dist/utils/gbnfJson/terminals/GbnfObjectMap.d.ts +28 -0
  638. package/dist/utils/gbnfJson/terminals/GbnfObjectMap.js +88 -0
  639. package/dist/utils/gbnfJson/terminals/GbnfObjectMap.js.map +1 -0
  640. package/dist/utils/gbnfJson/terminals/GbnfOr.d.ts +9 -0
  641. package/dist/utils/gbnfJson/terminals/GbnfOr.js +34 -0
  642. package/dist/utils/gbnfJson/terminals/GbnfOr.js.map +1 -0
  643. package/dist/utils/gbnfJson/terminals/GbnfRef.d.ts +15 -0
  644. package/dist/utils/gbnfJson/terminals/GbnfRef.js +34 -0
  645. package/dist/utils/gbnfJson/terminals/GbnfRef.js.map +1 -0
  646. package/dist/utils/gbnfJson/terminals/GbnfRepetition.d.ts +15 -0
  647. package/dist/utils/gbnfJson/terminals/GbnfRepetition.js +86 -0
  648. package/dist/utils/gbnfJson/terminals/GbnfRepetition.js.map +1 -0
  649. package/dist/utils/gbnfJson/terminals/GbnfString.d.ts +12 -0
  650. package/dist/utils/gbnfJson/terminals/GbnfString.js +43 -0
  651. package/dist/utils/gbnfJson/terminals/GbnfString.js.map +1 -0
  652. package/dist/utils/gbnfJson/terminals/GbnfStringValue.d.ts +8 -0
  653. package/dist/utils/gbnfJson/terminals/GbnfStringValue.js +26 -0
  654. package/dist/utils/gbnfJson/terminals/GbnfStringValue.js.map +1 -0
  655. package/dist/utils/gbnfJson/terminals/GbnfVerbatimText.d.ts +6 -0
  656. package/dist/utils/gbnfJson/terminals/GbnfVerbatimText.js +21 -0
  657. package/dist/utils/gbnfJson/terminals/GbnfVerbatimText.js.map +1 -0
  658. package/dist/utils/gbnfJson/terminals/GbnfWhitespace.d.ts +13 -0
  659. package/dist/utils/gbnfJson/terminals/GbnfWhitespace.js +67 -0
  660. package/dist/utils/gbnfJson/terminals/GbnfWhitespace.js.map +1 -0
  661. package/dist/utils/gbnfJson/terminals/gbnfConsts.d.ts +30 -0
  662. package/dist/utils/gbnfJson/terminals/gbnfConsts.js +72 -0
  663. package/dist/utils/gbnfJson/terminals/gbnfConsts.js.map +1 -0
  664. package/dist/utils/gbnfJson/types.d.ts +213 -0
  665. package/dist/utils/gbnfJson/types.js +30 -0
  666. package/dist/utils/gbnfJson/types.js.map +1 -0
  667. package/dist/utils/gbnfJson/utils/GbnfJsonScopeState.d.ts +10 -0
  668. package/dist/utils/gbnfJson/utils/GbnfJsonScopeState.js +15 -0
  669. package/dist/utils/gbnfJson/utils/GbnfJsonScopeState.js.map +1 -0
  670. package/dist/utils/gbnfJson/utils/defsScope.d.ts +7 -0
  671. package/dist/utils/gbnfJson/utils/defsScope.js +17 -0
  672. package/dist/utils/gbnfJson/utils/defsScope.js.map +1 -0
  673. package/dist/utils/gbnfJson/utils/getGbnfJsonTerminalForGbnfJsonSchema.d.ts +5 -0
  674. package/dist/utils/gbnfJson/utils/getGbnfJsonTerminalForGbnfJsonSchema.js +143 -0
  675. package/dist/utils/gbnfJson/utils/getGbnfJsonTerminalForGbnfJsonSchema.js.map +1 -0
  676. package/dist/utils/gbnfJson/utils/getGbnfJsonTerminalForLiteral.d.ts +5 -0
  677. package/dist/utils/gbnfJson/utils/getGbnfJsonTerminalForLiteral.js +16 -0
  678. package/dist/utils/gbnfJson/utils/getGbnfJsonTerminalForLiteral.js.map +1 -0
  679. package/dist/utils/gbnfJson/utils/validateObjectAgainstGbnfSchema.d.ts +8 -0
  680. package/dist/utils/gbnfJson/utils/validateObjectAgainstGbnfSchema.js +242 -0
  681. package/dist/utils/gbnfJson/utils/validateObjectAgainstGbnfSchema.js.map +1 -0
  682. package/dist/utils/getBuildDefaults.d.ts +5 -0
  683. package/dist/utils/getBuildDefaults.js +9 -0
  684. package/dist/utils/getBuildDefaults.js.map +1 -0
  685. package/dist/utils/getChatWrapperSegmentDefinition.d.ts +2 -0
  686. package/dist/utils/getChatWrapperSegmentDefinition.js +9 -0
  687. package/dist/utils/getChatWrapperSegmentDefinition.js.map +1 -0
  688. package/dist/utils/getConsoleLogPrefix.d.ts +1 -0
  689. package/dist/utils/getConsoleLogPrefix.js +10 -0
  690. package/dist/utils/getConsoleLogPrefix.js.map +1 -0
  691. package/dist/utils/getFirstWritableDir.d.ts +8 -0
  692. package/dist/utils/getFirstWritableDir.js +60 -0
  693. package/dist/utils/getFirstWritableDir.js.map +1 -0
  694. package/dist/utils/getGrammarsFolder.d.ts +2 -0
  695. package/dist/utils/getGrammarsFolder.js +19 -0
  696. package/dist/utils/getGrammarsFolder.js.map +1 -0
  697. package/dist/utils/getLlamaClasses.d.ts +9 -0
  698. package/dist/utils/getLlamaClasses.js +14 -0
  699. package/dist/utils/getLlamaClasses.js.map +1 -0
  700. package/dist/utils/getModuleVersion.d.ts +1 -0
  701. package/dist/utils/getModuleVersion.js +13 -0
  702. package/dist/utils/getModuleVersion.js.map +1 -0
  703. package/dist/utils/getQueuedTokensBeforeStopTrigger.d.ts +6 -0
  704. package/dist/utils/getQueuedTokensBeforeStopTrigger.js +22 -0
  705. package/dist/utils/getQueuedTokensBeforeStopTrigger.js.map +1 -0
  706. package/dist/utils/getReadableContextSize.d.ts +1 -0
  707. package/dist/utils/getReadableContextSize.js +7 -0
  708. package/dist/utils/getReadableContextSize.js.map +1 -0
  709. package/dist/utils/getTempDir.d.ts +10 -0
  710. package/dist/utils/getTempDir.js +121 -0
  711. package/dist/utils/getTempDir.js.map +1 -0
  712. package/dist/utils/getTypeScriptTypeStringForGbnfJsonSchema.d.ts +2 -0
  713. package/dist/utils/getTypeScriptTypeStringForGbnfJsonSchema.js +205 -0
  714. package/dist/utils/getTypeScriptTypeStringForGbnfJsonSchema.js.map +1 -0
  715. package/dist/utils/gitReleaseBundles.d.ts +2 -0
  716. package/dist/utils/gitReleaseBundles.js +132 -0
  717. package/dist/utils/gitReleaseBundles.js.map +1 -0
  718. package/dist/utils/hashString.d.ts +1 -0
  719. package/dist/utils/hashString.js +8 -0
  720. package/dist/utils/hashString.js.map +1 -0
  721. package/dist/utils/includesText.d.ts +1 -0
  722. package/dist/utils/includesText.js +12 -0
  723. package/dist/utils/includesText.js.map +1 -0
  724. package/dist/utils/isLockfileActive.d.ts +4 -0
  725. package/dist/utils/isLockfileActive.js +12 -0
  726. package/dist/utils/isLockfileActive.js.map +1 -0
  727. package/dist/utils/isToken.d.ts +2 -0
  728. package/dist/utils/isToken.js +4 -0
  729. package/dist/utils/isToken.js.map +1 -0
  730. package/dist/utils/isUrl.d.ts +1 -0
  731. package/dist/utils/isUrl.js +15 -0
  732. package/dist/utils/isUrl.js.map +1 -0
  733. package/dist/utils/mergeUnionTypes.d.ts +10 -0
  734. package/dist/utils/mergeUnionTypes.js +2 -0
  735. package/dist/utils/mergeUnionTypes.js.map +1 -0
  736. package/dist/utils/modelDownloadEndpoints.d.ts +13 -0
  737. package/dist/utils/modelDownloadEndpoints.js +27 -0
  738. package/dist/utils/modelDownloadEndpoints.js.map +1 -0
  739. package/dist/utils/modelFileAccessTokens.d.ts +5 -0
  740. package/dist/utils/modelFileAccessTokens.js +41 -0
  741. package/dist/utils/modelFileAccessTokens.js.map +1 -0
  742. package/dist/utils/optionsMatrix.d.ts +58 -0
  743. package/dist/utils/optionsMatrix.js +97 -0
  744. package/dist/utils/optionsMatrix.js.map +1 -0
  745. package/dist/utils/parseModelFileName.d.ts +15 -0
  746. package/dist/utils/parseModelFileName.js +132 -0
  747. package/dist/utils/parseModelFileName.js.map +1 -0
  748. package/dist/utils/parseModelUri.d.ts +40 -0
  749. package/dist/utils/parseModelUri.js +346 -0
  750. package/dist/utils/parseModelUri.js.map +1 -0
  751. package/dist/utils/parseTextTemplate.d.ts +66 -0
  752. package/dist/utils/parseTextTemplate.js +116 -0
  753. package/dist/utils/parseTextTemplate.js.map +1 -0
  754. package/dist/utils/prettyPrintObject.d.ts +10 -0
  755. package/dist/utils/prettyPrintObject.js +84 -0
  756. package/dist/utils/prettyPrintObject.js.map +1 -0
  757. package/dist/utils/pushAll.d.ts +6 -0
  758. package/dist/utils/pushAll.js +11 -0
  759. package/dist/utils/pushAll.js.map +1 -0
  760. package/dist/utils/removeNullFields.d.ts +2 -0
  761. package/dist/utils/removeNullFields.js +17 -0
  762. package/dist/utils/removeNullFields.js.map +1 -0
  763. package/dist/utils/resolveGithubRelease.d.ts +2 -0
  764. package/dist/utils/resolveGithubRelease.js +77 -0
  765. package/dist/utils/resolveGithubRelease.js.map +1 -0
  766. package/dist/utils/resolveLastTokens.d.ts +2 -0
  767. package/dist/utils/resolveLastTokens.js +12 -0
  768. package/dist/utils/resolveLastTokens.js.map +1 -0
  769. package/dist/utils/resolveModelDestination.d.ts +16 -0
  770. package/dist/utils/resolveModelDestination.js +54 -0
  771. package/dist/utils/resolveModelDestination.js.map +1 -0
  772. package/dist/utils/resolveModelFile.d.ts +142 -0
  773. package/dist/utils/resolveModelFile.js +201 -0
  774. package/dist/utils/resolveModelFile.js.map +1 -0
  775. package/dist/utils/runtime.d.ts +4 -0
  776. package/dist/utils/runtime.js +8 -0
  777. package/dist/utils/runtime.js.map +1 -0
  778. package/dist/utils/safeEventCallback.d.ts +6 -0
  779. package/dist/utils/safeEventCallback.js +29 -0
  780. package/dist/utils/safeEventCallback.js.map +1 -0
  781. package/dist/utils/signalSleep.d.ts +1 -0
  782. package/dist/utils/signalSleep.js +20 -0
  783. package/dist/utils/signalSleep.js.map +1 -0
  784. package/dist/utils/spawnCommand.d.ts +11 -0
  785. package/dist/utils/spawnCommand.js +89 -0
  786. package/dist/utils/spawnCommand.js.map +1 -0
  787. package/dist/utils/tokenizeInput.d.ts +3 -0
  788. package/dist/utils/tokenizeInput.js +14 -0
  789. package/dist/utils/tokenizeInput.js.map +1 -0
  790. package/dist/utils/tokenizerUtils.d.ts +12 -0
  791. package/dist/utils/tokenizerUtils.js +32 -0
  792. package/dist/utils/tokenizerUtils.js.map +1 -0
  793. package/dist/utils/transformPromisable.d.ts +54 -0
  794. package/dist/utils/transformPromisable.js +95 -0
  795. package/dist/utils/transformPromisable.js.map +1 -0
  796. package/dist/utils/truncateTextAndRoundToWords.d.ts +11 -0
  797. package/dist/utils/truncateTextAndRoundToWords.js +110 -0
  798. package/dist/utils/truncateTextAndRoundToWords.js.map +1 -0
  799. package/dist/utils/utilTypes.d.ts +18 -0
  800. package/dist/utils/utilTypes.js +2 -0
  801. package/dist/utils/utilTypes.js.map +1 -0
  802. package/dist/utils/waitForLockfileRelease.d.ts +5 -0
  803. package/dist/utils/waitForLockfileRelease.js +19 -0
  804. package/dist/utils/waitForLockfileRelease.js.map +1 -0
  805. package/dist/utils/withLockfile.d.ts +7 -0
  806. package/dist/utils/withLockfile.js +44 -0
  807. package/dist/utils/withLockfile.js.map +1 -0
  808. package/dist/utils/withOra.d.ts +7 -0
  809. package/dist/utils/withOra.js +37 -0
  810. package/dist/utils/withOra.js.map +1 -0
  811. package/dist/utils/withProgressLog.d.ts +22 -0
  812. package/dist/utils/withProgressLog.js +211 -0
  813. package/dist/utils/withProgressLog.js.map +1 -0
  814. package/dist/utils/withStatusLogs.d.ts +6 -0
  815. package/dist/utils/withStatusLogs.js +25 -0
  816. package/dist/utils/withStatusLogs.js.map +1 -0
  817. package/dist/utils/wrapAbortSignal.d.ts +1 -0
  818. package/dist/utils/wrapAbortSignal.js +14 -0
  819. package/dist/utils/wrapAbortSignal.js.map +1 -0
  820. package/llama/.clang-format +46 -0
  821. package/llama/CMakeLists.txt +141 -0
  822. package/llama/addon/AddonContext.cpp +1181 -0
  823. package/llama/addon/AddonContext.h +85 -0
  824. package/llama/addon/AddonGrammar.cpp +92 -0
  825. package/llama/addon/AddonGrammar.h +22 -0
  826. package/llama/addon/AddonGrammarEvaluationState.cpp +36 -0
  827. package/llama/addon/AddonGrammarEvaluationState.h +17 -0
  828. package/llama/addon/AddonModel.cpp +691 -0
  829. package/llama/addon/AddonModel.h +64 -0
  830. package/llama/addon/AddonModelData.cpp +25 -0
  831. package/llama/addon/AddonModelData.h +15 -0
  832. package/llama/addon/AddonModelLora.cpp +103 -0
  833. package/llama/addon/AddonModelLora.h +28 -0
  834. package/llama/addon/AddonSampler.cpp +669 -0
  835. package/llama/addon/AddonSampler.h +75 -0
  836. package/llama/addon/RingBuffer.h +109 -0
  837. package/llama/addon/addon.cpp +330 -0
  838. package/llama/addon/addonGlobals.cpp +22 -0
  839. package/llama/addon/addonGlobals.h +12 -0
  840. package/llama/addon/globals/addonLog.cpp +143 -0
  841. package/llama/addon/globals/addonLog.h +24 -0
  842. package/llama/addon/globals/addonProgress.cpp +15 -0
  843. package/llama/addon/globals/addonProgress.h +15 -0
  844. package/llama/addon/globals/getGpuInfo.cpp +146 -0
  845. package/llama/addon/globals/getGpuInfo.h +11 -0
  846. package/llama/addon/globals/getMemoryInfo.cpp +63 -0
  847. package/llama/addon/globals/getMemoryInfo.h +4 -0
  848. package/llama/addon/globals/getSwapInfo.cpp +69 -0
  849. package/llama/addon/globals/getSwapInfo.h +4 -0
  850. package/llama/binariesGithubRelease.json +3 -0
  851. package/llama/cmake/addVariantSuffix.cmake +21 -0
  852. package/llama/cmake/win32.ensureNinjaPath.cmake +68 -0
  853. package/llama/cmake/win32.ensureNodeLib.cmake +34 -0
  854. package/llama/cmake/win32.llvmApplyGnuModeAdaptations.cmake +12 -0
  855. package/llama/cmake/win32.llvmEnsureCmakeAr.cmake +37 -0
  856. package/llama/cmake/win32.llvmUseGnuModeCompilers.cmake +87 -0
  857. package/llama/cmake/win32.programFilesPaths.cmake +35 -0
  858. package/llama/gpuInfo/vulkan-gpu-info.cpp +207 -0
  859. package/llama/gpuInfo/vulkan-gpu-info.h +9 -0
  860. package/llama/package.json +5 -0
  861. package/llama/profiles/llvm.win32.host-arm64.target-arm64.cmake +14 -0
  862. package/llama/profiles/llvm.win32.host-x64.target-arm64.cmake +14 -0
  863. package/llama/profiles/llvm.win32.host-x64.target-x64.cmake +14 -0
  864. package/llama/toolchains/darwin.host-x64.target-arm64.cmake +8 -0
  865. package/llama/toolchains/linux.host-arm64.target-x64.cmake +5 -0
  866. package/llama/toolchains/linux.host-x64.target-arm64.cmake +5 -0
  867. package/llama/toolchains/linux.host-x64.target-arm71.cmake +5 -0
  868. package/llama/toolchains/llvm.win32.host-x64.target-x64.cmake +20 -0
  869. package/llama/toolchains/win32.host-arm64.target-arm64.cmake +21 -0
  870. package/llama/toolchains/win32.host-x64.target-arm64.cmake +21 -0
  871. package/llama/xpack/package.json +10 -0
  872. package/package.json +241 -0
  873. package/templates/README.md +6 -0
  874. package/templates/package.json +10 -0
  875. package/templates/packed/electron-typescript-react.json +1 -0
  876. package/templates/packed/node-typescript.json +1 -0
@@ -0,0 +1 @@
1
+ {"version":3,"file":"getRamUsageFromUnifiedVram.js","sourceRoot":"","sources":["../../../../src/gguf/insights/utils/getRamUsageFromUnifiedVram.ts"],"names":[],"mappings":"AAAA,MAAM,UAAU,0BAA0B,CAAC,SAAiB,EAAE,SAA6D;IACvH,MAAM,YAAY,GAAG,SAAS,CAAC,KAAK,GAAG,SAAS,CAAC,WAAW,CAAC;IAC7D,MAAM,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,SAAS,CAAC,KAAK,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC;IAEpE,MAAM,eAAe,GAAG,IAAI,CAAC,GAAG,CAAC,SAAS,CAAC,WAAW,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,YAAY,GAAG,aAAa,CAAC,CAAC,CAAC,CAAC;IAE5H,OAAO,eAAe,CAAC;AAC3B,CAAC"}
@@ -0,0 +1,33 @@
1
+ import { LlamaContextOptions } from "../../../evaluator/LlamaContext/types.js";
2
+ import { GgufInsights } from "../GgufInsights.js";
3
+ import { BuildGpu } from "../../../bindings/types.js";
4
+ import type { GgmlType } from "../../types/GgufTensorInfoTypes.js";
5
+ export declare function resolveContextContextSizeOption({ contextSize, batchSize, sequences, modelFileInsights, modelGpuLayers, modelTrainContextSize, flashAttention, kvCacheKeyType, kvCacheValueType, swaFullCache, getVramState, getRamState, getSwapState, ignoreMemorySafetyChecks, isEmbeddingContext, maxContextSizeSwapUse }: {
6
+ contextSize?: LlamaContextOptions["contextSize"];
7
+ batchSize?: LlamaContextOptions["batchSize"];
8
+ sequences: number;
9
+ modelFileInsights: GgufInsights;
10
+ modelGpuLayers: number;
11
+ modelTrainContextSize: number;
12
+ flashAttention: boolean;
13
+ kvCacheKeyType?: GgmlType;
14
+ kvCacheValueType?: GgmlType;
15
+ swaFullCache: boolean;
16
+ getVramState(): Promise<{
17
+ total: number;
18
+ free: number;
19
+ unifiedSize: number;
20
+ }>;
21
+ getRamState(): Promise<{
22
+ total: number;
23
+ free: number;
24
+ }>;
25
+ getSwapState(): Promise<{
26
+ total: number;
27
+ free: number;
28
+ }>;
29
+ llamaGpu: BuildGpu;
30
+ ignoreMemorySafetyChecks?: boolean;
31
+ isEmbeddingContext?: boolean;
32
+ maxContextSizeSwapUse?: number;
33
+ }): Promise<number>;
@@ -0,0 +1,117 @@
1
+ import { minAllowedContextSizeInCalculations } from "../../../config.js";
2
+ import { getDefaultContextBatchSize, getDefaultModelContextSize } from "../../../evaluator/LlamaContext/LlamaContext.js";
3
+ import { InsufficientMemoryError } from "../../../utils/InsufficientMemoryError.js";
4
+ import { getRamUsageFromUnifiedVram } from "./getRamUsageFromUnifiedVram.js";
5
+ const defaultMaxContextSizeSwapUse = 2048;
6
+ export async function resolveContextContextSizeOption({ contextSize, batchSize, sequences, modelFileInsights, modelGpuLayers, modelTrainContextSize, flashAttention, kvCacheKeyType, kvCacheValueType, swaFullCache, getVramState, getRamState, getSwapState, ignoreMemorySafetyChecks = false, isEmbeddingContext = false, maxContextSizeSwapUse = defaultMaxContextSizeSwapUse }) {
7
+ if (contextSize == null)
8
+ contextSize = "auto";
9
+ if (typeof contextSize === "number") {
10
+ const resolvedContextSize = Math.max(1, Math.floor(contextSize));
11
+ if (ignoreMemorySafetyChecks)
12
+ return resolvedContextSize;
13
+ const [vramState, ramState, swapState] = await Promise.all([
14
+ getVramState(),
15
+ getRamState(),
16
+ getSwapState()
17
+ ]);
18
+ const contextResourceRequirements = modelFileInsights.estimateContextResourceRequirements({
19
+ contextSize: resolvedContextSize,
20
+ batchSize: batchSize ?? getDefaultContextBatchSize({ contextSize: resolvedContextSize, sequences }),
21
+ modelGpuLayers: modelGpuLayers,
22
+ sequences,
23
+ flashAttention,
24
+ kvCacheKeyType,
25
+ kvCacheValueType,
26
+ swaFullCache,
27
+ isEmbeddingContext
28
+ });
29
+ if (contextResourceRequirements.gpuVram > vramState.free)
30
+ throw new InsufficientMemoryError(`A context size of ${resolvedContextSize}${sequences > 1 ? ` with ${sequences} sequences` : ""} is too large for the available VRAM`);
31
+ else if (contextResourceRequirements.cpuRam > (ramState.free + swapState.free - getRamUsageFromUnifiedVram(contextResourceRequirements.gpuVram, vramState)))
32
+ throw new InsufficientMemoryError(`A context size of ${resolvedContextSize}${sequences > 1 ? ` with ${sequences} sequences` : ""} is too large for the available RAM${swapState.total > 0 ? " (including swap)" : ""}`);
33
+ return resolvedContextSize;
34
+ }
35
+ else if (contextSize === "auto" || typeof contextSize === "object") {
36
+ const [vramState, ramState, swapState] = await Promise.all([
37
+ getVramState(),
38
+ getRamState(),
39
+ getSwapState()
40
+ ]);
41
+ const maxContextSize = contextSize === "auto"
42
+ ? getDefaultModelContextSize({ trainContextSize: modelTrainContextSize })
43
+ : Math.min(contextSize.max ?? getDefaultModelContextSize({ trainContextSize: modelTrainContextSize }), getDefaultModelContextSize({ trainContextSize: modelTrainContextSize }));
44
+ const minContextSize = contextSize === "auto"
45
+ ? minAllowedContextSizeInCalculations
46
+ : Math.max(contextSize.min ?? minAllowedContextSizeInCalculations, minAllowedContextSizeInCalculations);
47
+ let highestCompatibleContextSize = null;
48
+ let step = -Math.max(1, Math.floor((maxContextSize - minContextSize) / 4));
49
+ for (let testContextSize = maxContextSize; testContextSize >= minContextSize && testContextSize <= maxContextSize;) {
50
+ const contextResourceRequirements = modelFileInsights.estimateContextResourceRequirements({
51
+ contextSize: testContextSize,
52
+ batchSize: batchSize ?? getDefaultContextBatchSize({ contextSize: testContextSize, sequences }),
53
+ modelGpuLayers: modelGpuLayers,
54
+ sequences,
55
+ flashAttention,
56
+ kvCacheKeyType,
57
+ kvCacheValueType,
58
+ swaFullCache,
59
+ isEmbeddingContext
60
+ });
61
+ if (contextResourceRequirements.gpuVram <= vramState.free &&
62
+ contextResourceRequirements.cpuRam <= (ramState.free - getRamUsageFromUnifiedVram(contextResourceRequirements.gpuVram, vramState) + (testContextSize <= maxContextSizeSwapUse
63
+ ? swapState.free
64
+ : 0))) {
65
+ if (highestCompatibleContextSize == null || testContextSize >= highestCompatibleContextSize) {
66
+ highestCompatibleContextSize = testContextSize;
67
+ if (step === -1)
68
+ break;
69
+ else if (step < 0)
70
+ step = Math.max(1, Math.floor(-step / 2));
71
+ }
72
+ }
73
+ else if (step > 0)
74
+ step = -Math.max(1, Math.floor(step / 2));
75
+ if (testContextSize == minContextSize && step === -1)
76
+ break;
77
+ testContextSize += step;
78
+ if (testContextSize < minContextSize) {
79
+ testContextSize = minContextSize;
80
+ step = Math.max(1, Math.floor(Math.abs(step) / 2));
81
+ }
82
+ else if (testContextSize > maxContextSize) {
83
+ testContextSize = maxContextSize;
84
+ step = -Math.max(1, Math.floor(Math.abs(step) / 2));
85
+ }
86
+ }
87
+ if (highestCompatibleContextSize != null)
88
+ return highestCompatibleContextSize;
89
+ if (ignoreMemorySafetyChecks)
90
+ return minContextSize;
91
+ const minContextSizeResourceRequirements = modelFileInsights.estimateContextResourceRequirements({
92
+ contextSize: minContextSize,
93
+ batchSize: batchSize ?? getDefaultContextBatchSize({ contextSize: minContextSize, sequences }),
94
+ modelGpuLayers: modelGpuLayers,
95
+ sequences,
96
+ flashAttention,
97
+ kvCacheKeyType,
98
+ kvCacheValueType,
99
+ swaFullCache,
100
+ isEmbeddingContext
101
+ });
102
+ const unifiedRamUsage = getRamUsageFromUnifiedVram(minContextSizeResourceRequirements.gpuVram, vramState);
103
+ if (minContextSizeResourceRequirements.gpuVram > vramState.free &&
104
+ minContextSizeResourceRequirements.cpuRam > ramState.free + swapState.free - unifiedRamUsage)
105
+ throw new InsufficientMemoryError(`A context size of ${minContextSize}${sequences > 1 ? ` with ${sequences} sequences` : ""} is too large for the available VRAM and RAM${swapState.total > 0 ? " (including swap)" : ""}`);
106
+ else if (minContextSizeResourceRequirements.gpuVram > vramState.free)
107
+ throw new InsufficientMemoryError(`A context size of ${minContextSize}${sequences > 1 ? ` with ${sequences} sequences` : ""} is too large for the available VRAM`);
108
+ else if (minContextSizeResourceRequirements.cpuRam > ramState.free + swapState.free - unifiedRamUsage)
109
+ throw new InsufficientMemoryError(`A context size of ${minContextSize}${sequences > 1 ? ` with ${sequences} sequences` : ""} is too large for the available RAM${swapState.total > 0 ? " (including swap)" : ""}`);
110
+ else if (minContextSizeResourceRequirements.cpuRam > ramState.free - unifiedRamUsage)
111
+ throw new InsufficientMemoryError(`A context size of ${minContextSize}${sequences > 1 ? ` with ${sequences} sequences` : ""} is too large for the available RAM`);
112
+ else
113
+ throw new InsufficientMemoryError(`A context size of ${minContextSize}${sequences > 1 ? ` with ${sequences} sequences` : ""} is too large for the available resources`);
114
+ }
115
+ throw new Error(`Invalid context size: "${contextSize}"`);
116
+ }
117
+ //# 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;AACvH,OAAO,EAAC,uBAAuB,EAAC,MAAM,2CAA2C,CAAC;AAClF,OAAO,EAAC,0BAA0B,EAAC,MAAM,iCAAiC,CAAC;AAG3E,MAAM,4BAA4B,GAAG,IAAI,CAAC;AAE1C,MAAM,CAAC,KAAK,UAAU,+BAA+B,CAAC,EAClD,WAAW,EAAE,SAAS,EAAE,SAAS,EAAE,iBAAiB,EAAE,cAAc,EAAE,qBAAqB,EAAE,cAAc,EAC3G,cAAc,EAAE,gBAAgB,EAAE,YAAY,EAC9C,YAAY,EAAE,WAAW,EAAE,YAAY,EAAE,wBAAwB,GAAG,KAAK,EAAE,kBAAkB,GAAG,KAAK,EACrG,qBAAqB,GAAG,4BAA4B,EAmBvD;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,CACF,SAAS,EACT,QAAQ,EACR,SAAS,CACZ,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;YAClB,YAAY,EAAE;YACd,WAAW,EAAE;YACb,YAAY,EAAE;SACjB,CAAC,CAAC;QACH,MAAM,2BAA2B,GAAG,iBAAiB,CAAC,mCAAmC,CAAC;YACtF,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,cAAc;YACd,gBAAgB;YAChB,YAAY;YACZ,kBAAkB;SACrB,CAAC,CAAC;QAEH,IAAI,2BAA2B,CAAC,OAAO,GAAG,SAAS,CAAC,IAAI;YACpD,MAAM,IAAI,uBAAuB,CAAC,qBAAqB,mBAAmB,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,SAAS,YAAY,CAAC,CAAC,CAAC,EAAE,sCAAsC,CAAC,CAAC;aACvK,IAAI,2BAA2B,CAAC,MAAM,GAAG,CAC1C,QAAQ,CAAC,IAAI,GAAG,SAAS,CAAC,IAAI,GAAG,0BAA0B,CAAC,2BAA2B,CAAC,OAAO,EAAE,SAAS,CAAC,CAC9G;YACG,MAAM,IAAI,uBAAuB,CAAC,qBAAqB,mBAAmB,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,SAAS,YAAY,CAAC,CAAC,CAAC,EAAE,sCAAsC,SAAS,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,mBAAmB,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;QAE5N,OAAO,mBAAmB,CAAC;IAC/B,CAAC;SAAM,IAAI,WAAW,KAAK,MAAM,IAAI,OAAO,WAAW,KAAK,QAAQ,EAAE,CAAC;QACnE,MAAM,CACF,SAAS,EACT,QAAQ,EACR,SAAS,CACZ,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;YAClB,YAAY,EAAE;YACd,WAAW,EAAE;YACb,YAAY,EAAE;SACjB,CAAC,CAAC;QAEH,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,2BAA2B,GAAG,iBAAiB,CAAC,mCAAmC,CAAC;gBACtF,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,cAAc;gBACd,gBAAgB;gBAChB,YAAY;gBACZ,kBAAkB;aACrB,CAAC,CAAC;YAEH,IAAI,2BAA2B,CAAC,OAAO,IAAI,SAAS,CAAC,IAAI;gBACrD,2BAA2B,CAAC,MAAM,IAAI,CAClC,QAAQ,CAAC,IAAI,GAAG,0BAA0B,CAAC,2BAA2B,CAAC,OAAO,EAAE,SAAS,CAAC,GAAG,CACzF,eAAe,IAAI,qBAAqB;oBACpC,CAAC,CAAC,SAAS,CAAC,IAAI;oBAChB,CAAC,CAAC,CAAC,CACV,CACJ,EACH,CAAC;gBACC,IAAI,4BAA4B,IAAI,IAAI,IAAI,eAAe,IAAI,4BAA4B,EAAE,CAAC;oBAC1F,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,kCAAkC,GAAG,iBAAiB,CAAC,mCAAmC,CAAC;YAC7F,WAAW,EAAE,cAAc;YAC3B,SAAS,EAAE,SAAS,IAAI,0BAA0B,CAAC,EAAC,WAAW,EAAE,cAAc,EAAE,SAAS,EAAC,CAAC;YAC5F,cAAc,EAAE,cAAc;YAC9B,SAAS;YACT,cAAc;YACd,cAAc;YACd,gBAAgB;YAChB,YAAY;YACZ,kBAAkB;SACrB,CAAC,CAAC;QAEH,MAAM,eAAe,GAAG,0BAA0B,CAAC,kCAAkC,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC;QAC1G,IAAI,kCAAkC,CAAC,OAAO,GAAG,SAAS,CAAC,IAAI;YAC3D,kCAAkC,CAAC,MAAM,GAAG,QAAQ,CAAC,IAAI,GAAG,SAAS,CAAC,IAAI,GAAG,eAAe;YAE5F,MAAM,IAAI,uBAAuB,CAAC,qBAAqB,cAAc,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,SAAS,YAAY,CAAC,CAAC,CAAC,EAAE,+CAA+C,SAAS,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,mBAAmB,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;aAC3N,IAAI,kCAAkC,CAAC,OAAO,GAAG,SAAS,CAAC,IAAI;YAChE,MAAM,IAAI,uBAAuB,CAAC,qBAAqB,cAAc,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,SAAS,YAAY,CAAC,CAAC,CAAC,EAAE,sCAAsC,CAAC,CAAC;aAClK,IAAI,kCAAkC,CAAC,MAAM,GAAG,QAAQ,CAAC,IAAI,GAAG,SAAS,CAAC,IAAI,GAAG,eAAe;YACjG,MAAM,IAAI,uBAAuB,CAAC,qBAAqB,cAAc,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,SAAS,YAAY,CAAC,CAAC,CAAC,EAAE,sCAAsC,SAAS,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,mBAAmB,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;aAClN,IAAI,kCAAkC,CAAC,MAAM,GAAG,QAAQ,CAAC,IAAI,GAAG,eAAe;YAChF,MAAM,IAAI,uBAAuB,CAAC,qBAAqB,cAAc,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,SAAS,YAAY,CAAC,CAAC,CAAC,EAAE,qCAAqC,CAAC,CAAC;;YAElK,MAAM,IAAI,uBAAuB,CAAC,qBAAqB,cAAc,GAAG,SAAS,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,SAAS,YAAY,CAAC,CAAC,CAAC,EAAE,2CAA2C,CAAC,CAAC;IAChL,CAAC;IAED,MAAM,IAAI,KAAK,CAAC,0BAA0B,WAAW,GAAG,CAAC,CAAC;AAC9D,CAAC"}
@@ -0,0 +1,20 @@
1
+ import { LlamaModelOptions } from "../../../evaluator/LlamaModel/LlamaModel.js";
2
+ import { BuildGpu } from "../../../bindings/types.js";
3
+ import type { GgmlType } from "../../types/GgufTensorInfoTypes.js";
4
+ import type { GgufInsights } from "../GgufInsights.js";
5
+ export declare function resolveModelGpuLayersOption(gpuLayers: LlamaModelOptions["gpuLayers"], { ggufInsights, ignoreMemorySafetyChecks, getVramState, llamaVramPaddingSize, llamaGpu, llamaSupportsGpuOffloading, defaultContextFlashAttention, defaultContextKvCacheKeyType, defaultContextKvCacheValueType, defaultContextSwaFullCache, useMmap }: {
6
+ ggufInsights: GgufInsights;
7
+ ignoreMemorySafetyChecks?: boolean;
8
+ getVramState(): Promise<{
9
+ total: number;
10
+ free: number;
11
+ }>;
12
+ llamaVramPaddingSize: number;
13
+ llamaGpu: BuildGpu;
14
+ llamaSupportsGpuOffloading: boolean;
15
+ defaultContextFlashAttention: boolean;
16
+ defaultContextKvCacheKeyType?: GgmlType;
17
+ defaultContextKvCacheValueType?: GgmlType;
18
+ defaultContextSwaFullCache: boolean;
19
+ useMmap?: boolean;
20
+ }): Promise<number>;
@@ -0,0 +1,251 @@
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, defaultContextKvCacheKeyType, defaultContextKvCacheValueType, defaultContextSwaFullCache, useMmap }) {
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
+ defaultContextKvCacheKeyType,
25
+ defaultContextKvCacheValueType,
26
+ defaultContextSwaFullCache,
27
+ useMmap
28
+ });
29
+ if (maxLayersRequirements == null)
30
+ throw new InsufficientMemoryError("Not enough VRAM to fit the model with the specified settings");
31
+ return resolvedGpuLayers;
32
+ }
33
+ else if (gpuLayers === "auto" || typeof gpuLayers === "object") {
34
+ if (llamaGpu === false)
35
+ return 0;
36
+ const vramState = await getVramState();
37
+ if (vramState.total === 0)
38
+ return 0;
39
+ let freeVram = vramState.free;
40
+ if (typeof gpuLayers === "object" && gpuLayers.fitContext?.contextSize != null) {
41
+ freeVram -= llamaVramPaddingSize * fitContextExtraMemoryPaddingPercentage;
42
+ if (freeVram < 0)
43
+ freeVram = 0;
44
+ }
45
+ const bestGpuLayersOption = getBestGpuLayersForFreeVram({
46
+ ggufInsights,
47
+ freeVram,
48
+ fitContext: typeof gpuLayers === "object"
49
+ ? gpuLayers.fitContext
50
+ : undefined,
51
+ minGpuLayers: typeof gpuLayers === "object"
52
+ ? gpuLayers.min
53
+ : undefined,
54
+ maxGpuLayers: typeof gpuLayers === "object"
55
+ ? gpuLayers.max
56
+ : undefined,
57
+ defaultContextFlashAttention,
58
+ defaultContextKvCacheKeyType,
59
+ defaultContextKvCacheValueType,
60
+ defaultContextSwaFullCache,
61
+ useMmap
62
+ });
63
+ const hasGpuLayersRequirements = typeof gpuLayers === "object" &&
64
+ (gpuLayers.min != null || gpuLayers.max != null || gpuLayers.fitContext?.contextSize != null);
65
+ if (!ignoreMemorySafetyChecks && bestGpuLayersOption == null && hasGpuLayersRequirements)
66
+ throw new InsufficientMemoryError("Not enough VRAM to fit the model with the specified settings");
67
+ return bestGpuLayersOption ?? 0;
68
+ }
69
+ throw new Error(`Invalid gpuLayers value: ${gpuLayers}`);
70
+ }
71
+ function getBestGpuLayersForFreeVram({ ggufInsights, freeVram, fitContext, minGpuLayers, maxGpuLayers, defaultContextFlashAttention, defaultContextKvCacheKeyType, defaultContextKvCacheValueType, defaultContextSwaFullCache, useMmap }) {
72
+ return findBestOption({
73
+ *generator() {
74
+ const minLayers = Math.floor(Math.max(0, minGpuLayers ?? 0));
75
+ const maxLayers = Math.floor(Math.min(ggufInsights.totalLayers, maxGpuLayers ?? ggufInsights.totalLayers));
76
+ for (let layers = maxLayers; layers >= minLayers; layers--) {
77
+ yield {
78
+ gpuLayers: layers
79
+ };
80
+ }
81
+ },
82
+ score(option) {
83
+ const layersRequirements = getVramRequiredForGpuLayers({
84
+ gpuLayers: option.gpuLayers,
85
+ ggufInsights,
86
+ currentVram: freeVram,
87
+ fitContext,
88
+ defaultContextFlashAttention,
89
+ defaultContextSwaFullCache,
90
+ defaultContextKvCacheKeyType,
91
+ defaultContextKvCacheValueType,
92
+ useMmap
93
+ });
94
+ if (layersRequirements == null)
95
+ return null;
96
+ return scoreGpuLayersAndContextCombination({ gpuLayers: option.gpuLayers, contextSize: layersRequirements.contextSize }, {
97
+ totalGpuLayers: ggufInsights.totalLayers,
98
+ trainContextSize: getDefaultModelContextSize({ trainContextSize: ggufInsights.trainContextSize })
99
+ });
100
+ }
101
+ })?.gpuLayers ?? null;
102
+ }
103
+ function scoreGpuLayersAndContextCombination({ gpuLayers, contextSize }, { totalGpuLayers, trainContextSize }) {
104
+ function scoreGpuLayers() {
105
+ return scoreLevels(gpuLayers, [{
106
+ start: 0,
107
+ points: 4
108
+ }, {
109
+ start: 1,
110
+ points: 26
111
+ }, {
112
+ start: totalGpuLayers,
113
+ points: 14,
114
+ end: totalGpuLayers
115
+ }]);
116
+ }
117
+ function scoreContextSize() {
118
+ const gpuLayersPercentage = gpuLayers / totalGpuLayers;
119
+ return scoreLevels(contextSize, [{
120
+ start: 0,
121
+ points: 2
122
+ }, {
123
+ start: 1024,
124
+ points: 4
125
+ }, {
126
+ start: 2048,
127
+ points: gpuLayersPercentage < 0.1 ? 1 : 8
128
+ }, {
129
+ start: 4096,
130
+ points: gpuLayersPercentage < 0.3 ? 4 : 16
131
+ }, {
132
+ start: 8192,
133
+ points: gpuLayersPercentage < 0.6 ? 1 : 8,
134
+ end: Math.max(trainContextSize, 16384)
135
+ }]);
136
+ }
137
+ return scoreGpuLayers() + scoreContextSize();
138
+ }
139
+ function getVramRequiredForGpuLayers({ gpuLayers, ggufInsights, currentVram, fitContext, defaultContextFlashAttention = false, defaultContextKvCacheKeyType, defaultContextKvCacheValueType, defaultContextSwaFullCache = false, useMmap }) {
140
+ const modelVram = ggufInsights.estimateModelResourceRequirements({
141
+ gpuLayers,
142
+ useMmap
143
+ }).gpuVram;
144
+ if (modelVram > currentVram)
145
+ return null;
146
+ if (fitContext != null && fitContext.contextSize != null) {
147
+ const contextVram = ggufInsights.estimateContextResourceRequirements({
148
+ contextSize: fitContext.contextSize,
149
+ batchSize: getDefaultContextBatchSize({ contextSize: fitContext.contextSize, sequences: 1 }),
150
+ modelGpuLayers: gpuLayers,
151
+ sequences: 1,
152
+ isEmbeddingContext: fitContext.embeddingContext ?? false,
153
+ flashAttention: defaultContextFlashAttention,
154
+ kvCacheKeyType: defaultContextKvCacheKeyType,
155
+ kvCacheValueType: defaultContextKvCacheValueType,
156
+ swaFullCache: defaultContextSwaFullCache
157
+ }).gpuVram;
158
+ const totalVram = modelVram + contextVram;
159
+ if (totalVram > currentVram)
160
+ return null;
161
+ return {
162
+ contextSize: fitContext.contextSize,
163
+ contextVram,
164
+ totalVram
165
+ };
166
+ }
167
+ const maxContext = findMaxPossibleContextSizeForVram({
168
+ gpuLayers,
169
+ ggufInsights,
170
+ vram: currentVram - modelVram,
171
+ isEmbeddingContext: fitContext?.embeddingContext ?? false,
172
+ flashAttention: defaultContextFlashAttention,
173
+ kvCacheKeyType: defaultContextKvCacheKeyType,
174
+ kvCacheValueType: defaultContextKvCacheValueType,
175
+ swaFullCache: defaultContextSwaFullCache
176
+ });
177
+ if (maxContext == null || modelVram + maxContext.vram > currentVram)
178
+ return null;
179
+ return {
180
+ contextSize: maxContext.contextSize,
181
+ contextVram: maxContext.vram,
182
+ totalVram: modelVram + maxContext.vram
183
+ };
184
+ }
185
+ function findMaxPossibleContextSizeForVram({ gpuLayers, ggufInsights, vram, isEmbeddingContext, flashAttention, kvCacheKeyType, kvCacheValueType, swaFullCache }) {
186
+ const maxContextSize = getDefaultModelContextSize({ trainContextSize: ggufInsights.trainContextSize });
187
+ return findMaxValidValue({
188
+ maxValue: maxContextSize,
189
+ minValue: minAllowedContextSizeInCalculations,
190
+ minStep: 1,
191
+ test(contextSize) {
192
+ const contextVram = ggufInsights.estimateContextResourceRequirements({
193
+ contextSize,
194
+ batchSize: getDefaultContextBatchSize({ contextSize, sequences: 1 }),
195
+ modelGpuLayers: gpuLayers,
196
+ sequences: 1,
197
+ isEmbeddingContext,
198
+ flashAttention,
199
+ kvCacheKeyType,
200
+ kvCacheValueType,
201
+ swaFullCache
202
+ }).gpuVram;
203
+ if (contextVram <= vram)
204
+ return {
205
+ contextSize,
206
+ vram: contextVram
207
+ };
208
+ return null;
209
+ }
210
+ });
211
+ }
212
+ function findMaxValidValue({ maxValue, minValue, minStep = 1, test }) {
213
+ let step = -Math.max(minStep, Math.floor((maxValue - minValue) / 4));
214
+ let bestValue = null;
215
+ for (let value = maxValue; value >= minValue;) {
216
+ const result = (bestValue != null && value === bestValue.value)
217
+ ? bestValue.result
218
+ : test(value);
219
+ if (result != null) {
220
+ if (bestValue == null || value >= bestValue.value) {
221
+ bestValue = { value: value, result: result };
222
+ if (step === -minStep)
223
+ break;
224
+ else if (step < 0)
225
+ step = Math.max(minStep, Math.floor(-step / 2));
226
+ }
227
+ }
228
+ else if (bestValue != null && value < bestValue.value) {
229
+ value = bestValue.value;
230
+ step = Math.max(minStep, Math.floor(Math.abs(step) / 2));
231
+ continue;
232
+ }
233
+ else if (step > 0)
234
+ step = -Math.max(minStep, Math.floor(step / 2));
235
+ if (value === minValue && step === -minStep)
236
+ break;
237
+ value += step;
238
+ if (value < minValue) {
239
+ value = minValue;
240
+ step = Math.max(minStep, Math.floor(Math.abs(step) / 2));
241
+ }
242
+ else if (value > maxValue) {
243
+ value = maxValue;
244
+ step = -Math.max(minStep, Math.floor(Math.abs(step) / 2));
245
+ }
246
+ }
247
+ if (bestValue != null)
248
+ return bestValue.result;
249
+ return null;
250
+ }
251
+ //# 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;AAI7C,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,EAClE,4BAA4B,EAAE,8BAA8B,EAAE,0BAA0B,EAAE,OAAO,EAOpG;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;YAC5B,4BAA4B;YAC5B,8BAA8B;YAC9B,0BAA0B;YAC1B,OAAO;SACV,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;YAC5B,4BAA4B;YAC5B,8BAA8B;YAC9B,0BAA0B;YAC1B,OAAO;SACV,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,EAC5B,4BAA4B,EAC5B,8BAA8B,EAC9B,0BAA0B,EAC1B,OAAO,EAYV;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;gBAC5B,0BAA0B;gBAC1B,4BAA4B;gBAC5B,8BAA8B;gBAC9B,OAAO;aACV,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,EACtF,4BAA4B,EAAE,8BAA8B,EAAE,0BAA0B,GAAG,KAAK,EAAE,OAAO,EAK5G;IACG,MAAM,SAAS,GAAG,YAAY,CAAC,iCAAiC,CAAC;QAC7D,SAAS;QACT,OAAO;KACV,CAAC,CAAC,OAAO,CAAC;IAEX,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;YAC5C,cAAc,EAAE,4BAA4B;YAC5C,gBAAgB,EAAE,8BAA8B;YAChD,YAAY,EAAE,0BAA0B;SAC3C,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;QAC5C,cAAc,EAAE,4BAA4B;QAC5C,gBAAgB,EAAE,8BAA8B;QAChD,YAAY,EAAE,0BAA0B;KAC3C,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,EACvC,SAAS,EAAE,YAAY,EAAE,IAAI,EAAE,kBAAkB,EAAE,cAAc,EAAE,cAAc,EAAE,gBAAgB,EAAE,YAAY,EAIpH;IACG,MAAM,cAAc,GAAG,0BAA0B,CAAC,EAAC,gBAAgB,EAAE,YAAY,CAAC,gBAAgB,EAAC,CAAC,CAAC;IAErG,OAAO,iBAAiB,CAAC;QACrB,QAAQ,EAAE,cAAc;QACxB,QAAQ,EAAE,mCAAmC;QAC7C,OAAO,EAAE,CAAC;QACV,IAAI,CAAC,WAAW;YACZ,MAAM,WAAW,GAAG,YAAY,CAAC,mCAAmC,CAAC;gBACjE,WAAW;gBACX,SAAS,EAAE,0BAA0B,CAAC,EAAC,WAAW,EAAE,SAAS,EAAE,CAAC,EAAC,CAAC;gBAClE,cAAc,EAAE,SAAS;gBACzB,SAAS,EAAE,CAAC;gBACZ,kBAAkB;gBAClB,cAAc;gBACd,cAAc;gBACd,gBAAgB;gBAChB,YAAY;aACf,CAAC,CAAC,OAAO,CAAC;YAEX,IAAI,WAAW,IAAI,IAAI;gBACnB,OAAO;oBACH,WAAW;oBACX,IAAI,EAAE,WAAW;iBACpB,CAAC;YAEN,OAAO,IAAI,CAAC;QAChB,CAAC;KACJ,CAAC,CAAC;AACP,CAAC;AAED,SAAS,iBAAiB,CAAI,EAC1B,QAAQ,EACR,QAAQ,EACR,OAAO,GAAG,CAAC,EACX,IAAI,EAMP;IACG,IAAI,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,QAAQ,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;IACrE,IAAI,SAAS,GAAsC,IAAI,CAAC;IAExD,KAAK,IAAI,KAAK,GAAG,QAAQ,EAAE,KAAK,IAAI,QAAQ,GAAG,CAAC;QAC5C,MAAM,MAAM,GAAa,CAAC,SAAS,IAAI,IAAI,IAAI,KAAK,KAAK,SAAS,CAAC,KAAK,CAAC;YACrE,CAAC,CAAC,SAAS,CAAC,MAAM;YAClB,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAElB,IAAI,MAAM,IAAI,IAAI,EAAE,CAAC;YACjB,IAAI,SAAS,IAAI,IAAI,IAAI,KAAK,IAAI,SAAS,CAAC,KAAK,EAAE,CAAC;gBAChD,SAAS,GAAG,EAAC,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAC,CAAC;gBAE3C,IAAI,IAAI,KAAK,CAAC,OAAO;oBACjB,MAAM;qBACL,IAAI,IAAI,GAAG,CAAC;oBACb,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC;YACxD,CAAC;QACL,CAAC;aAAM,IAAI,SAAS,IAAI,IAAI,IAAI,KAAK,GAAG,SAAS,CAAC,KAAK,EAAE,CAAC;YACtD,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC;YACxB,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;YACzD,SAAS;QACb,CAAC;aAAM,IAAI,IAAI,GAAG,CAAC;YACf,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC;QAEpD,IAAI,KAAK,KAAK,QAAQ,IAAI,IAAI,KAAK,CAAC,OAAO;YACvC,MAAM;QAEV,KAAK,IAAI,IAAI,CAAC;QACd,IAAI,KAAK,GAAG,QAAQ,EAAE,CAAC;YACnB,KAAK,GAAG,QAAQ,CAAC;YACjB,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;QAC7D,CAAC;aAAM,IAAI,KAAK,GAAG,QAAQ,EAAE,CAAC;YAC1B,KAAK,GAAG,QAAQ,CAAC;YACjB,IAAI,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,OAAO,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;QAC9D,CAAC;IACL,CAAC;IAED,IAAI,SAAS,IAAI,IAAI;QACjB,OAAO,SAAS,CAAC,MAAM,CAAC;IAE5B,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,MAAuD;IAC5F,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,CAAE,CAAC;QACzB,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"}
@@ -0,0 +1,20 @@
1
+ import { GgufReadOffset } from "../utils/GgufReadOffset.js";
2
+ import { GgufValueType, GgufVersionParserOptions, GgufVersionParserResult, MetadataKeyValueRecord, MetadataValue } from "../types/GgufFileInfoTypes.js";
3
+ import { Promisable } from "../../utils/transformPromisable.js";
4
+ export declare class GgufV2Parser {
5
+ private readonly _fileReader;
6
+ private readonly _shouldReadTensorInfo;
7
+ private readonly _ignoreKeys;
8
+ private readonly _readOffset;
9
+ private readonly _logWarnings;
10
+ constructor({ fileReader, readTensorInfo, ignoreKeys, readOffset, logWarnings }: GgufVersionParserOptions);
11
+ parse(): Promise<GgufVersionParserResult>;
12
+ protected _readGgufValue(type: GgufValueType, offset: number | GgufReadOffset): Promisable<MetadataValue>;
13
+ protected _readStringValue(offset: number | GgufReadOffset): Promisable<string>;
14
+ protected _readRawHeader(readOffset: GgufReadOffset): Promise<{
15
+ tensorCount: number | bigint;
16
+ metadata: MetadataKeyValueRecord;
17
+ headerSize: number;
18
+ }>;
19
+ private _readTensorInfo;
20
+ }