gpt-tokenizer 2.0.0 → 2.1.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (328) hide show
  1. package/README.md +54 -8
  2. package/cjs/GptEncoding.d.ts +33 -6
  3. package/cjs/GptEncoding.js +79 -34
  4. package/cjs/GptEncoding.js.map +1 -1
  5. package/cjs/GptEncoding.test.js +54 -0
  6. package/cjs/GptEncoding.test.js.map +1 -1
  7. package/cjs/encoding/cl100k_base.d.ts +2 -2
  8. package/cjs/encoding/cl100k_base.js +4 -2
  9. package/cjs/encoding/cl100k_base.js.map +1 -1
  10. package/cjs/encoding/p50k_base.d.ts +1 -1
  11. package/cjs/encoding/p50k_edit.d.ts +1 -1
  12. package/cjs/encoding/r50k_base.d.ts +1 -1
  13. package/cjs/main.d.ts +1 -0
  14. package/cjs/main.js +7 -0
  15. package/cjs/main.js.map +1 -1
  16. package/cjs/mapping.d.ts +37 -0
  17. package/cjs/mapping.js +34 -2
  18. package/cjs/mapping.js.map +1 -1
  19. package/cjs/model/ada.d.ts +1 -0
  20. package/cjs/model/ada.js +7 -0
  21. package/cjs/model/ada.js.map +1 -1
  22. package/cjs/model/babbage.d.ts +1 -0
  23. package/cjs/model/babbage.js +7 -0
  24. package/cjs/model/babbage.js.map +1 -1
  25. package/cjs/model/code-cushman-001.d.ts +1 -0
  26. package/cjs/model/code-cushman-001.js +7 -0
  27. package/cjs/model/code-cushman-001.js.map +1 -1
  28. package/cjs/model/code-cushman-002.d.ts +1 -0
  29. package/cjs/model/code-cushman-002.js +7 -0
  30. package/cjs/model/code-cushman-002.js.map +1 -1
  31. package/cjs/model/code-davinci-001.d.ts +1 -0
  32. package/cjs/model/code-davinci-001.js +7 -0
  33. package/cjs/model/code-davinci-001.js.map +1 -1
  34. package/cjs/model/code-davinci-002.d.ts +1 -0
  35. package/cjs/model/code-davinci-002.js +7 -0
  36. package/cjs/model/code-davinci-002.js.map +1 -1
  37. package/cjs/model/code-davinci-edit-001.d.ts +1 -0
  38. package/cjs/model/code-davinci-edit-001.js +7 -0
  39. package/cjs/model/code-davinci-edit-001.js.map +1 -1
  40. package/cjs/model/code-search-ada-code-001.d.ts +1 -0
  41. package/cjs/model/code-search-ada-code-001.js +7 -0
  42. package/cjs/model/code-search-ada-code-001.js.map +1 -1
  43. package/cjs/model/code-search-babbage-code-001.d.ts +1 -0
  44. package/cjs/model/code-search-babbage-code-001.js +7 -0
  45. package/cjs/model/code-search-babbage-code-001.js.map +1 -1
  46. package/cjs/model/curie.d.ts +1 -0
  47. package/cjs/model/curie.js +7 -0
  48. package/cjs/model/curie.js.map +1 -1
  49. package/cjs/model/cushman-codex.d.ts +1 -0
  50. package/cjs/model/cushman-codex.js +7 -0
  51. package/cjs/model/cushman-codex.js.map +1 -1
  52. package/cjs/model/davinci-codex.d.ts +1 -0
  53. package/cjs/model/davinci-codex.js +7 -0
  54. package/cjs/model/davinci-codex.js.map +1 -1
  55. package/cjs/model/davinci.d.ts +1 -0
  56. package/cjs/model/davinci.js +7 -0
  57. package/cjs/model/davinci.js.map +1 -1
  58. package/cjs/model/gpt-3.5-turbo-0301.d.ts +6 -0
  59. package/cjs/model/gpt-3.5-turbo-0301.js +38 -0
  60. package/cjs/model/gpt-3.5-turbo-0301.js.map +1 -0
  61. package/cjs/model/gpt-3.5-turbo.d.ts +6 -1
  62. package/cjs/model/gpt-3.5-turbo.js +21 -1
  63. package/cjs/model/gpt-3.5-turbo.js.map +1 -1
  64. package/cjs/model/gpt-4-0314.d.ts +6 -0
  65. package/cjs/model/gpt-4-0314.js +38 -0
  66. package/cjs/model/gpt-4-0314.js.map +1 -0
  67. package/cjs/model/gpt-4-32k-0314.d.ts +6 -0
  68. package/cjs/model/gpt-4-32k-0314.js +38 -0
  69. package/cjs/model/gpt-4-32k-0314.js.map +1 -0
  70. package/cjs/model/gpt-4-32k.d.ts +6 -0
  71. package/cjs/model/gpt-4-32k.js +38 -0
  72. package/cjs/model/gpt-4-32k.js.map +1 -0
  73. package/cjs/model/gpt-4.d.ts +6 -1
  74. package/cjs/model/gpt-4.js +21 -1
  75. package/cjs/model/gpt-4.js.map +1 -1
  76. package/cjs/model/text-ada-001.d.ts +1 -0
  77. package/cjs/model/text-ada-001.js +7 -0
  78. package/cjs/model/text-ada-001.js.map +1 -1
  79. package/cjs/model/text-babbage-001.d.ts +1 -0
  80. package/cjs/model/text-babbage-001.js +7 -0
  81. package/cjs/model/text-babbage-001.js.map +1 -1
  82. package/cjs/model/text-curie-001.d.ts +1 -0
  83. package/cjs/model/text-curie-001.js +7 -0
  84. package/cjs/model/text-curie-001.js.map +1 -1
  85. package/cjs/model/text-davinci-001.d.ts +1 -0
  86. package/cjs/model/text-davinci-001.js +7 -0
  87. package/cjs/model/text-davinci-001.js.map +1 -1
  88. package/cjs/model/text-davinci-002.d.ts +1 -0
  89. package/cjs/model/text-davinci-002.js +7 -0
  90. package/cjs/model/text-davinci-002.js.map +1 -1
  91. package/cjs/model/text-davinci-003.d.ts +1 -0
  92. package/cjs/model/text-davinci-003.js +7 -0
  93. package/cjs/model/text-davinci-003.js.map +1 -1
  94. package/cjs/model/text-davinci-edit-001.d.ts +1 -0
  95. package/cjs/model/text-davinci-edit-001.js +7 -0
  96. package/cjs/model/text-davinci-edit-001.js.map +1 -1
  97. package/cjs/model/text-embedding-ada-002.d.ts +1 -0
  98. package/cjs/model/text-embedding-ada-002.js +7 -0
  99. package/cjs/model/text-embedding-ada-002.js.map +1 -1
  100. package/cjs/model/text-search-ada-doc-001.d.ts +1 -0
  101. package/cjs/model/text-search-ada-doc-001.js +7 -0
  102. package/cjs/model/text-search-ada-doc-001.js.map +1 -1
  103. package/cjs/model/text-search-babbage-doc-001.d.ts +1 -0
  104. package/cjs/model/text-search-babbage-doc-001.js +7 -0
  105. package/cjs/model/text-search-babbage-doc-001.js.map +1 -1
  106. package/cjs/model/text-search-curie-doc-001.d.ts +1 -0
  107. package/cjs/model/text-search-curie-doc-001.js +7 -0
  108. package/cjs/model/text-search-curie-doc-001.js.map +1 -1
  109. package/cjs/model/text-search-davinci-doc-001.d.ts +1 -0
  110. package/cjs/model/text-search-davinci-doc-001.js +7 -0
  111. package/cjs/model/text-search-davinci-doc-001.js.map +1 -1
  112. package/cjs/model/text-similarity-ada-001.d.ts +1 -0
  113. package/cjs/model/text-similarity-ada-001.js +7 -0
  114. package/cjs/model/text-similarity-ada-001.js.map +1 -1
  115. package/cjs/model/text-similarity-babbage-001.d.ts +1 -0
  116. package/cjs/model/text-similarity-babbage-001.js +7 -0
  117. package/cjs/model/text-similarity-babbage-001.js.map +1 -1
  118. package/cjs/model/text-similarity-curie-001.d.ts +1 -0
  119. package/cjs/model/text-similarity-curie-001.js +7 -0
  120. package/cjs/model/text-similarity-curie-001.js.map +1 -1
  121. package/cjs/model/text-similarity-davinci-001.d.ts +1 -0
  122. package/cjs/model/text-similarity-davinci-001.js +7 -0
  123. package/cjs/model/text-similarity-davinci-001.js.map +1 -1
  124. package/cjs/modelParams.d.ts +5 -4
  125. package/cjs/modelParams.js +7 -4
  126. package/cjs/modelParams.js.map +1 -1
  127. package/cjs/specialTokens.d.ts +3 -0
  128. package/cjs/specialTokens.js +4 -1
  129. package/cjs/specialTokens.js.map +1 -1
  130. package/cjs/util.d.ts +2 -0
  131. package/cjs/util.js +19 -0
  132. package/cjs/util.js.map +1 -0
  133. package/dist/cl100k_base.js +1 -1
  134. package/dist/cl100k_base.js.map +1 -1
  135. package/dist/p50k_base.js +1 -1
  136. package/dist/p50k_base.js.map +1 -1
  137. package/dist/p50k_edit.js +1 -1
  138. package/dist/p50k_edit.js.map +1 -1
  139. package/dist/r50k_base.js +1 -1
  140. package/dist/r50k_base.js.map +1 -1
  141. package/esm/GptEncoding.d.ts +33 -6
  142. package/esm/GptEncoding.js +81 -36
  143. package/esm/GptEncoding.js.map +1 -1
  144. package/esm/GptEncoding.test.js +55 -1
  145. package/esm/GptEncoding.test.js.map +1 -1
  146. package/esm/codegen/generateByModel.js +19 -0
  147. package/esm/codegen/generateByModel.js.map +1 -0
  148. package/esm/{generateJsModels.js → codegen/generateJsEncodings.js} +4 -4
  149. package/esm/codegen/generateJsEncodings.js.map +1 -0
  150. package/esm/encoding/cl100k_base.d.ts +2 -2
  151. package/esm/encoding/cl100k_base.js +2 -2
  152. package/esm/encoding/cl100k_base.js.map +1 -1
  153. package/esm/encoding/p50k_base.d.ts +1 -1
  154. package/esm/encoding/p50k_edit.d.ts +1 -1
  155. package/esm/encoding/r50k_base.d.ts +1 -1
  156. package/esm/main.d.ts +1 -0
  157. package/esm/main.js +2 -0
  158. package/esm/main.js.map +1 -1
  159. package/esm/mapping.d.ts +37 -0
  160. package/esm/mapping.js +32 -0
  161. package/esm/mapping.js.map +1 -1
  162. package/esm/model/ada.d.ts +1 -0
  163. package/esm/model/ada.js +2 -0
  164. package/esm/model/ada.js.map +1 -1
  165. package/esm/model/babbage.d.ts +1 -0
  166. package/esm/model/babbage.js +2 -0
  167. package/esm/model/babbage.js.map +1 -1
  168. package/esm/model/code-cushman-001.d.ts +1 -0
  169. package/esm/model/code-cushman-001.js +2 -0
  170. package/esm/model/code-cushman-001.js.map +1 -1
  171. package/esm/model/code-cushman-002.d.ts +1 -0
  172. package/esm/model/code-cushman-002.js +2 -0
  173. package/esm/model/code-cushman-002.js.map +1 -1
  174. package/esm/model/code-davinci-001.d.ts +1 -0
  175. package/esm/model/code-davinci-001.js +2 -0
  176. package/esm/model/code-davinci-001.js.map +1 -1
  177. package/esm/model/code-davinci-002.d.ts +1 -0
  178. package/esm/model/code-davinci-002.js +2 -0
  179. package/esm/model/code-davinci-002.js.map +1 -1
  180. package/esm/model/code-davinci-edit-001.d.ts +1 -0
  181. package/esm/model/code-davinci-edit-001.js +2 -0
  182. package/esm/model/code-davinci-edit-001.js.map +1 -1
  183. package/esm/model/code-search-ada-code-001.d.ts +1 -0
  184. package/esm/model/code-search-ada-code-001.js +2 -0
  185. package/esm/model/code-search-ada-code-001.js.map +1 -1
  186. package/esm/model/code-search-babbage-code-001.d.ts +1 -0
  187. package/esm/model/code-search-babbage-code-001.js +2 -0
  188. package/esm/model/code-search-babbage-code-001.js.map +1 -1
  189. package/esm/model/curie.d.ts +1 -0
  190. package/esm/model/curie.js +2 -0
  191. package/esm/model/curie.js.map +1 -1
  192. package/esm/model/cushman-codex.d.ts +1 -0
  193. package/esm/model/cushman-codex.js +2 -0
  194. package/esm/model/cushman-codex.js.map +1 -1
  195. package/esm/model/davinci-codex.d.ts +1 -0
  196. package/esm/model/davinci-codex.js +2 -0
  197. package/esm/model/davinci-codex.js.map +1 -1
  198. package/esm/model/davinci.d.ts +1 -0
  199. package/esm/model/davinci.js +2 -0
  200. package/esm/model/davinci.js.map +1 -1
  201. package/esm/model/gpt-3.5-turbo-0301.d.ts +6 -0
  202. package/esm/model/gpt-3.5-turbo-0301.js +11 -0
  203. package/esm/model/gpt-3.5-turbo-0301.js.map +1 -0
  204. package/esm/model/gpt-3.5-turbo.d.ts +6 -1
  205. package/esm/model/gpt-3.5-turbo.js +10 -1
  206. package/esm/model/gpt-3.5-turbo.js.map +1 -1
  207. package/esm/model/gpt-4-0314.d.ts +6 -0
  208. package/esm/model/gpt-4-0314.js +11 -0
  209. package/esm/model/gpt-4-0314.js.map +1 -0
  210. package/esm/model/gpt-4-32k-0314.d.ts +6 -0
  211. package/esm/model/gpt-4-32k-0314.js +11 -0
  212. package/esm/model/gpt-4-32k-0314.js.map +1 -0
  213. package/esm/model/gpt-4-32k.d.ts +6 -0
  214. package/esm/model/gpt-4-32k.js +11 -0
  215. package/esm/model/gpt-4-32k.js.map +1 -0
  216. package/esm/model/gpt-4.d.ts +6 -1
  217. package/esm/model/gpt-4.js +10 -1
  218. package/esm/model/gpt-4.js.map +1 -1
  219. package/esm/model/text-ada-001.d.ts +1 -0
  220. package/esm/model/text-ada-001.js +2 -0
  221. package/esm/model/text-ada-001.js.map +1 -1
  222. package/esm/model/text-babbage-001.d.ts +1 -0
  223. package/esm/model/text-babbage-001.js +2 -0
  224. package/esm/model/text-babbage-001.js.map +1 -1
  225. package/esm/model/text-curie-001.d.ts +1 -0
  226. package/esm/model/text-curie-001.js +2 -0
  227. package/esm/model/text-curie-001.js.map +1 -1
  228. package/esm/model/text-davinci-001.d.ts +1 -0
  229. package/esm/model/text-davinci-001.js +2 -0
  230. package/esm/model/text-davinci-001.js.map +1 -1
  231. package/esm/model/text-davinci-002.d.ts +1 -0
  232. package/esm/model/text-davinci-002.js +2 -0
  233. package/esm/model/text-davinci-002.js.map +1 -1
  234. package/esm/model/text-davinci-003.d.ts +1 -0
  235. package/esm/model/text-davinci-003.js +2 -0
  236. package/esm/model/text-davinci-003.js.map +1 -1
  237. package/esm/model/text-davinci-edit-001.d.ts +1 -0
  238. package/esm/model/text-davinci-edit-001.js +2 -0
  239. package/esm/model/text-davinci-edit-001.js.map +1 -1
  240. package/esm/model/text-embedding-ada-002.d.ts +1 -0
  241. package/esm/model/text-embedding-ada-002.js +2 -0
  242. package/esm/model/text-embedding-ada-002.js.map +1 -1
  243. package/esm/model/text-search-ada-doc-001.d.ts +1 -0
  244. package/esm/model/text-search-ada-doc-001.js +2 -0
  245. package/esm/model/text-search-ada-doc-001.js.map +1 -1
  246. package/esm/model/text-search-babbage-doc-001.d.ts +1 -0
  247. package/esm/model/text-search-babbage-doc-001.js +2 -0
  248. package/esm/model/text-search-babbage-doc-001.js.map +1 -1
  249. package/esm/model/text-search-curie-doc-001.d.ts +1 -0
  250. package/esm/model/text-search-curie-doc-001.js +2 -0
  251. package/esm/model/text-search-curie-doc-001.js.map +1 -1
  252. package/esm/model/text-search-davinci-doc-001.d.ts +1 -0
  253. package/esm/model/text-search-davinci-doc-001.js +2 -0
  254. package/esm/model/text-search-davinci-doc-001.js.map +1 -1
  255. package/esm/model/text-similarity-ada-001.d.ts +1 -0
  256. package/esm/model/text-similarity-ada-001.js +2 -0
  257. package/esm/model/text-similarity-ada-001.js.map +1 -1
  258. package/esm/model/text-similarity-babbage-001.d.ts +1 -0
  259. package/esm/model/text-similarity-babbage-001.js +2 -0
  260. package/esm/model/text-similarity-babbage-001.js.map +1 -1
  261. package/esm/model/text-similarity-curie-001.d.ts +1 -0
  262. package/esm/model/text-similarity-curie-001.js +2 -0
  263. package/esm/model/text-similarity-curie-001.js.map +1 -1
  264. package/esm/model/text-similarity-davinci-001.d.ts +1 -0
  265. package/esm/model/text-similarity-davinci-001.js +2 -0
  266. package/esm/model/text-similarity-davinci-001.js.map +1 -1
  267. package/esm/modelParams.d.ts +5 -4
  268. package/esm/modelParams.js +6 -3
  269. package/esm/modelParams.js.map +1 -1
  270. package/esm/specialTokens.d.ts +3 -0
  271. package/esm/specialTokens.js +3 -0
  272. package/esm/specialTokens.js.map +1 -1
  273. package/esm/util.d.ts +2 -0
  274. package/esm/util.js +14 -0
  275. package/esm/util.js.map +1 -0
  276. package/package.json +8 -6
  277. package/src/GptEncoding.test.ts +78 -1
  278. package/src/GptEncoding.ts +125 -44
  279. package/src/__snapshots__/GptEncoding.test.ts.snap +811 -0
  280. package/src/codegen/generateByModel.ts +35 -0
  281. package/src/{generateJsModels.ts → codegen/generateJsEncodings.ts} +3 -3
  282. package/src/encoding/cl100k_base.ts +4 -0
  283. package/src/main.ts +2 -0
  284. package/src/mapping.ts +42 -0
  285. package/src/model/ada.ts +2 -0
  286. package/src/model/babbage.ts +2 -0
  287. package/src/model/code-cushman-001.ts +2 -0
  288. package/src/model/code-cushman-002.ts +2 -0
  289. package/src/model/code-davinci-001.ts +2 -0
  290. package/src/model/code-davinci-002.ts +2 -0
  291. package/src/model/code-davinci-edit-001.ts +2 -0
  292. package/src/model/code-search-ada-code-001.ts +2 -0
  293. package/src/model/code-search-babbage-code-001.ts +2 -0
  294. package/src/model/curie.ts +2 -0
  295. package/src/model/cushman-codex.ts +2 -0
  296. package/src/model/davinci-codex.ts +2 -0
  297. package/src/model/davinci.ts +2 -0
  298. package/src/model/gpt-3.5-turbo-0301.ts +32 -0
  299. package/src/model/gpt-3.5-turbo.ts +32 -1
  300. package/src/model/gpt-4-0314.ts +32 -0
  301. package/src/model/gpt-4-32k-0314.ts +32 -0
  302. package/src/model/gpt-4-32k.ts +32 -0
  303. package/src/model/gpt-4.ts +32 -1
  304. package/src/model/text-ada-001.ts +2 -0
  305. package/src/model/text-babbage-001.ts +2 -0
  306. package/src/model/text-curie-001.ts +2 -0
  307. package/src/model/text-davinci-001.ts +2 -0
  308. package/src/model/text-davinci-002.ts +2 -0
  309. package/src/model/text-davinci-003.ts +2 -0
  310. package/src/model/text-davinci-edit-001.ts +2 -0
  311. package/src/model/text-embedding-ada-002.ts +2 -0
  312. package/src/model/text-search-ada-doc-001.ts +2 -0
  313. package/src/model/text-search-babbage-doc-001.ts +2 -0
  314. package/src/model/text-search-curie-doc-001.ts +2 -0
  315. package/src/model/text-search-davinci-doc-001.ts +2 -0
  316. package/src/model/text-similarity-ada-001.ts +2 -0
  317. package/src/model/text-similarity-babbage-001.ts +2 -0
  318. package/src/model/text-similarity-curie-001.ts +2 -0
  319. package/src/model/text-similarity-davinci-001.ts +2 -0
  320. package/src/modelParams.ts +17 -10
  321. package/src/specialTokens.ts +3 -0
  322. package/src/util.ts +15 -0
  323. package/esm/generateByModel.js +0 -13
  324. package/esm/generateByModel.js.map +0 -1
  325. package/esm/generateJsModels.js.map +0 -1
  326. package/src/generateByModel.ts +0 -20
  327. /package/esm/{generateByModel.d.ts → codegen/generateByModel.d.ts} +0 -0
  328. /package/esm/{generateJsModels.d.ts → codegen/generateJsEncodings.d.ts} +0 -0
@@ -1,6 +1,6 @@
1
1
  "use strict";
2
2
  Object.defineProperty(exports, "__esModule", { value: true });
3
- exports.getModelParamsAsync = exports.getModelParams = void 0;
3
+ exports.getModelParamsAsync = exports.getEncodingParams = void 0;
4
4
  const specialTokens_js_1 = require("./specialTokens.js");
5
5
  const tokenSplitRegex = /'s|'t|'re|'ve|'m|'ll|'d| ?\p{L}+| ?\p{N}+| ?[^\s\p{L}\p{N}]+|\s+(?!\S)|\s+/gu;
6
6
  function R50KBase(mergeableBytePairRanks) {
@@ -38,6 +38,9 @@ function Cl100KBase(mergeableBytePairRanks) {
38
38
  [specialTokens_js_1.FimPrefix, 100258],
39
39
  [specialTokens_js_1.FimMiddle, 100259],
40
40
  [specialTokens_js_1.FimSuffix, 100260],
41
+ [specialTokens_js_1.ImStart, 100264],
42
+ [specialTokens_js_1.ImEnd, 100265],
43
+ [specialTokens_js_1.ImSep, 100266],
41
44
  [specialTokens_js_1.EndOfPrompt, 100276],
42
45
  ]);
43
46
  return {
@@ -46,7 +49,7 @@ function Cl100KBase(mergeableBytePairRanks) {
46
49
  specialTokenMapping,
47
50
  };
48
51
  }
49
- function getModelParams(encodingName, getMergeableRanks) {
52
+ function getEncodingParams(encodingName, getMergeableRanks) {
50
53
  const mergeableBytePairRanks = getMergeableRanks(encodingName);
51
54
  switch (encodingName.toLowerCase()) {
52
55
  case 'r50k_base':
@@ -61,10 +64,10 @@ function getModelParams(encodingName, getMergeableRanks) {
61
64
  throw new Error(`Unknown encoding name: ${encodingName}`);
62
65
  }
63
66
  }
64
- exports.getModelParams = getModelParams;
67
+ exports.getEncodingParams = getEncodingParams;
65
68
  async function getModelParamsAsync(encodingName, getMergeableRanks) {
66
69
  const mergeableBytePairRanks = await getMergeableRanks(encodingName);
67
- return getModelParams(encodingName, () => mergeableBytePairRanks);
70
+ return getEncodingParams(encodingName, () => mergeableBytePairRanks);
68
71
  }
69
72
  exports.getModelParamsAsync = getModelParamsAsync;
70
73
  //# sourceMappingURL=modelParams.js.map
@@ -1 +1 @@
1
- {"version":3,"file":"modelParams.js","sourceRoot":"","sources":["../src/modelParams.ts"],"names":[],"mappings":";;;AAGA,yDAM2B;AAmB3B,MAAM,eAAe,GACnB,8EAA8E,CAAA;AAEhF,SAAS,QAAQ,CAAC,sBAAkC;IAClD,OAAO;QACL,sBAAsB,EAAE,KAAM;QAC9B,eAAe;QACf,sBAAsB;QACtB,mBAAmB,EAAE,IAAI,GAAG,CAAiB,CAAC,CAAC,4BAAS,EAAE,KAAM,CAAC,CAAC,CAAC;KACpE,CAAA;AACH,CAAC;AAED,SAAS,QAAQ,CAAC,sBAAkC;IAClD,OAAO;QACL,sBAAsB,EAAE,KAAM;QAC9B,eAAe;QACf,sBAAsB;QACtB,mBAAmB,EAAE,IAAI,GAAG,CAAiB,CAAC,CAAC,4BAAS,EAAE,KAAM,CAAC,CAAC,CAAC;KACpE,CAAA;AACH,CAAC;AAED,SAAS,QAAQ,CAAC,sBAAkC;IAClD,MAAM,mBAAmB,GAAG,IAAI,GAAG,CAAiB;QAClD,CAAC,4BAAS,EAAE,KAAM,CAAC;QACnB,CAAC,4BAAS,EAAE,KAAM,CAAC;QACnB,CAAC,4BAAS,EAAE,KAAM,CAAC;QACnB,CAAC,4BAAS,EAAE,KAAM,CAAC;KACpB,CAAC,CAAA;IAEF,OAAO;QACL,eAAe;QACf,sBAAsB;QACtB,mBAAmB;KACpB,CAAA;AACH,CAAC;AAED,SAAS,UAAU,CAAC,sBAAkC;IACpD,MAAM,mBAAmB,GAAG,IAAI,GAAG,CAAiB;QAClD,CAAC,4BAAS,EAAE,MAAO,CAAC;QACpB,CAAC,4BAAS,EAAE,MAAO,CAAC;QACpB,CAAC,4BAAS,EAAE,MAAO,CAAC;QACpB,CAAC,4BAAS,EAAE,MAAO,CAAC;QACpB,CAAC,8BAAW,EAAE,MAAO,CAAC;KACvB,CAAC,CAAA;IAEF,OAAO;QACL,eAAe,EACb,uHAAuH;QACzH,sBAAsB;QACtB,mBAAmB;KACpB,CAAA;AACH,CAAC;AAOD,SAAgB,cAAc,CAC5B,YAA0B,EAC1B,iBAAsC;IAEtC,MAAM,sBAAsB,GAAG,iBAAiB,CAAC,YAAY,CAAC,CAAA;IAC9D,QAAQ,YAAY,CAAC,WAAW,EAAE,EAAE;QAClC,KAAK,WAAW;YACd,OAAO,QAAQ,CAAC,sBAAsB,CAAC,CAAA;QAEzC,KAAK,WAAW;YACd,OAAO,QAAQ,CAAC,sBAAsB,CAAC,CAAA;QAEzC,KAAK,WAAW;YACd,OAAO,QAAQ,CAAC,sBAAsB,CAAC,CAAA;QAEzC,KAAK,aAAa;YAChB,OAAO,UAAU,CAAC,sBAAsB,CAAC,CAAA;QAE3C;YACE,MAAM,IAAI,KAAK,CAAC,0BAA0B,YAAY,EAAE,CAAC,CAAA;KAC5D;AACH,CAAC;AArBD,wCAqBC;AAEM,KAAK,UAAU,mBAAmB,CACvC,YAA0B,EAC1B,iBAA2C;IAE3C,MAAM,sBAAsB,GAAG,MAAM,iBAAiB,CAAC,YAAY,CAAC,CAAA;IACpE,OAAO,cAAc,CAAC,YAAY,EAAE,GAAG,EAAE,CAAC,sBAAsB,CAAC,CAAA;AACnE,CAAC;AAND,kDAMC"}
1
+ {"version":3,"file":"modelParams.js","sourceRoot":"","sources":["../src/modelParams.ts"],"names":[],"mappings":";;;AAGA,yDAS2B;AAoB3B,MAAM,eAAe,GACnB,8EAA8E,CAAA;AAEhF,SAAS,QAAQ,CAAC,sBAAkC;IAClD,OAAO;QACL,sBAAsB,EAAE,KAAM;QAC9B,eAAe;QACf,sBAAsB;QACtB,mBAAmB,EAAE,IAAI,GAAG,CAAiB,CAAC,CAAC,4BAAS,EAAE,KAAM,CAAC,CAAC,CAAC;KACpE,CAAA;AACH,CAAC;AAED,SAAS,QAAQ,CAAC,sBAAkC;IAClD,OAAO;QACL,sBAAsB,EAAE,KAAM;QAC9B,eAAe;QACf,sBAAsB;QACtB,mBAAmB,EAAE,IAAI,GAAG,CAAiB,CAAC,CAAC,4BAAS,EAAE,KAAM,CAAC,CAAC,CAAC;KACpE,CAAA;AACH,CAAC;AAED,SAAS,QAAQ,CAAC,sBAAkC;IAClD,MAAM,mBAAmB,GAAG,IAAI,GAAG,CAAiB;QAClD,CAAC,4BAAS,EAAE,KAAM,CAAC;QACnB,CAAC,4BAAS,EAAE,KAAM,CAAC;QACnB,CAAC,4BAAS,EAAE,KAAM,CAAC;QACnB,CAAC,4BAAS,EAAE,KAAM,CAAC;KACpB,CAAC,CAAA;IAEF,OAAO;QACL,eAAe;QACf,sBAAsB;QACtB,mBAAmB;KACpB,CAAA;AACH,CAAC;AAED,SAAS,UAAU,CAAC,sBAAkC;IACpD,MAAM,mBAAmB,GAAG,IAAI,GAAG,CAAiB;QAClD,CAAC,4BAAS,EAAE,MAAO,CAAC;QACpB,CAAC,4BAAS,EAAE,MAAO,CAAC;QACpB,CAAC,4BAAS,EAAE,MAAO,CAAC;QACpB,CAAC,4BAAS,EAAE,MAAO,CAAC;QACpB,CAAC,0BAAO,EAAE,MAAO,CAAC;QAClB,CAAC,wBAAK,EAAE,MAAO,CAAC;QAChB,CAAC,wBAAK,EAAE,MAAO,CAAC;QAChB,CAAC,8BAAW,EAAE,MAAO,CAAC;KACvB,CAAC,CAAA;IAEF,OAAO;QACL,eAAe,EACb,uHAAuH;QACzH,sBAAsB;QACtB,mBAAmB;KACpB,CAAA;AACH,CAAC;AAOD,SAAgB,iBAAiB,CAC/B,YAA0B,EAC1B,iBAAsC;IAEtC,MAAM,sBAAsB,GAAG,iBAAiB,CAAC,YAAY,CAAC,CAAA;IAC9D,QAAQ,YAAY,CAAC,WAAW,EAAE,EAAE;QAClC,KAAK,WAAW;YACd,OAAO,QAAQ,CAAC,sBAAsB,CAAC,CAAA;QAEzC,KAAK,WAAW;YACd,OAAO,QAAQ,CAAC,sBAAsB,CAAC,CAAA;QAEzC,KAAK,WAAW;YACd,OAAO,QAAQ,CAAC,sBAAsB,CAAC,CAAA;QAEzC,KAAK,aAAa;YAChB,OAAO,UAAU,CAAC,sBAAsB,CAAC,CAAA;QAE3C;YACE,MAAM,IAAI,KAAK,CAAC,0BAA0B,YAAY,EAAE,CAAC,CAAA;KAC5D;AACH,CAAC;AArBD,8CAqBC;AAEM,KAAK,UAAU,mBAAmB,CACvC,YAA0B,EAC1B,iBAA2C;IAE3C,MAAM,sBAAsB,GAAG,MAAM,iBAAiB,CAAC,YAAY,CAAC,CAAA;IACpE,OAAO,iBAAiB,CAAC,YAAY,EAAE,GAAG,EAAE,CAAC,sBAAsB,CAAC,CAAA;AACtE,CAAC;AAND,kDAMC"}
@@ -2,4 +2,7 @@ export declare const EndOfText = "<|endoftext|>";
2
2
  export declare const FimPrefix = "<|fim_prefix|>";
3
3
  export declare const FimMiddle = "<|fim_middle|>";
4
4
  export declare const FimSuffix = "<|fim_suffix|>";
5
+ export declare const ImStart = "<|im_start|>";
6
+ export declare const ImEnd = "<|im_end|>";
7
+ export declare const ImSep = "<|im_sep|>";
5
8
  export declare const EndOfPrompt = "<|endofprompt|>";
@@ -1,9 +1,12 @@
1
1
  "use strict";
2
2
  Object.defineProperty(exports, "__esModule", { value: true });
3
- exports.EndOfPrompt = exports.FimSuffix = exports.FimMiddle = exports.FimPrefix = exports.EndOfText = void 0;
3
+ exports.EndOfPrompt = exports.ImSep = exports.ImEnd = exports.ImStart = exports.FimSuffix = exports.FimMiddle = exports.FimPrefix = exports.EndOfText = void 0;
4
4
  exports.EndOfText = '<|endoftext|>';
5
5
  exports.FimPrefix = '<|fim_prefix|>';
6
6
  exports.FimMiddle = '<|fim_middle|>';
7
7
  exports.FimSuffix = '<|fim_suffix|>';
8
+ exports.ImStart = '<|im_start|>'; // 100264
9
+ exports.ImEnd = '<|im_end|>'; // 100265
10
+ exports.ImSep = '<|im_sep|>'; // 100266
8
11
  exports.EndOfPrompt = '<|endofprompt|>';
9
12
  //# sourceMappingURL=specialTokens.js.map
@@ -1 +1 @@
1
- {"version":3,"file":"specialTokens.js","sourceRoot":"","sources":["../src/specialTokens.ts"],"names":[],"mappings":";;;AAAa,QAAA,SAAS,GAAG,eAAe,CAAA;AAC3B,QAAA,SAAS,GAAG,gBAAgB,CAAA;AAC5B,QAAA,SAAS,GAAG,gBAAgB,CAAA;AAC5B,QAAA,SAAS,GAAG,gBAAgB,CAAA;AAC5B,QAAA,WAAW,GAAG,iBAAiB,CAAA"}
1
+ {"version":3,"file":"specialTokens.js","sourceRoot":"","sources":["../src/specialTokens.ts"],"names":[],"mappings":";;;AAAa,QAAA,SAAS,GAAG,eAAe,CAAA;AAC3B,QAAA,SAAS,GAAG,gBAAgB,CAAA;AAC5B,QAAA,SAAS,GAAG,gBAAgB,CAAA;AAC5B,QAAA,SAAS,GAAG,gBAAgB,CAAA;AAC5B,QAAA,OAAO,GAAG,cAAc,CAAA,CAAC,SAAS;AAClC,QAAA,KAAK,GAAG,YAAY,CAAA,CAAC,SAAS;AAC9B,QAAA,KAAK,GAAG,YAAY,CAAA,CAAC,SAAS;AAC9B,QAAA,WAAW,GAAG,iBAAiB,CAAA"}
package/cjs/util.d.ts ADDED
@@ -0,0 +1,2 @@
1
+ export declare function getMaxValueFromMap(map: Map<unknown, number>): number;
2
+ export declare function getSpecialTokenRegex(tokens: Set<string>): RegExp;
package/cjs/util.js ADDED
@@ -0,0 +1,19 @@
1
+ "use strict";
2
+ Object.defineProperty(exports, "__esModule", { value: true });
3
+ exports.getSpecialTokenRegex = exports.getMaxValueFromMap = void 0;
4
+ const escapeRegExp_js_1 = require("./escapeRegExp.js");
5
+ function getMaxValueFromMap(map) {
6
+ let max = 0;
7
+ map.forEach((val) => {
8
+ max = Math.max(max, val);
9
+ });
10
+ return max;
11
+ }
12
+ exports.getMaxValueFromMap = getMaxValueFromMap;
13
+ function getSpecialTokenRegex(tokens) {
14
+ const escapedTokens = [...tokens].map(escapeRegExp_js_1.escapeRegExp);
15
+ const inner = escapedTokens.join('|');
16
+ return new RegExp(`(${inner})`);
17
+ }
18
+ exports.getSpecialTokenRegex = getSpecialTokenRegex;
19
+ //# sourceMappingURL=util.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"util.js","sourceRoot":"","sources":["../src/util.ts"],"names":[],"mappings":";;;AAAA,uDAAgD;AAEhD,SAAgB,kBAAkB,CAAC,GAAyB;IAC1D,IAAI,GAAG,GAAG,CAAC,CAAA;IACX,GAAG,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE;QAClB,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,GAAG,CAAC,CAAA;IAC1B,CAAC,CAAC,CAAA;IACF,OAAO,GAAG,CAAA;AACZ,CAAC;AAND,gDAMC;AAED,SAAgB,oBAAoB,CAAC,MAAmB;IACtD,MAAM,aAAa,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC,GAAG,CAAC,8BAAY,CAAC,CAAA;IACnD,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,CAAC,GAAG,CAAC,CAAA;IACrC,OAAO,IAAI,MAAM,CAAC,IAAI,KAAK,GAAG,CAAC,CAAA;AACjC,CAAC;AAJD,oDAIC"}