FluidInference/kokoro-82m-coreml

es_lexicon_cache.json: regenerate against the shipped Swift rules (adds ser, coda-d, 2,145 proper nouns; 51,539 exceptions) (FluidAudio #950)

fr_lexicon_cache.json: add espeak-ng fr-fr pronunciations for 9,043 proper nouns missing from ipa-dict (FluidAudio #950)

Add fr/es lexicon caches for Kokoro ANE Spanish/French variants (FluidAudio #926)

Add Mandarin jieba HMM segmentation tables

ANE-ja: Japanese G2P assets (unidic-lite trimmed to pos1,pron,kana + Cutlet ja_words) for FluidAudio #914

Add COLA-corrected KokoroTail_v2 for ANE/, ANE-ja/, ANE-zh/ (issue #852)

fix(vocoder): rank-stable anchor for macOS 14 Espresso (FluidAudio#836)

Upload 7 files

Upload 8 files

Upload 8 files

Upload 41 files

Upload 9 files

Upload 2 files

Upload weight.bin

Upload model.mlmodel

Delete ANE-zh/KokoroNoise.mlpackage/Manifest.json.txt

Update ANE-zh/KokoroNoise.mlpackage/Manifest.json

Upload Manifest.json.txt

Upload coremldata.bin

Upload weight.bin

Upload 3 files

Upload coremldata.bin

Upload 3 files

Delete ANE-zh/metadata.json

Delete ANE-zh/coremldata.bin

Delete ANE-zh/model.mil

Upload 3 files

Upload 103 files

Upload 62 files

Delete ANE/ANE-zh

Upload 62 files

Update README.md

Upload 61 files

Upload 16 files

Update config.json

Upload 16 files

g2p-vocab (#3)

g2p-model (#2)

Upload 8 files

Upload 5 files

Upload 5 files

Upload 6 files

Update README.md

Delete word_phonemes.json

Delete word_frames_phonemes.json

Delete ESpeakNG.xcframework.zip

Delete kokoro_completev21.mlpackage

Delete kokoro_completev21.mlmodelc

Delete espeak-ng.zip

Delete ESpeakNG.xcframework