kokoro: unidic-lite trim + MeCab reference reader for the Swift Japanese frontend (FluidAudio #914) - #90
Open
Alex-Wengg wants to merge 1 commit into
Open
kokoro: unidic-lite trim + MeCab reference reader for the Swift Japanese frontend (FluidAudio #914)#90Alex-Wengg wants to merge 1 commit into
Alex-Wengg wants to merge 1 commit into
Conversation
Alex-Wengg
force-pushed
the
feat/kokoro-ja-unidic-trim
branch
from
September 13, 2026 17:49
0dc0a07 to
4040a39
Compare
…ese frontend (FluidAudio #914) convert_unidic_lite.py rewrites unidic-lite's feature strings to pos1,pron,kana (188 MB -> 41 MB) keeping the double array and token table intact; mecab_reference.py validates the trimmed dictionary against fugashi.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Companion to FluidInference/FluidAudio#915.
models/tts/kokoro/coreml/g2p/japanese/:convert_unidic_lite.pyrewrites unidic-lite's MeCab dictionary feature strings topos1,pron,kana(188 MB → 41 MB) while keeping the double array and token table byte-for-byte, and applies the same trim tounk.dic;matrix.binandchar.binare copied.mecab_reference.pyis a pure-Python reader of the MeCab binary layout with a MeCab-style Viterbi; it printsSAME/DIFFper sentence against fugashi so a trimmed dictionary can be validated before publishing. The SwiftJapaneseTokenizerin FluidAudio mirrors it.README.mddocuments the build, validation and publication (FluidInference/kokoro-82m-coreml/ANE-ja/assets/, uploaded as HF commit c9fdfdac).