kodama-model
The model files Kodama downloads, laid out as Kodama keeps them:
<engine>/<model>/… for a model and _assets/<engine>/<asset>/… for a file several models share.
Kodama tries this repository first and falls back to each model's official repository, file by file;
every file is checked against the SHA-256 pinned in Kodama's catalogue.
kodama config set downloads.repository kodama-model
MIRROR.json lists every file with its size, SHA-256 and origin (repository, revision, path).
Files marked archive there are kept as published and unpacked by Kodama after download.
Licences
The files are unmodified copies; each keeps the licence of the repository it comes from. Using them means
accepting those licences (LICENSES.md). In particular:
- f5tts (
f5tts/): MIT (code); CC-BY-NC-4.0 (pretrained weights, trained on Emilia). The pretrained F5-TTS weights are CC-BY-NC-4.0: not for commercial use. - fishspeech (
fishspeech/): Fish Audio Research License (fishspeech/LICENSE.md,fishspeech/NOTICE.txt). Not for commercial use. Built with Fish Audio. - indextts (
indextts/): bilibili Model Use License (indextts/LICENSE.txt). Cloning a voice without its owner's permission is forbidden by the IndexTTS disclaimer.
Contents (82.7 GB)
| Folder | Engine licence | Files | Size | Origin |
|---|---|---|---|---|
asr/whisper-large-v3-turbo |
MIT (Whisper weights) | 11 | 1.62 GB | openai/whisper-large-v3-turbo@41f01f3fe87f |
cosyvoice/cosyvoice3-0.5b |
Apache License 2.0 | 13 | 4.44 GB | FunAudioLLM/Fun-CosyVoice3-0.5B-2512@29e01c4e8d00 |
cosyvoice/cosyvoice2-0.5b |
Apache License 2.0 | 13 | 3.09 GB | FunAudioLLM/CosyVoice2-0.5B@eec1ae6c7987 |
cosyvoice/cosyvoice-300m |
Apache License 2.0 | 8 | 2.30 GB | FunAudioLLM/CosyVoice-300M@24c40509c3c5 |
cosyvoice/cosyvoice-300m-sft |
Apache License 2.0 | 9 | 2.30 GB | FunAudioLLM/CosyVoice-300M-SFT@fbb71de2afe3 |
cosyvoice/cosyvoice-300m-instruct |
Apache License 2.0 | 9 | 2.30 GB | FunAudioLLM/CosyVoice-300M-Instruct@706bee1915e9 |
_assets/cosyvoice/wetext |
Apache License 2.0 | 4 | 0.01 GB | pengzhendong/wetext@030bb1febce0 |
f5tts/v1-base |
MIT (code); CC-BY-NC-4.0 (pretrained weights, trained on Emilia) | 4 | 1.40 GB | SWivid/F5-TTS@84e5a410d9ce, charactr/vocos-mel-24khz@0feb3fdd929b |
f5tts/base |
MIT (code); CC-BY-NC-4.0 (pretrained weights, trained on Emilia) | 4 | 1.40 GB | SWivid/F5-TTS@84e5a410d9ce, charactr/vocos-mel-24khz@0feb3fdd929b |
f5tts/e2-base |
MIT (code); CC-BY-NC-4.0 (pretrained weights, trained on Emilia) | 3 | 1.39 GB | SWivid/E2-TTS@851141880b5c, charactr/vocos-mel-24khz@0feb3fdd929b |
fishspeech/s2-pro |
Fish Audio Research License | 9 | 11.01 GB | fishaudio/s2-pro@1de9996b6be3 |
gptsovits/v2-base |
MIT License | 2 | 0.26 GB | lj1995/GPT-SoVITS@8f8857563c04 |
gptsovits/v2pro-base |
MIT License | 2 | 0.32 GB | lj1995/GPT-SoVITS@8f8857563c04 |
gptsovits/v2proplus-base |
MIT License | 2 | 0.36 GB | lj1995/GPT-SoVITS@8f8857563c04 |
gptsovits/v4-base |
MIT License | 3 | 0.98 GB | lj1995/GPT-SoVITS@8f8857563c04 |
gptsovits/v5turbo-base |
MIT License | 3 | 0.98 GB | lj1995/GPT-SoVITS@8f8857563c04 |
_assets/gptsovits/encoders |
MIT License | 8 | 1.08 GB | lj1995/GPT-SoVITS@8f8857563c04, XXXXRT/GPT-SoVITS-Pretrained@0c47645e02a7 |
_assets/gptsovits/g2pw |
MIT License | 1 | 0.59 GB | XXXXRT/GPT-SoVITS-Pretrained@0c47645e02a7 |
_assets/gptsovits/nltk-data |
MIT License | 1 | 0.01 GB | XXXXRT/GPT-SoVITS-Pretrained@0c47645e02a7 |
_assets/gptsovits/open-jtalk |
MIT License | 1 | 0.02 GB | XXXXRT/GPT-SoVITS-Pretrained@0c47645e02a7 |
_assets/gptsovits/s2d-v2 |
MIT License | 1 | 0.09 GB | lj1995/GPT-SoVITS@8f8857563c04 |
_assets/gptsovits/s2d-v2pro |
MIT License | 1 | 0.13 GB | lj1995/GPT-SoVITS@8f8857563c04 |
_assets/gptsovits/s2d-v2proplus |
MIT License | 1 | 0.13 GB | lj1995/GPT-SoVITS@8f8857563c04 |
indextts/v1.5 |
bilibili Model Use License | 4 | 1.71 GB | IndexTeam/IndexTTS-1.5@25851a6036df |
indextts/v2 |
bilibili Model Use License | 18 | 5.90 GB | IndexTeam/IndexTTS-2@740dcaff3962 |
indextts/v2.5 |
bilibili Model Use License | 19 | 5.49 GB | IndexTeam/IndexTTS-2.5@c39ce5ba9815 |
_assets/indextts/w2v-bert |
bilibili Model Use License | 3 | 2.32 GB | facebook/w2v-bert-2.0@da985ba0987f |
_assets/indextts/semantic-codec |
bilibili Model Use License | 1 | 0.18 GB | amphion/MaskGCT@265c6cef0762 |
_assets/indextts/campplus |
bilibili Model Use License | 1 | 0.03 GB | funasr/campplus@e4b6ede7ce16 |
_assets/indextts/bigvgan |
bilibili Model Use License | 2 | 0.45 GB | nvidia/bigvgan_v2_22khz_80band_256x@633ff708ed5b |
omnivoice/omnivoice |
Apache License 2.0 | 9 | 3.27 GB | k2-fsa/OmniVoice@c5fdb5ccb189 |
qwen3tts/1.7b-base |
Apache License 2.0 | 11 | 4.54 GB | Qwen/Qwen3-TTS-12Hz-1.7B-Base@fd4b25438912 |
qwen3tts/0.6b-base |
Apache License 2.0 | 11 | 2.52 GB | Qwen/Qwen3-TTS-12Hz-0.6B-Base@5d83992436ea |
qwen3tts/1.7b-custom-voice |
Apache License 2.0 | 11 | 4.52 GB | Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice@0c0e3051f131 |
qwen3tts/0.6b-custom-voice |
Apache License 2.0 | 11 | 2.50 GB | Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice@85e237c12c02 |
qwen3tts/1.7b-voice-design |
Apache License 2.0 | 11 | 4.52 GB | Qwen/Qwen3-TTS-12Hz-1.7B-VoiceDesign@5ecdb67327fd |
voxcpm/2 |
Apache License 2.0 (code and weights) | 5 | 4.96 GB | openbmb/VoxCPM2@32279effe8c1 |
voxcpm/1.5 |
Apache License 2.0 (code and weights) | 5 | 1.95 GB | openbmb/VoxCPM1.5@8cdc403854cd |
voxcpm/0.5b |
Apache License 2.0 (code and weights) | 5 | 1.61 GB | openbmb/VoxCPM-0.5B@b2e656f75243 |