MCPcopy Create free account
hub / github.com/catboost/catboost / BuildBpe

Method BuildBpe

library/cpp/text_processing/app_helpers/app_helpers.cpp:196–227  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

194}
195
196TIntrusivePtr<TBpeDictionary> NTextProcessing::NDictionary::BuildBpe(
197 const TString& inputFilePath,
198 const TDictionaryBuilderOptions& dictionaryBuilderOptions,
199 const TDictionaryOptions& dictionaryOptions,
200 const TBpeDictionaryOptions& bpeOptions,
201 const TTokenizerOptions& tokenizerOptions,
202 bool useTokenizer,
203 bool verbose
204) {
205 if (dictionaryOptions.TokenLevelType == ETokenLevelType::Word || bpeOptions.ContextLevel == EContextLevel::Sentence) {
206 return BuildBpeWord(
207 dictionaryBuilderOptions,
208 dictionaryOptions,
209 bpeOptions,
210 inputFilePath,
211 tokenizerOptions,
212 useTokenizer,
213 verbose
214 );
215 } else {
216 Y_ASSERT(dictionaryOptions.TokenLevelType == ETokenLevelType::Letter);
217 return BuildBpeLetter(
218 dictionaryBuilderOptions,
219 dictionaryOptions,
220 bpeOptions,
221 inputFilePath,
222 tokenizerOptions,
223 useTokenizer,
224 verbose
225 );
226 }
227}
228
229void NTextProcessing::NDictionary::ApplyDictionaryToFile(
230 const TString& inputPath,

Callers

nothing calls this directly

Calls 2

BuildBpeWordFunction · 0.85
BuildBpeLetterFunction · 0.85

Tested by

no test coverage detected