Instructions to use Infinix-io/ZONOS3-Mini with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Zonos
How to use Infinix-io/ZONOS3-Mini with Zonos:
# pip install git+https://github.com/Zyphra/Zonos.git import torchaudio from zonos.model import Zonos from zonos.conditioning import make_cond_dict model = Zonos.from_pretrained("Infinix-io/ZONOS3-Mini", device="cuda") wav, sr = torchaudio.load("speaker.wav") # 5-10s reference clip speaker = model.make_speaker_embedding(wav, sr) cond = make_cond_dict(text="Hello, world!", speaker=speaker, language="en-us") codes = model.generate(model.prepare_conditioning(cond)) audio = model.autoencoder.decode(codes)[0].cpu() torchaudio.save("sample.wav", audio, model.autoencoder.sampling_rate) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use Infinix-io/ZONOS3-Mini with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf Infinix-io/ZONOS3-Mini # Run inference directly in the terminal: llama cli -hf Infinix-io/ZONOS3-Mini
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf Infinix-io/ZONOS3-Mini # Run inference directly in the terminal: llama cli -hf Infinix-io/ZONOS3-Mini
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-app # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama serve -hf Infinix-io/ZONOS3-Mini # Run inference directly in the terminal: ./build/bin/llama cli -hf Infinix-io/ZONOS3-Mini
Use Docker
docker model run hf.co/Infinix-io/ZONOS3-Mini
- Llama new
- LM Studio
- Jan
- Ollama
How to use Infinix-io/ZONOS3-Mini with Ollama:
ollama run hf.co/Infinix-io/ZONOS3-Mini
- Unsloth Desktop
- Docker Model Runner
How to use Infinix-io/ZONOS3-Mini with Docker Model Runner:
docker model run hf.co/Infinix-io/ZONOS3-Mini
- Lemonade
How to use Infinix-io/ZONOS3-Mini with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull Infinix-io/ZONOS3-Mini
Run and chat with the model
lemonade run user.ZONOS3-Mini-{{QUANT_TAG}}List all available models
lemonade list
- Atomic Chat
Commit ·
ab402f8
0
Parent(s):
Duplicate from Zyphra/ZONOS3-Mini
Browse filesCo-authored-by: Gabriel Clark <gabrielclark3330@users.noreply.huggingface.co>
This view is limited to 50 files because it contains too many changes. See raw diff
- .gitattributes +164 -0
- README.md +204 -0
- assets/arch_zac1_dark.png +3 -0
- assets/eval_pareto.png +3 -0
- assets/zonos3_mini_arloop_animated.gif +3 -0
- assets/zonos3_mini_hero.webp +3 -0
- autoencoder/generator.pt +3 -0
- autoencoder/heads/full_mr/config_full_unfreezeall_mr.json +455 -0
- autoencoder/heads/full_mr/step_0050000.pt +3 -0
- autoencoder/heads/keep2_plain/config_full_unfreezeall_mr_drop30_4x10.json +473 -0
- autoencoder/heads/keep2_plain/step_0050000.pt +3 -0
- autoencoder/teacher/config.json +81 -0
- autoencoder/teacher/model.safetensors +3 -0
- gguf/backbone-ptq-edit.gguf +3 -0
- gguf/backbone-ptq.gguf +3 -0
- gguf/edit/accents.gguf +3 -0
- gguf/spk-encoder.gguf +3 -0
- gguf/ssl-teacher.gguf +3 -0
- gguf/steer.gguf +3 -0
- gguf/voices/English/Cillian_en-Irish_M_1.codes.npy +3 -0
- gguf/voices/English/Cillian_en-Irish_M_1.embedding.npy +3 -0
- gguf/voices/English/Cillian_en-Irish_M_1.wav +3 -0
- gguf/voices/English/Imogen_en-British_F_1.codes.npy +3 -0
- gguf/voices/English/Imogen_en-British_F_1.embedding.npy +3 -0
- gguf/voices/English/Imogen_en-British_F_1.wav +3 -0
- gguf/voices/English/Jett_en-energetic_M_1.codes.npy +3 -0
- gguf/voices/English/Jett_en-energetic_M_1.embedding.npy +3 -0
- gguf/voices/English/Jett_en-energetic_M_1.wav +3 -0
- gguf/voices/English/Morgana_en-witch_F_1.codes.npy +3 -0
- gguf/voices/English/Morgana_en-witch_F_1.embedding.npy +3 -0
- gguf/voices/English/Morgana_en-witch_F_1.mp3 +3 -0
- gguf/voices/English/Rupert_en-British_M_1.codes.npy +3 -0
- gguf/voices/English/Rupert_en-British_M_1.embedding.npy +3 -0
- gguf/voices/English/Rupert_en-British_M_1.wav +3 -0
- gguf/voices/English/Sakura_en-anime_F_1.codes.npy +3 -0
- gguf/voices/English/Sakura_en-anime_F_1.embedding.npy +3 -0
- gguf/voices/English/Sakura_en-anime_F_1.wav +3 -0
- gguf/voices/English/Savannah_en-American_F_1.codes.npy +3 -0
- gguf/voices/English/Savannah_en-American_F_1.embedding.npy +3 -0
- gguf/voices/English/Savannah_en-American_F_1.mp3 +3 -0
- gguf/voices/English/Thaddeus_en-announcervoice_M_1.codes.npy +3 -0
- gguf/voices/English/Thaddeus_en-announcervoice_M_1.embedding.npy +3 -0
- gguf/voices/English/Thaddeus_en-announcervoice_M_1.mp3 +3 -0
- gguf/voices/English/Willow_en-asrm_F_1.codes.npy +3 -0
- gguf/voices/English/Willow_en-asrm_F_1.embedding.npy +3 -0
- gguf/voices/English/Willow_en-asrm_F_1.wav +3 -0
- gguf/voices/English/Wyatt_en-American_M_1.codes.npy +3 -0
- gguf/voices/English/Wyatt_en-American_M_1.embedding.npy +3 -0
- gguf/voices/English/Wyatt_en-American_M_1.ogg +3 -0
- gguf/voices/French/Gaspard_fr_M_1.codes.npy +3 -0
.gitattributes
ADDED
|
@@ -0,0 +1,164 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 8 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 9 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 10 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 11 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 12 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
| 13 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 14 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 15 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 16 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 17 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
| 18 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
| 19 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
| 20 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
| 21 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
| 22 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
| 23 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
| 24 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
| 25 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
| 26 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
| 27 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
| 28 |
+
*.tar filter=lfs diff=lfs merge=lfs -text
|
| 29 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
| 30 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
| 31 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
| 32 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
| 33 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
assets/eval_itw.png filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
assets/eval_clean.png filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
assets/eval_ranking.png filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
assets/arch_zonos3_mini.png filter=lfs diff=lfs merge=lfs -text
|
| 40 |
+
assets/arch_zac1.png filter=lfs diff=lfs merge=lfs -text
|
| 41 |
+
assets/zonos3_mini_arloop_animated.gif filter=lfs diff=lfs merge=lfs -text
|
| 42 |
+
assets/arch_zac1_dark.png filter=lfs diff=lfs merge=lfs -text
|
| 43 |
+
assets/eval_pareto.png filter=lfs diff=lfs merge=lfs -text
|
| 44 |
+
assets/zonos3_mini_hero.webp filter=lfs diff=lfs merge=lfs -text
|
| 45 |
+
gguf/edit/accents.gguf filter=lfs diff=lfs merge=lfs -text
|
| 46 |
+
gguf/spk-encoder.gguf filter=lfs diff=lfs merge=lfs -text
|
| 47 |
+
gguf/ssl-teacher.gguf filter=lfs diff=lfs merge=lfs -text
|
| 48 |
+
gguf/steer.gguf filter=lfs diff=lfs merge=lfs -text
|
| 49 |
+
gguf/voices/Ananya_hi_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 50 |
+
gguf/voices/Arjun_hi_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 51 |
+
gguf/voices/Chaitanya_te_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 52 |
+
gguf/voices/Cillian_en-Irish_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 53 |
+
gguf/voices/Gaspard_fr_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 54 |
+
gguf/voices/Ginevra_it_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 55 |
+
gguf/voices/Haoran_zh-energetic_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 56 |
+
gguf/voices/Imogen_en-British_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 57 |
+
gguf/voices/Jett_en-energetic_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 58 |
+
gguf/voices/Leonor_pt_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 59 |
+
gguf/voices/Liesel_de_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 60 |
+
gguf/voices/Lorenzo_it_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 61 |
+
gguf/voices/Meilin_zh-energetic_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 62 |
+
gguf/voices/Minjun_ko_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 63 |
+
gguf/voices/Morgana_en-witch_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 64 |
+
gguf/voices/Rodrigo_es_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 65 |
+
gguf/voices/Rupert_en-British_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 66 |
+
gguf/voices/Sakura_en-anime_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 67 |
+
gguf/voices/Savannah_en-American_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 68 |
+
gguf/voices/Seoyeon_ko_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 69 |
+
gguf/voices/Sravya_te_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 70 |
+
gguf/voices/Thaddeus_en-announcervoice_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 71 |
+
gguf/voices/Thiago_pt_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 72 |
+
gguf/voices/Willow_en-asrm_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 73 |
+
gguf/voices/Wolfgang_de_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 74 |
+
gguf/voices/Wyatt_en-American_M_1.ogg filter=lfs diff=lfs merge=lfs -text
|
| 75 |
+
gguf/voices/Ximena_es_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 76 |
+
gguf/zac-encoder.gguf filter=lfs diff=lfs merge=lfs -text
|
| 77 |
+
gguf/zac-keep2.gguf filter=lfs diff=lfs merge=lfs -text
|
| 78 |
+
gguf/zac.gguf filter=lfs diff=lfs merge=lfs -text
|
| 79 |
+
gguf/zonos3-ptq-edit.gguf filter=lfs diff=lfs merge=lfs -text
|
| 80 |
+
gguf/zonos3-ptq.gguf filter=lfs diff=lfs merge=lfs -text
|
| 81 |
+
web/companions/steer.gguf filter=lfs diff=lfs merge=lfs -text
|
| 82 |
+
web/voices/Ananya_hi_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 83 |
+
web/voices/Arjun_hi_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 84 |
+
web/voices/Chaitanya_te_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 85 |
+
web/voices/Cillian_en-Irish_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 86 |
+
web/voices/Gaspard_fr_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 87 |
+
web/voices/Ginevra_it_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 88 |
+
web/voices/Haoran_zh-energetic_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 89 |
+
web/voices/Imogen_en-British_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 90 |
+
web/voices/Jett_en-energetic_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 91 |
+
web/voices/Leonor_pt_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 92 |
+
web/voices/Liesel_de_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 93 |
+
web/voices/Lorenzo_it_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 94 |
+
web/voices/Meilin_zh-energetic_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 95 |
+
web/voices/Minjun_ko_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 96 |
+
web/voices/Morgana_en-witch_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 97 |
+
web/voices/Rodrigo_es_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 98 |
+
web/voices/Rupert_en-British_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 99 |
+
web/voices/Sakura_en-anime_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 100 |
+
web/voices/Savannah_en-American_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 101 |
+
web/voices/Seoyeon_ko_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 102 |
+
web/voices/Sravya_te_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 103 |
+
web/voices/Thaddeus_en-announcervoice_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 104 |
+
web/voices/Thiago_pt_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 105 |
+
web/voices/Willow_en-asrm_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 106 |
+
web/voices/Wolfgang_de_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 107 |
+
web/voices/Wyatt_en-American_M_1.ogg filter=lfs diff=lfs merge=lfs -text
|
| 108 |
+
web/voices/Ximena_es_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 109 |
+
gguf/voices/English/Cillian_en-Irish_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 110 |
+
gguf/voices/English/Imogen_en-British_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 111 |
+
gguf/voices/English/Jett_en-energetic_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 112 |
+
gguf/voices/English/Morgana_en-witch_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 113 |
+
gguf/voices/English/Rupert_en-British_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 114 |
+
gguf/voices/English/Sakura_en-anime_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 115 |
+
gguf/voices/English/Savannah_en-American_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 116 |
+
gguf/voices/English/Thaddeus_en-announcervoice_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 117 |
+
gguf/voices/English/Willow_en-asrm_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 118 |
+
gguf/voices/English/Wyatt_en-American_M_1.ogg filter=lfs diff=lfs merge=lfs -text
|
| 119 |
+
gguf/voices/French/Gaspard_fr_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 120 |
+
gguf/voices/German/Liesel_de_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 121 |
+
gguf/voices/German/Wolfgang_de_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 122 |
+
gguf/voices/Hindi/Ananya_hi_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 123 |
+
gguf/voices/Hindi/Arjun_hi_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 124 |
+
gguf/voices/Italian/Ginevra_it_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 125 |
+
gguf/voices/Italian/Lorenzo_it_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 126 |
+
gguf/voices/Korean/Minjun_ko_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 127 |
+
gguf/voices/Korean/Seoyeon_ko_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 128 |
+
gguf/voices/Mandarin[[:space:]]Chinese/Haoran_zh-energetic_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 129 |
+
gguf/voices/Mandarin[[:space:]]Chinese/Meilin_zh-energetic_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 130 |
+
gguf/voices/Portuguese/Leonor_pt_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 131 |
+
gguf/voices/Portuguese/Thiago_pt_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 132 |
+
gguf/voices/Spanish/Rodrigo_es_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 133 |
+
gguf/voices/Spanish/Ximena_es_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 134 |
+
gguf/voices/Telugu/Chaitanya_te_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 135 |
+
gguf/voices/Telugu/Sravya_te_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 136 |
+
web/voices/English/Cillian_en-Irish_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 137 |
+
web/voices/English/Imogen_en-British_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 138 |
+
web/voices/English/Jett_en-energetic_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 139 |
+
web/voices/English/Morgana_en-witch_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 140 |
+
web/voices/English/Rupert_en-British_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 141 |
+
web/voices/English/Sakura_en-anime_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 142 |
+
web/voices/English/Savannah_en-American_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 143 |
+
web/voices/English/Thaddeus_en-announcervoice_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 144 |
+
web/voices/English/Willow_en-asrm_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 145 |
+
web/voices/English/Wyatt_en-American_M_1.ogg filter=lfs diff=lfs merge=lfs -text
|
| 146 |
+
web/voices/French/Gaspard_fr_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 147 |
+
web/voices/German/Liesel_de_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 148 |
+
web/voices/German/Wolfgang_de_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 149 |
+
web/voices/Hindi/Ananya_hi_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 150 |
+
web/voices/Hindi/Arjun_hi_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 151 |
+
web/voices/Italian/Ginevra_it_F_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 152 |
+
web/voices/Italian/Lorenzo_it_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 153 |
+
web/voices/Korean/Minjun_ko_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 154 |
+
web/voices/Korean/Seoyeon_ko_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 155 |
+
web/voices/Mandarin[[:space:]]Chinese/Haoran_zh-energetic_M_1.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 156 |
+
web/voices/Mandarin[[:space:]]Chinese/Meilin_zh-energetic_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 157 |
+
web/voices/Portuguese/Leonor_pt_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 158 |
+
web/voices/Portuguese/Thiago_pt_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 159 |
+
web/voices/Spanish/Rodrigo_es_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 160 |
+
web/voices/Spanish/Ximena_es_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 161 |
+
web/voices/Telugu/Chaitanya_te_M_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 162 |
+
web/voices/Telugu/Sravya_te_F_1.wav filter=lfs diff=lfs merge=lfs -text
|
| 163 |
+
gguf/backbone-ptq.gguf filter=lfs diff=lfs merge=lfs -text
|
| 164 |
+
gguf/backbone-ptq-edit.gguf filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,204 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
pipeline_tag: text-to-speech
|
| 4 |
+
library_name: zonos3
|
| 5 |
+
language:
|
| 6 |
+
- en
|
| 7 |
+
- fr
|
| 8 |
+
- es
|
| 9 |
+
- de
|
| 10 |
+
- it
|
| 11 |
+
- pt
|
| 12 |
+
- nl
|
| 13 |
+
- pl
|
| 14 |
+
- ru
|
| 15 |
+
- sv
|
| 16 |
+
- tr
|
| 17 |
+
- uk
|
| 18 |
+
- cs
|
| 19 |
+
- da
|
| 20 |
+
- fi
|
| 21 |
+
- el
|
| 22 |
+
- hu
|
| 23 |
+
- "no"
|
| 24 |
+
- ro
|
| 25 |
+
- sk
|
| 26 |
+
- bg
|
| 27 |
+
- zh
|
| 28 |
+
- ja
|
| 29 |
+
- ko
|
| 30 |
+
- ar
|
| 31 |
+
- he
|
| 32 |
+
- vi
|
| 33 |
+
- id
|
| 34 |
+
- th
|
| 35 |
+
- tl
|
| 36 |
+
- ms
|
| 37 |
+
- ka
|
| 38 |
+
- hi
|
| 39 |
+
- bn
|
| 40 |
+
- ta
|
| 41 |
+
- te
|
| 42 |
+
- mr
|
| 43 |
+
- gu
|
| 44 |
+
- kn
|
| 45 |
+
- ml
|
| 46 |
+
- pa
|
| 47 |
+
- or
|
| 48 |
+
- ur
|
| 49 |
+
tags:
|
| 50 |
+
- text-to-speech
|
| 51 |
+
- tts
|
| 52 |
+
- speech-synthesis
|
| 53 |
+
- voice-cloning
|
| 54 |
+
- multilingual
|
| 55 |
+
- gguf
|
| 56 |
+
- zonos
|
| 57 |
+
- zonos3
|
| 58 |
+
- zyphra
|
| 59 |
+
---
|
| 60 |
+
|
| 61 |
+
# ZONOS3-Mini
|
| 62 |
+
|
| 63 |
+
<p align="center">
|
| 64 |
+
<img src="./assets/zonos3_mini_hero.webp" alt="ZONOS3-Mini title card" width="750" />
|
| 65 |
+
</p>
|
| 66 |
+
|
| 67 |
+
<div align="center">
|
| 68 |
+
<a href="https://arxiv.org/abs/2611.16525" target="_blank">
|
| 69 |
+
<img src="https://img.shields.io/badge/arXiv-2611.16525-B31B1B?style=for-the-badge&logo=arxiv&logoColor=white" alt="arXiv">
|
| 70 |
+
</a>
|
| 71 |
+
<a href="https://github.com/Zyphra/ZONOS3-Mini" target="_blank">
|
| 72 |
+
<img src="https://img.shields.io/badge/GitHub-ZONOS3--Mini-181717?style=for-the-badge&logo=github&logoColor=white" alt="GitHub">
|
| 73 |
+
</a>
|
| 74 |
+
<a href="https://discord.gg/gTW9JwST8q" target="_blank">
|
| 75 |
+
<img src="https://img.shields.io/badge/Join%20Our%20Discord-7289DA?style=for-the-badge&logo=discord&logoColor=white" alt="Discord">
|
| 76 |
+
</a>
|
| 77 |
+
</div>
|
| 78 |
+
|
| 79 |
+
---
|
| 80 |
+
|
| 81 |
+
ZONOS3-Mini is Zyphra’s compact, open-source text-to-speech model, delivering natural reliable speech and high-fidelity voice
|
| 82 |
+
cloning with just 100 million active parameters. It runs in real time on phones and laptops, including directly in the browser.
|
| 83 |
+
|
| 84 |
+
Clone a voice from an audio sample, or design a character by shaping age, accent and vocal texture. Steer emotion, pace and
|
| 85 |
+
delivery down to individual words for games, narration and voice assistants.
|
| 86 |
+
|
| 87 |
+
Run ZONOS3-Mini locally with our browser and native runtimes, or access it through [Zyphra Cloud](https://cloud.zyphra.com/models/zyphra/ZONOS3-Mini). Model weights are released
|
| 88 |
+
under Apache 2.0.
|
| 89 |
+
|
| 90 |
+
**Model details, samples and evaluations are in the [blog post](https://www.zyphra.com/our-work/zonos3-mini).**
|
| 91 |
+
**A hosted version is available at [Zyphra Cloud](https://cloud.zyphra.com/models/zyphra/ZONOS3-Mini).**
|
| 92 |
+
|
| 93 |
+
## Languages
|
| 94 |
+
|
| 95 |
+
ZONOS3-Mini supports 43 languages.
|
| 96 |
+
|
| 97 |
+
| Region | Languages |
|
| 98 |
+
|---|---|
|
| 99 |
+
| Western and Southern Europe | English, French, Spanish, German, Italian, Portuguese, Dutch, Greek |
|
| 100 |
+
| Northern Europe | Swedish, Danish, Norwegian, Finnish |
|
| 101 |
+
| Central and Eastern Europe | Polish, Czech, Slovak, Hungarian, Romanian, Bulgarian, Russian, Ukrainian |
|
| 102 |
+
| Middle East and Caucasus | Turkish, Arabic, Hebrew, Georgian |
|
| 103 |
+
| East Asia | Mandarin Chinese, Japanese, Korean |
|
| 104 |
+
| Southeast Asia | Vietnamese, Indonesian, Malay, Thai, Tagalog |
|
| 105 |
+
| South Asia | Hindi, Bengali, Tamil, Telugu, Marathi, Gujarati, Kannada, Malayalam, Punjabi, Odia, Urdu |
|
| 106 |
+
|
| 107 |
+
## Quick start
|
| 108 |
+
|
| 109 |
+
### In your browser
|
| 110 |
+
|
| 111 |
+
ZONOS3-Mini runs entirely on your own GPU through WebGPU: no install and no server. The first visit downloads the model, and
|
| 112 |
+
the browser keeps it cached for later visits. You need a browser with WebGPU support.
|
| 113 |
+
|
| 114 |
+
- [Zonos Voice Studio](https://zyphra.com/zonos3-mini/localplayground#cast): the full studio.
|
| 115 |
+
- [Lite](https://zyphra.com/zonos3-mini/localplayground/lite): quick voice cloning. Pick, upload or record a voice and
|
| 116 |
+
speak with it.
|
| 117 |
+
|
| 118 |
+
### With SGLang
|
| 119 |
+
|
| 120 |
+
[SGLang-Omni](https://github.com/sgl-project/sglang-omni) serves ZONOS3-Mini on NVIDIA GPUs behind the OpenAI-compatible
|
| 121 |
+
`/v1/audio/speech` endpoint, with streaming and voice cloning from a reference clip (no transcript needed). Install SGLang-Omni
|
| 122 |
+
by following its [installation guide](https://github.com/sgl-project/sglang-omni/blob/main/docs/get_started/installation.md),
|
| 123 |
+
then start a server:
|
| 124 |
+
|
| 125 |
+
```bash
|
| 126 |
+
sgl-omni serve --model-path Zyphra/ZONOS3-Mini --port 8000
|
| 127 |
+
```
|
| 128 |
+
|
| 129 |
+
Once `curl http://localhost:8000/health` reports `"healthy"`, generate speech with the default voice from another terminal:
|
| 130 |
+
|
| 131 |
+
```bash
|
| 132 |
+
curl -X POST http://localhost:8000/v1/audio/speech \
|
| 133 |
+
-H "Content-Type: application/json" \
|
| 134 |
+
-d '{"input": "Hello from ZONOS3-Mini!"}' \
|
| 135 |
+
--output output.wav
|
| 136 |
+
```
|
| 137 |
+
|
| 138 |
+
Add `"references": [{"audio_path": "<reference clip>"}]` to clone a voice, or `"stream": true, "response_format": "pcm"` to
|
| 139 |
+
stream 44.1 kHz audio as it is generated. The [ZONOS3 cookbook](https://github.com/sgl-project/sglang-omni/blob/main/docs/cookbook/zonos3.md)
|
| 140 |
+
covers voice cloning, streaming and the sampling options.
|
| 141 |
+
|
| 142 |
+
### Native runtimes
|
| 143 |
+
|
| 144 |
+
The [ZONOS3-Mini GitHub repository](https://github.com/Zyphra/ZONOS3-Mini) holds Zyphra's own runtimes, which share one HTTP
|
| 145 |
+
API and serve the Zonos Voice Studio: a CUDA server for NVIDIA GPUs, an MLX server for Apple Silicon, a C++/ggml server and CLI
|
| 146 |
+
(CPU, CUDA, Metal or Vulkan), and the in-browser WebGPU build.
|
| 147 |
+
They load the GGUF files in [`gguf/`](https://huggingface.co/Zyphra/ZONOS3-Mini/tree/main/gguf) and the WebGPU shards in
|
| 148 |
+
[`web/`](https://huggingface.co/Zyphra/ZONOS3-Mini/tree/main/web); the 27 default voices, grouped by language, are in
|
| 149 |
+
[`gguf/voices/`](https://huggingface.co/Zyphra/ZONOS3-Mini/blob/main/gguf/voices/README.md).
|
| 150 |
+
|
| 151 |
+
## Architecture (rough outline)
|
| 152 |
+
|
| 153 |
+
Text is NeMo-normalized (English, Mandarin, Japanese, Korean, German, Spanish, French, Italian, Portuguese; other languages
|
| 154 |
+
pass through as written) and encoded with an 8,192-entry SuperBPE tokenizer. The language model autoregressively generates 15 Hz
|
| 155 |
+
dual-codec frames (10 codebooks of 4,096 entries each: one semantic codebook distilled from the Wav2Vec2-BERT teacher plus nine
|
| 156 |
+
acoustic codebooks), conditioned on the text, a speaker prompt and three bucketed acoustic controls. A causal multi-resolution
|
| 157 |
+
iSTFT head renders the frames to 44.1 kHz audio as they are produced.
|
| 158 |
+
|
| 159 |
+

|
| 160 |
+
|
| 161 |
+
*Inference pipeline. Speaker conditioning (raw reference codes or a speaker-embedding row), control rows and SuperBPE text enter
|
| 162 |
+
the hybrid GDN/attention MoE backbone, which predicts all ten codebooks of the next frame under the delay pattern (four drawn);
|
| 163 |
+
as soon as a frame's last codebook is generated, the codec's causal iSTFT head renders that frame to audio.*
|
| 164 |
+
|
| 165 |
+
| | ZONOS3-Mini |
|
| 166 |
+
|---|---|
|
| 167 |
+
| Backbone | 11 layers, model dim 1,024; 800M parameters total, about 100M active per token (latent MoE: 64 experts, top-2 routing, from layer 1) |
|
| 168 |
+
| Sequence mixer | pattern `gdn3_global1`: 9 Gated DeltaNet (GDN) layers + 2 global attention layers (16 heads x 32 dims, 4 KV heads), attention residuals (`full_attnres`); context 32,768 positions |
|
| 169 |
+
| Text | SuperBPE, 8,192 entries (+4 specials), NeMo text normalization at inference |
|
| 170 |
+
| Audio tokens | 15 Hz, 44.1 kHz dual-encoding codec, 10 x 4,096 codebooks (1 semantic + 9 acoustic); hop 2,940 samples; up to 10 minutes of audio per generation |
|
| 171 |
+
| Speaker prompt | raw reference clip (encoded by the codec into a token prefix, up to 60 s, with speaker CFG) **or** a single 192-d MECT-B2 speaker embedding (`ownmect_s2`) |
|
| 172 |
+
| Controls | loudness (LUFS, 10 buckets), bandwidth f_max (Hz, 10 buckets), **speaking rate (UTF-8 bytes per second, 30 buckets)**; each row has an "unknown" index; bucket edges in `params.json` (`condition_edges`) |
|
| 173 |
+
| Decoder | causal multi-resolution iSTFT head: `full_mr` (default) or `keep2_plain` (pruned, about 3x faster decode) |
|
| 174 |
+
| Precision | bf16 weights (`model.pth`); decoder heads run in fp32 |
|
| 175 |
+
|
| 176 |
+

|
| 177 |
+
|
| 178 |
+
*ZAC 1 (Zyphra Audio Codec 1).*
|
| 179 |
+
|
| 180 |
+
## Evaluations
|
| 181 |
+
|
| 182 |
+
Full results and the ZTTS2 capability index are in the [blog post](https://www.zyphra.com/our-work/zonos3-mini#evaluation).
|
| 183 |
+
|
| 184 |
+
Quality against speed on one H100 80GB. Quality combines macro speaker similarity and macro WER over the 55 language subsets of
|
| 185 |
+
[ZTTS2-Eval](https://zyphra.github.io/ZTTS2-Eval/) (benchmark version 2: FLEURS-R read speech in 35 languages and VoxBlink2 in-the-wild recordings in 20 languages) as
|
| 186 |
+
½ z(similarity) − ½ z(WER) across the systems shown; ZONOS3-Mini is measured with the speaker-embedding prompt. Left: median
|
| 187 |
+
time to first audio with 8 concurrent requests (80 ms for ZONOS3-Mini). Right: peak throughput at each model's best concurrency
|
| 188 |
+
(457 seconds of audio per second for ZONOS3-Mini).
|
| 189 |
+
|
| 190 |
+

|
| 191 |
+
|
| 192 |
+
## Citation
|
| 193 |
+
If you find this model useful in an academic context, please cite the technical report:
|
| 194 |
+
```bibtex
|
| 195 |
+
@misc{zyphra2026zonos3mini,
|
| 196 |
+
title = {{ZONOS3-Mini Technical Report}},
|
| 197 |
+
author = {Gabriel Clark and Sofian Mejjoute and George Close and Beren Millidge},
|
| 198 |
+
year = {2026},
|
| 199 |
+
eprint = {2611.16525},
|
| 200 |
+
archivePrefix = {arXiv},
|
| 201 |
+
primaryClass = {cs.SD},
|
| 202 |
+
url = {https://arxiv.org/abs/2611.16525},
|
| 203 |
+
}
|
| 204 |
+
```
|
assets/arch_zac1_dark.png
ADDED
|
Git LFS Details
|
assets/eval_pareto.png
ADDED
|
Git LFS Details
|
assets/zonos3_mini_arloop_animated.gif
ADDED
|
Git LFS Details
|
assets/zonos3_mini_hero.webp
ADDED
|
Git LFS Details
|
autoencoder/generator.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:75a587a3eda9ca3d345f0c305756a7ff8619b01685c261b08ae243c69c2caf9a
|
| 3 |
+
size 7960647220
|
autoencoder/heads/full_mr/config_full_unfreezeall_mr.json
ADDED
|
@@ -0,0 +1,455 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": {
|
| 3 |
+
"architecture": "s1_semantic_rvq_moss_v2_hybrid",
|
| 4 |
+
"sample_rate": 44100,
|
| 5 |
+
"internal_sample_rate": 44160,
|
| 6 |
+
"channels": 1,
|
| 7 |
+
"channel_interleave": false,
|
| 8 |
+
"frame_rate": 15,
|
| 9 |
+
"internal_hop_length": 2944,
|
| 10 |
+
"external_samples_per_frame": 2940,
|
| 11 |
+
"encoder_stages": [
|
| 12 |
+
{
|
| 13 |
+
"type": "S1CausalWNConv1d",
|
| 14 |
+
"in_channels": 1,
|
| 15 |
+
"out_channels": 64,
|
| 16 |
+
"kernel_size": 7,
|
| 17 |
+
"stride": 1
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"type": "S1EncoderBlock",
|
| 21 |
+
"in_channels": 64,
|
| 22 |
+
"out_channels": 128,
|
| 23 |
+
"stride": 2,
|
| 24 |
+
"kernel_size": 4,
|
| 25 |
+
"residual_dilations": [
|
| 26 |
+
1,
|
| 27 |
+
3,
|
| 28 |
+
9
|
| 29 |
+
],
|
| 30 |
+
"residual_kernel_size": 7,
|
| 31 |
+
"activation": "Snake1d",
|
| 32 |
+
"transformer_layers": 0
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"type": "S1EncoderBlock",
|
| 36 |
+
"in_channels": 128,
|
| 37 |
+
"out_channels": 256,
|
| 38 |
+
"stride": 4,
|
| 39 |
+
"kernel_size": 8,
|
| 40 |
+
"residual_dilations": [
|
| 41 |
+
1,
|
| 42 |
+
3,
|
| 43 |
+
9
|
| 44 |
+
],
|
| 45 |
+
"residual_kernel_size": 7,
|
| 46 |
+
"activation": "Snake1d",
|
| 47 |
+
"transformer_layers": 0
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"type": "PatchedPretransform",
|
| 51 |
+
"patch_size": 23,
|
| 52 |
+
"direction": "down"
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"type": "MossProjectedTransformer",
|
| 56 |
+
"stage": 2,
|
| 57 |
+
"input_dimension": 5888,
|
| 58 |
+
"output_dimension": 384,
|
| 59 |
+
"d_model": 768,
|
| 60 |
+
"num_heads": 12,
|
| 61 |
+
"num_layers": 12,
|
| 62 |
+
"dim_feedforward": 3072,
|
| 63 |
+
"causal": true,
|
| 64 |
+
"norm": "layer_norm",
|
| 65 |
+
"positional_embedding": "rope",
|
| 66 |
+
"max_period": 10000,
|
| 67 |
+
"gating": "none",
|
| 68 |
+
"activation": "gelu",
|
| 69 |
+
"layer_scale": 0.01,
|
| 70 |
+
"attention_dropout": 0.1,
|
| 71 |
+
"ffn_dropout": 0.1,
|
| 72 |
+
"frame_rate": 240,
|
| 73 |
+
"context_duration": 2,
|
| 74 |
+
"context_frames": 480
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"type": "PatchedPretransform",
|
| 78 |
+
"patch_size": 2,
|
| 79 |
+
"direction": "down"
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"type": "MossProjectedTransformer",
|
| 83 |
+
"stage": 3,
|
| 84 |
+
"input_dimension": 768,
|
| 85 |
+
"output_dimension": 384,
|
| 86 |
+
"d_model": 768,
|
| 87 |
+
"num_heads": 12,
|
| 88 |
+
"num_layers": 12,
|
| 89 |
+
"dim_feedforward": 3072,
|
| 90 |
+
"causal": true,
|
| 91 |
+
"norm": "layer_norm",
|
| 92 |
+
"positional_embedding": "rope",
|
| 93 |
+
"max_period": 10000,
|
| 94 |
+
"gating": "none",
|
| 95 |
+
"activation": "gelu",
|
| 96 |
+
"layer_scale": 0.01,
|
| 97 |
+
"attention_dropout": 0.1,
|
| 98 |
+
"ffn_dropout": 0.1,
|
| 99 |
+
"frame_rate": 120,
|
| 100 |
+
"context_duration": 4,
|
| 101 |
+
"context_frames": 480
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"type": "PatchedPretransform",
|
| 105 |
+
"patch_size": 2,
|
| 106 |
+
"direction": "down"
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"type": "MossProjectedTransformer",
|
| 110 |
+
"stage": 4,
|
| 111 |
+
"input_dimension": 768,
|
| 112 |
+
"output_dimension": 384,
|
| 113 |
+
"d_model": 768,
|
| 114 |
+
"num_heads": 12,
|
| 115 |
+
"num_layers": 12,
|
| 116 |
+
"dim_feedforward": 3072,
|
| 117 |
+
"causal": true,
|
| 118 |
+
"norm": "layer_norm",
|
| 119 |
+
"positional_embedding": "rope",
|
| 120 |
+
"max_period": 10000,
|
| 121 |
+
"gating": "none",
|
| 122 |
+
"activation": "gelu",
|
| 123 |
+
"layer_scale": 0.01,
|
| 124 |
+
"attention_dropout": 0.1,
|
| 125 |
+
"ffn_dropout": 0.1,
|
| 126 |
+
"frame_rate": 60,
|
| 127 |
+
"context_duration": 8,
|
| 128 |
+
"context_frames": 480
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"type": "PatchedPretransform",
|
| 132 |
+
"patch_size": 2,
|
| 133 |
+
"direction": "down"
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"type": "MossProjectedTransformer",
|
| 137 |
+
"stage": 5,
|
| 138 |
+
"input_dimension": 768,
|
| 139 |
+
"output_dimension": 640,
|
| 140 |
+
"d_model": 768,
|
| 141 |
+
"num_heads": 12,
|
| 142 |
+
"num_layers": 12,
|
| 143 |
+
"dim_feedforward": 3072,
|
| 144 |
+
"causal": true,
|
| 145 |
+
"norm": "layer_norm",
|
| 146 |
+
"positional_embedding": "rope",
|
| 147 |
+
"max_period": 10000,
|
| 148 |
+
"gating": "none",
|
| 149 |
+
"activation": "gelu",
|
| 150 |
+
"layer_scale": 0.01,
|
| 151 |
+
"attention_dropout": 0.1,
|
| 152 |
+
"ffn_dropout": 0.1,
|
| 153 |
+
"frame_rate": 30,
|
| 154 |
+
"context_duration": 10,
|
| 155 |
+
"context_frames": 300
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"type": "PatchedPretransform",
|
| 159 |
+
"patch_size": 2,
|
| 160 |
+
"direction": "down"
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"type": "MossProjectedTransformer",
|
| 164 |
+
"stage": 6,
|
| 165 |
+
"input_dimension": 1280,
|
| 166 |
+
"output_dimension": 1024,
|
| 167 |
+
"d_model": 1280,
|
| 168 |
+
"num_heads": 20,
|
| 169 |
+
"num_layers": 32,
|
| 170 |
+
"dim_feedforward": 5120,
|
| 171 |
+
"causal": true,
|
| 172 |
+
"norm": "layer_norm",
|
| 173 |
+
"positional_embedding": "rope",
|
| 174 |
+
"max_period": 10000,
|
| 175 |
+
"gating": "none",
|
| 176 |
+
"activation": "gelu",
|
| 177 |
+
"layer_scale": 0.01,
|
| 178 |
+
"attention_dropout": 0.1,
|
| 179 |
+
"ffn_dropout": 0.1,
|
| 180 |
+
"frame_rate": 15,
|
| 181 |
+
"context_duration": 10,
|
| 182 |
+
"context_frames": 150
|
| 183 |
+
}
|
| 184 |
+
],
|
| 185 |
+
"decoder_stages": [
|
| 186 |
+
{
|
| 187 |
+
"type": "MossProjectedTransformer",
|
| 188 |
+
"stage": 6,
|
| 189 |
+
"input_dimension": 1024,
|
| 190 |
+
"output_dimension": 1280,
|
| 191 |
+
"d_model": 1280,
|
| 192 |
+
"num_heads": 20,
|
| 193 |
+
"num_layers": 32,
|
| 194 |
+
"dim_feedforward": 5120,
|
| 195 |
+
"causal": true,
|
| 196 |
+
"norm": "layer_norm",
|
| 197 |
+
"positional_embedding": "rope",
|
| 198 |
+
"max_period": 10000,
|
| 199 |
+
"gating": "none",
|
| 200 |
+
"activation": "gelu",
|
| 201 |
+
"layer_scale": 0.01,
|
| 202 |
+
"attention_dropout": 0.1,
|
| 203 |
+
"ffn_dropout": 0.1,
|
| 204 |
+
"frame_rate": 15,
|
| 205 |
+
"context_duration": 10,
|
| 206 |
+
"context_frames": 150
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"type": "PatchedPretransform",
|
| 210 |
+
"patch_size": 2,
|
| 211 |
+
"direction": "up"
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"type": "MossProjectedTransformer",
|
| 215 |
+
"stage": 5,
|
| 216 |
+
"input_dimension": 640,
|
| 217 |
+
"output_dimension": 768,
|
| 218 |
+
"d_model": 768,
|
| 219 |
+
"num_heads": 12,
|
| 220 |
+
"num_layers": 12,
|
| 221 |
+
"dim_feedforward": 3072,
|
| 222 |
+
"causal": true,
|
| 223 |
+
"norm": "layer_norm",
|
| 224 |
+
"positional_embedding": "rope",
|
| 225 |
+
"max_period": 10000,
|
| 226 |
+
"gating": "none",
|
| 227 |
+
"activation": "gelu",
|
| 228 |
+
"layer_scale": 0.01,
|
| 229 |
+
"attention_dropout": 0.1,
|
| 230 |
+
"ffn_dropout": 0.1,
|
| 231 |
+
"frame_rate": 30,
|
| 232 |
+
"context_duration": 10,
|
| 233 |
+
"context_frames": 300
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"type": "PatchedPretransform",
|
| 237 |
+
"patch_size": 2,
|
| 238 |
+
"direction": "up"
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"type": "MossProjectedTransformer",
|
| 242 |
+
"stage": 4,
|
| 243 |
+
"input_dimension": 384,
|
| 244 |
+
"output_dimension": 768,
|
| 245 |
+
"d_model": 768,
|
| 246 |
+
"num_heads": 12,
|
| 247 |
+
"num_layers": 12,
|
| 248 |
+
"dim_feedforward": 3072,
|
| 249 |
+
"causal": true,
|
| 250 |
+
"norm": "layer_norm",
|
| 251 |
+
"positional_embedding": "rope",
|
| 252 |
+
"max_period": 10000,
|
| 253 |
+
"gating": "none",
|
| 254 |
+
"activation": "gelu",
|
| 255 |
+
"layer_scale": 0.01,
|
| 256 |
+
"attention_dropout": 0.1,
|
| 257 |
+
"ffn_dropout": 0.1,
|
| 258 |
+
"frame_rate": 60,
|
| 259 |
+
"context_duration": 8,
|
| 260 |
+
"context_frames": 480
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"type": "PatchedPretransform",
|
| 264 |
+
"patch_size": 2,
|
| 265 |
+
"direction": "up"
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"type": "MossProjectedTransformer",
|
| 269 |
+
"stage": 3,
|
| 270 |
+
"input_dimension": 384,
|
| 271 |
+
"output_dimension": 768,
|
| 272 |
+
"d_model": 768,
|
| 273 |
+
"num_heads": 12,
|
| 274 |
+
"num_layers": 12,
|
| 275 |
+
"dim_feedforward": 3072,
|
| 276 |
+
"causal": true,
|
| 277 |
+
"norm": "layer_norm",
|
| 278 |
+
"positional_embedding": "rope",
|
| 279 |
+
"max_period": 10000,
|
| 280 |
+
"gating": "none",
|
| 281 |
+
"activation": "gelu",
|
| 282 |
+
"layer_scale": 0.01,
|
| 283 |
+
"attention_dropout": 0.1,
|
| 284 |
+
"ffn_dropout": 0.1,
|
| 285 |
+
"frame_rate": 120,
|
| 286 |
+
"context_duration": 4,
|
| 287 |
+
"context_frames": 480
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"type": "PatchedPretransform",
|
| 291 |
+
"patch_size": 2,
|
| 292 |
+
"direction": "up"
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"type": "MossProjectedTransformer",
|
| 296 |
+
"stage": 2,
|
| 297 |
+
"input_dimension": 384,
|
| 298 |
+
"output_dimension": 5888,
|
| 299 |
+
"d_model": 768,
|
| 300 |
+
"num_heads": 12,
|
| 301 |
+
"num_layers": 12,
|
| 302 |
+
"dim_feedforward": 3072,
|
| 303 |
+
"causal": true,
|
| 304 |
+
"norm": "layer_norm",
|
| 305 |
+
"positional_embedding": "rope",
|
| 306 |
+
"max_period": 10000,
|
| 307 |
+
"gating": "none",
|
| 308 |
+
"activation": "gelu",
|
| 309 |
+
"layer_scale": 0.01,
|
| 310 |
+
"attention_dropout": 0.1,
|
| 311 |
+
"ffn_dropout": 0.1,
|
| 312 |
+
"frame_rate": 240,
|
| 313 |
+
"context_duration": 2,
|
| 314 |
+
"context_frames": 480
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"type": "PatchedPretransform",
|
| 318 |
+
"patch_size": 23,
|
| 319 |
+
"direction": "up"
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"type": "S1DecoderBlock",
|
| 323 |
+
"in_channels": 256,
|
| 324 |
+
"out_channels": 128,
|
| 325 |
+
"stride": 4,
|
| 326 |
+
"kernel_size": 8,
|
| 327 |
+
"residual_dilations": [
|
| 328 |
+
1,
|
| 329 |
+
3,
|
| 330 |
+
9
|
| 331 |
+
],
|
| 332 |
+
"residual_kernel_size": 7,
|
| 333 |
+
"activation": "Snake1d",
|
| 334 |
+
"transformer_layers": 0
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"type": "S1DecoderBlock",
|
| 338 |
+
"in_channels": 128,
|
| 339 |
+
"out_channels": 64,
|
| 340 |
+
"stride": 2,
|
| 341 |
+
"kernel_size": 4,
|
| 342 |
+
"residual_dilations": [
|
| 343 |
+
1,
|
| 344 |
+
3,
|
| 345 |
+
9
|
| 346 |
+
],
|
| 347 |
+
"residual_kernel_size": 7,
|
| 348 |
+
"activation": "Snake1d",
|
| 349 |
+
"transformer_layers": 0
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"type": "S1WaveformHead",
|
| 353 |
+
"in_channels": 64,
|
| 354 |
+
"out_channels": 1,
|
| 355 |
+
"kernel_size": 7,
|
| 356 |
+
"activation_before": "Snake1d",
|
| 357 |
+
"activation_after": "tanh"
|
| 358 |
+
}
|
| 359 |
+
],
|
| 360 |
+
"bottleneck": {
|
| 361 |
+
"type": "dual_encoding_semantic_residual_rvq",
|
| 362 |
+
"input_dim": 1024,
|
| 363 |
+
"semantic_codebooks": 1,
|
| 364 |
+
"residual_codebooks": 9,
|
| 365 |
+
"semantic_codebook_size": 4096,
|
| 366 |
+
"residual_codebook_size": 4096,
|
| 367 |
+
"codebook_dim": 8,
|
| 368 |
+
"quantizer_dropout": 0.5,
|
| 369 |
+
"semantic_quantizer_dropout": 0.0,
|
| 370 |
+
"downsample_factors": [],
|
| 371 |
+
"upsample_factors": [],
|
| 372 |
+
"pre_module": "identity",
|
| 373 |
+
"post_module": "identity",
|
| 374 |
+
"lookup_and_vq_loss_dtype": "float32",
|
| 375 |
+
"semantic_head": {
|
| 376 |
+
"type": "ConvNeXtDecoder",
|
| 377 |
+
"width": 768,
|
| 378 |
+
"hidden": 2048,
|
| 379 |
+
"depth": 4,
|
| 380 |
+
"out_channels": 1024,
|
| 381 |
+
"causal": true
|
| 382 |
+
},
|
| 383 |
+
"distillation_input": "decoded_ssl_semantic_codebook"
|
| 384 |
+
},
|
| 385 |
+
"boundary_resampler": {
|
| 386 |
+
"input_rate": 44100,
|
| 387 |
+
"codec_rate": 44160,
|
| 388 |
+
"output_rate": 44100,
|
| 389 |
+
"encoder_ratio": [
|
| 390 |
+
736,
|
| 391 |
+
735
|
| 392 |
+
],
|
| 393 |
+
"decoder_ratio": [
|
| 394 |
+
735,
|
| 395 |
+
736
|
| 396 |
+
],
|
| 397 |
+
"method": "sinc_interp_hann",
|
| 398 |
+
"lowpass_filter_width": 16,
|
| 399 |
+
"rolloff": 0.99,
|
| 400 |
+
"streaming_policy": "finite_fir_lookahead_with_delay_compensated_timestamps",
|
| 401 |
+
"training_policy": "time_aligned_full_crop_reference",
|
| 402 |
+
"losses_and_teacher_use_external_rate": true
|
| 403 |
+
},
|
| 404 |
+
"checkpoint_transformers": true,
|
| 405 |
+
"attention_implementation": "pytorch_sdpa_chunked_causal_local_with_dropout",
|
| 406 |
+
"residual_codebooks": 9,
|
| 407 |
+
"semantic_codebook_size": 4096,
|
| 408 |
+
"residual_codebook_size": 4096,
|
| 409 |
+
"codebook_dim": 8,
|
| 410 |
+
"quantizer_dropout": 0.5,
|
| 411 |
+
"decoder_mirrors_encoder": true,
|
| 412 |
+
"dual_encoding": {
|
| 413 |
+
"width": 768,
|
| 414 |
+
"hidden": 2048,
|
| 415 |
+
"depth": 4,
|
| 416 |
+
"causal": true,
|
| 417 |
+
"window_ratio": 2.0,
|
| 418 |
+
"strategy": "frozen_ssl_to_q1_decoded_semantic_residual",
|
| 419 |
+
"normalization": "unchanged_baseline"
|
| 420 |
+
}
|
| 421 |
+
},
|
| 422 |
+
"head": {
|
| 423 |
+
"depth": 8,
|
| 424 |
+
"hidden": 2048,
|
| 425 |
+
"hop": 184,
|
| 426 |
+
"frame_rate": 240,
|
| 427 |
+
"trainable_stages": [
|
| 428 |
+
0,
|
| 429 |
+
1,
|
| 430 |
+
2,
|
| 431 |
+
3,
|
| 432 |
+
4,
|
| 433 |
+
5,
|
| 434 |
+
6,
|
| 435 |
+
7,
|
| 436 |
+
8
|
| 437 |
+
],
|
| 438 |
+
"parametrization": "per branch: log-magnitude + phase angle (Vocos), magnitude clipped at 1e2, periodic Hann, centered iSTFT; branches combined by a linear fp32 FIR mixer",
|
| 439 |
+
"recompute_trainable_stages": true,
|
| 440 |
+
"compile_trainable_stages": true,
|
| 441 |
+
"type": "mr_istft",
|
| 442 |
+
"n_ffts": [
|
| 443 |
+
736,
|
| 444 |
+
2208
|
| 445 |
+
],
|
| 446 |
+
"fir_kernel": 127,
|
| 447 |
+
"warm_blocks": 4
|
| 448 |
+
},
|
| 449 |
+
"init": {
|
| 450 |
+
"generator_checkpoint": "hf://Zyphra/ZONOS3-Mini/autoencoder/generator.pt",
|
| 451 |
+
"trainable_rename": {
|
| 452 |
+
"head.out.": "head.outs.0."
|
| 453 |
+
}
|
| 454 |
+
}
|
| 455 |
+
}
|
autoencoder/heads/full_mr/step_0050000.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9d2e2340c74f763340535f84e5dda59b5eabea668360400f8fac6b641598a1e9
|
| 3 |
+
size 4012134189
|
autoencoder/heads/keep2_plain/config_full_unfreezeall_mr_drop30_4x10.json
ADDED
|
@@ -0,0 +1,473 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model": {
|
| 3 |
+
"architecture": "s1_semantic_rvq_moss_v2_hybrid",
|
| 4 |
+
"sample_rate": 44100,
|
| 5 |
+
"internal_sample_rate": 44160,
|
| 6 |
+
"channels": 1,
|
| 7 |
+
"channel_interleave": false,
|
| 8 |
+
"frame_rate": 15,
|
| 9 |
+
"internal_hop_length": 2944,
|
| 10 |
+
"external_samples_per_frame": 2940,
|
| 11 |
+
"encoder_stages": [
|
| 12 |
+
{
|
| 13 |
+
"type": "S1CausalWNConv1d",
|
| 14 |
+
"in_channels": 1,
|
| 15 |
+
"out_channels": 64,
|
| 16 |
+
"kernel_size": 7,
|
| 17 |
+
"stride": 1
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"type": "S1EncoderBlock",
|
| 21 |
+
"in_channels": 64,
|
| 22 |
+
"out_channels": 128,
|
| 23 |
+
"stride": 2,
|
| 24 |
+
"kernel_size": 4,
|
| 25 |
+
"residual_dilations": [
|
| 26 |
+
1,
|
| 27 |
+
3,
|
| 28 |
+
9
|
| 29 |
+
],
|
| 30 |
+
"residual_kernel_size": 7,
|
| 31 |
+
"activation": "Snake1d",
|
| 32 |
+
"transformer_layers": 0
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"type": "S1EncoderBlock",
|
| 36 |
+
"in_channels": 128,
|
| 37 |
+
"out_channels": 256,
|
| 38 |
+
"stride": 4,
|
| 39 |
+
"kernel_size": 8,
|
| 40 |
+
"residual_dilations": [
|
| 41 |
+
1,
|
| 42 |
+
3,
|
| 43 |
+
9
|
| 44 |
+
],
|
| 45 |
+
"residual_kernel_size": 7,
|
| 46 |
+
"activation": "Snake1d",
|
| 47 |
+
"transformer_layers": 0
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"type": "PatchedPretransform",
|
| 51 |
+
"patch_size": 23,
|
| 52 |
+
"direction": "down"
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"type": "MossProjectedTransformer",
|
| 56 |
+
"stage": 2,
|
| 57 |
+
"input_dimension": 5888,
|
| 58 |
+
"output_dimension": 384,
|
| 59 |
+
"d_model": 768,
|
| 60 |
+
"num_heads": 12,
|
| 61 |
+
"num_layers": 12,
|
| 62 |
+
"dim_feedforward": 3072,
|
| 63 |
+
"causal": true,
|
| 64 |
+
"norm": "layer_norm",
|
| 65 |
+
"positional_embedding": "rope",
|
| 66 |
+
"max_period": 10000,
|
| 67 |
+
"gating": "none",
|
| 68 |
+
"activation": "gelu",
|
| 69 |
+
"layer_scale": 0.01,
|
| 70 |
+
"attention_dropout": 0.1,
|
| 71 |
+
"ffn_dropout": 0.1,
|
| 72 |
+
"frame_rate": 240,
|
| 73 |
+
"context_duration": 2,
|
| 74 |
+
"context_frames": 480
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"type": "PatchedPretransform",
|
| 78 |
+
"patch_size": 2,
|
| 79 |
+
"direction": "down"
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"type": "MossProjectedTransformer",
|
| 83 |
+
"stage": 3,
|
| 84 |
+
"input_dimension": 768,
|
| 85 |
+
"output_dimension": 384,
|
| 86 |
+
"d_model": 768,
|
| 87 |
+
"num_heads": 12,
|
| 88 |
+
"num_layers": 12,
|
| 89 |
+
"dim_feedforward": 3072,
|
| 90 |
+
"causal": true,
|
| 91 |
+
"norm": "layer_norm",
|
| 92 |
+
"positional_embedding": "rope",
|
| 93 |
+
"max_period": 10000,
|
| 94 |
+
"gating": "none",
|
| 95 |
+
"activation": "gelu",
|
| 96 |
+
"layer_scale": 0.01,
|
| 97 |
+
"attention_dropout": 0.1,
|
| 98 |
+
"ffn_dropout": 0.1,
|
| 99 |
+
"frame_rate": 120,
|
| 100 |
+
"context_duration": 4,
|
| 101 |
+
"context_frames": 480
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"type": "PatchedPretransform",
|
| 105 |
+
"patch_size": 2,
|
| 106 |
+
"direction": "down"
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"type": "MossProjectedTransformer",
|
| 110 |
+
"stage": 4,
|
| 111 |
+
"input_dimension": 768,
|
| 112 |
+
"output_dimension": 384,
|
| 113 |
+
"d_model": 768,
|
| 114 |
+
"num_heads": 12,
|
| 115 |
+
"num_layers": 12,
|
| 116 |
+
"dim_feedforward": 3072,
|
| 117 |
+
"causal": true,
|
| 118 |
+
"norm": "layer_norm",
|
| 119 |
+
"positional_embedding": "rope",
|
| 120 |
+
"max_period": 10000,
|
| 121 |
+
"gating": "none",
|
| 122 |
+
"activation": "gelu",
|
| 123 |
+
"layer_scale": 0.01,
|
| 124 |
+
"attention_dropout": 0.1,
|
| 125 |
+
"ffn_dropout": 0.1,
|
| 126 |
+
"frame_rate": 60,
|
| 127 |
+
"context_duration": 8,
|
| 128 |
+
"context_frames": 480
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"type": "PatchedPretransform",
|
| 132 |
+
"patch_size": 2,
|
| 133 |
+
"direction": "down"
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"type": "MossProjectedTransformer",
|
| 137 |
+
"stage": 5,
|
| 138 |
+
"input_dimension": 768,
|
| 139 |
+
"output_dimension": 640,
|
| 140 |
+
"d_model": 768,
|
| 141 |
+
"num_heads": 12,
|
| 142 |
+
"num_layers": 12,
|
| 143 |
+
"dim_feedforward": 3072,
|
| 144 |
+
"causal": true,
|
| 145 |
+
"norm": "layer_norm",
|
| 146 |
+
"positional_embedding": "rope",
|
| 147 |
+
"max_period": 10000,
|
| 148 |
+
"gating": "none",
|
| 149 |
+
"activation": "gelu",
|
| 150 |
+
"layer_scale": 0.01,
|
| 151 |
+
"attention_dropout": 0.1,
|
| 152 |
+
"ffn_dropout": 0.1,
|
| 153 |
+
"frame_rate": 30,
|
| 154 |
+
"context_duration": 10,
|
| 155 |
+
"context_frames": 300
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"type": "PatchedPretransform",
|
| 159 |
+
"patch_size": 2,
|
| 160 |
+
"direction": "down"
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"type": "MossProjectedTransformer",
|
| 164 |
+
"stage": 6,
|
| 165 |
+
"input_dimension": 1280,
|
| 166 |
+
"output_dimension": 1024,
|
| 167 |
+
"d_model": 1280,
|
| 168 |
+
"num_heads": 20,
|
| 169 |
+
"num_layers": 32,
|
| 170 |
+
"dim_feedforward": 5120,
|
| 171 |
+
"causal": true,
|
| 172 |
+
"norm": "layer_norm",
|
| 173 |
+
"positional_embedding": "rope",
|
| 174 |
+
"max_period": 10000,
|
| 175 |
+
"gating": "none",
|
| 176 |
+
"activation": "gelu",
|
| 177 |
+
"layer_scale": 0.01,
|
| 178 |
+
"attention_dropout": 0.1,
|
| 179 |
+
"ffn_dropout": 0.1,
|
| 180 |
+
"frame_rate": 15,
|
| 181 |
+
"context_duration": 10,
|
| 182 |
+
"context_frames": 150
|
| 183 |
+
}
|
| 184 |
+
],
|
| 185 |
+
"decoder_stages": [
|
| 186 |
+
{
|
| 187 |
+
"type": "MossProjectedTransformer",
|
| 188 |
+
"stage": 6,
|
| 189 |
+
"input_dimension": 1024,
|
| 190 |
+
"output_dimension": 1280,
|
| 191 |
+
"d_model": 1280,
|
| 192 |
+
"num_heads": 20,
|
| 193 |
+
"num_layers": 32,
|
| 194 |
+
"dim_feedforward": 5120,
|
| 195 |
+
"causal": true,
|
| 196 |
+
"norm": "layer_norm",
|
| 197 |
+
"positional_embedding": "rope",
|
| 198 |
+
"max_period": 10000,
|
| 199 |
+
"gating": "none",
|
| 200 |
+
"activation": "gelu",
|
| 201 |
+
"layer_scale": 0.01,
|
| 202 |
+
"attention_dropout": 0.1,
|
| 203 |
+
"ffn_dropout": 0.1,
|
| 204 |
+
"frame_rate": 15,
|
| 205 |
+
"context_duration": 10,
|
| 206 |
+
"context_frames": 150
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"type": "PatchedPretransform",
|
| 210 |
+
"patch_size": 2,
|
| 211 |
+
"direction": "up"
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"type": "MossProjectedTransformer",
|
| 215 |
+
"stage": 5,
|
| 216 |
+
"input_dimension": 640,
|
| 217 |
+
"output_dimension": 768,
|
| 218 |
+
"d_model": 768,
|
| 219 |
+
"num_heads": 12,
|
| 220 |
+
"num_layers": 12,
|
| 221 |
+
"dim_feedforward": 3072,
|
| 222 |
+
"causal": true,
|
| 223 |
+
"norm": "layer_norm",
|
| 224 |
+
"positional_embedding": "rope",
|
| 225 |
+
"max_period": 10000,
|
| 226 |
+
"gating": "none",
|
| 227 |
+
"activation": "gelu",
|
| 228 |
+
"layer_scale": 0.01,
|
| 229 |
+
"attention_dropout": 0.1,
|
| 230 |
+
"ffn_dropout": 0.1,
|
| 231 |
+
"frame_rate": 30,
|
| 232 |
+
"context_duration": 10,
|
| 233 |
+
"context_frames": 300
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"type": "PatchedPretransform",
|
| 237 |
+
"patch_size": 2,
|
| 238 |
+
"direction": "up"
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"type": "MossProjectedTransformer",
|
| 242 |
+
"stage": 4,
|
| 243 |
+
"input_dimension": 384,
|
| 244 |
+
"output_dimension": 768,
|
| 245 |
+
"d_model": 768,
|
| 246 |
+
"num_heads": 12,
|
| 247 |
+
"num_layers": 12,
|
| 248 |
+
"dim_feedforward": 3072,
|
| 249 |
+
"causal": true,
|
| 250 |
+
"norm": "layer_norm",
|
| 251 |
+
"positional_embedding": "rope",
|
| 252 |
+
"max_period": 10000,
|
| 253 |
+
"gating": "none",
|
| 254 |
+
"activation": "gelu",
|
| 255 |
+
"layer_scale": 0.01,
|
| 256 |
+
"attention_dropout": 0.1,
|
| 257 |
+
"ffn_dropout": 0.1,
|
| 258 |
+
"frame_rate": 60,
|
| 259 |
+
"context_duration": 8,
|
| 260 |
+
"context_frames": 480
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"type": "PatchedPretransform",
|
| 264 |
+
"patch_size": 2,
|
| 265 |
+
"direction": "up"
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"type": "MossProjectedTransformer",
|
| 269 |
+
"stage": 3,
|
| 270 |
+
"input_dimension": 384,
|
| 271 |
+
"output_dimension": 768,
|
| 272 |
+
"d_model": 768,
|
| 273 |
+
"num_heads": 12,
|
| 274 |
+
"num_layers": 12,
|
| 275 |
+
"dim_feedforward": 3072,
|
| 276 |
+
"causal": true,
|
| 277 |
+
"norm": "layer_norm",
|
| 278 |
+
"positional_embedding": "rope",
|
| 279 |
+
"max_period": 10000,
|
| 280 |
+
"gating": "none",
|
| 281 |
+
"activation": "gelu",
|
| 282 |
+
"layer_scale": 0.01,
|
| 283 |
+
"attention_dropout": 0.1,
|
| 284 |
+
"ffn_dropout": 0.1,
|
| 285 |
+
"frame_rate": 120,
|
| 286 |
+
"context_duration": 4,
|
| 287 |
+
"context_frames": 480
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"type": "PatchedPretransform",
|
| 291 |
+
"patch_size": 2,
|
| 292 |
+
"direction": "up"
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"type": "MossProjectedTransformer",
|
| 296 |
+
"stage": 2,
|
| 297 |
+
"input_dimension": 384,
|
| 298 |
+
"output_dimension": 5888,
|
| 299 |
+
"d_model": 768,
|
| 300 |
+
"num_heads": 12,
|
| 301 |
+
"num_layers": 12,
|
| 302 |
+
"dim_feedforward": 3072,
|
| 303 |
+
"causal": true,
|
| 304 |
+
"norm": "layer_norm",
|
| 305 |
+
"positional_embedding": "rope",
|
| 306 |
+
"max_period": 10000,
|
| 307 |
+
"gating": "none",
|
| 308 |
+
"activation": "gelu",
|
| 309 |
+
"layer_scale": 0.01,
|
| 310 |
+
"attention_dropout": 0.1,
|
| 311 |
+
"ffn_dropout": 0.1,
|
| 312 |
+
"frame_rate": 240,
|
| 313 |
+
"context_duration": 2,
|
| 314 |
+
"context_frames": 480
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"type": "PatchedPretransform",
|
| 318 |
+
"patch_size": 23,
|
| 319 |
+
"direction": "up"
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"type": "S1DecoderBlock",
|
| 323 |
+
"in_channels": 256,
|
| 324 |
+
"out_channels": 128,
|
| 325 |
+
"stride": 4,
|
| 326 |
+
"kernel_size": 8,
|
| 327 |
+
"residual_dilations": [
|
| 328 |
+
1,
|
| 329 |
+
3,
|
| 330 |
+
9
|
| 331 |
+
],
|
| 332 |
+
"residual_kernel_size": 7,
|
| 333 |
+
"activation": "Snake1d",
|
| 334 |
+
"transformer_layers": 0
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"type": "S1DecoderBlock",
|
| 338 |
+
"in_channels": 128,
|
| 339 |
+
"out_channels": 64,
|
| 340 |
+
"stride": 2,
|
| 341 |
+
"kernel_size": 4,
|
| 342 |
+
"residual_dilations": [
|
| 343 |
+
1,
|
| 344 |
+
3,
|
| 345 |
+
9
|
| 346 |
+
],
|
| 347 |
+
"residual_kernel_size": 7,
|
| 348 |
+
"activation": "Snake1d",
|
| 349 |
+
"transformer_layers": 0
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"type": "S1WaveformHead",
|
| 353 |
+
"in_channels": 64,
|
| 354 |
+
"out_channels": 1,
|
| 355 |
+
"kernel_size": 7,
|
| 356 |
+
"activation_before": "Snake1d",
|
| 357 |
+
"activation_after": "tanh"
|
| 358 |
+
}
|
| 359 |
+
],
|
| 360 |
+
"bottleneck": {
|
| 361 |
+
"type": "dual_encoding_semantic_residual_rvq",
|
| 362 |
+
"input_dim": 1024,
|
| 363 |
+
"semantic_codebooks": 1,
|
| 364 |
+
"residual_codebooks": 9,
|
| 365 |
+
"semantic_codebook_size": 4096,
|
| 366 |
+
"residual_codebook_size": 4096,
|
| 367 |
+
"codebook_dim": 8,
|
| 368 |
+
"quantizer_dropout": 0.5,
|
| 369 |
+
"semantic_quantizer_dropout": 0.0,
|
| 370 |
+
"downsample_factors": [],
|
| 371 |
+
"upsample_factors": [],
|
| 372 |
+
"pre_module": "identity",
|
| 373 |
+
"post_module": "identity",
|
| 374 |
+
"lookup_and_vq_loss_dtype": "float32",
|
| 375 |
+
"semantic_head": {
|
| 376 |
+
"type": "ConvNeXtDecoder",
|
| 377 |
+
"width": 768,
|
| 378 |
+
"hidden": 2048,
|
| 379 |
+
"depth": 4,
|
| 380 |
+
"out_channels": 1024,
|
| 381 |
+
"causal": true
|
| 382 |
+
},
|
| 383 |
+
"distillation_input": "decoded_ssl_semantic_codebook"
|
| 384 |
+
},
|
| 385 |
+
"boundary_resampler": {
|
| 386 |
+
"input_rate": 44100,
|
| 387 |
+
"codec_rate": 44160,
|
| 388 |
+
"output_rate": 44100,
|
| 389 |
+
"encoder_ratio": [
|
| 390 |
+
736,
|
| 391 |
+
735
|
| 392 |
+
],
|
| 393 |
+
"decoder_ratio": [
|
| 394 |
+
735,
|
| 395 |
+
736
|
| 396 |
+
],
|
| 397 |
+
"method": "sinc_interp_hann",
|
| 398 |
+
"lowpass_filter_width": 16,
|
| 399 |
+
"rolloff": 0.99,
|
| 400 |
+
"streaming_policy": "finite_fir_lookahead_with_delay_compensated_timestamps",
|
| 401 |
+
"training_policy": "time_aligned_full_crop_reference",
|
| 402 |
+
"losses_and_teacher_use_external_rate": true
|
| 403 |
+
},
|
| 404 |
+
"checkpoint_transformers": true,
|
| 405 |
+
"attention_implementation": "pytorch_sdpa_chunked_causal_local_with_dropout",
|
| 406 |
+
"residual_codebooks": 9,
|
| 407 |
+
"semantic_codebook_size": 4096,
|
| 408 |
+
"residual_codebook_size": 4096,
|
| 409 |
+
"codebook_dim": 8,
|
| 410 |
+
"quantizer_dropout": 0.5,
|
| 411 |
+
"decoder_mirrors_encoder": true,
|
| 412 |
+
"dual_encoding": {
|
| 413 |
+
"width": 768,
|
| 414 |
+
"hidden": 2048,
|
| 415 |
+
"depth": 4,
|
| 416 |
+
"causal": true,
|
| 417 |
+
"window_ratio": 2.0,
|
| 418 |
+
"strategy": "frozen_ssl_to_q1_decoded_semantic_residual",
|
| 419 |
+
"normalization": "unchanged_baseline"
|
| 420 |
+
}
|
| 421 |
+
},
|
| 422 |
+
"head": {
|
| 423 |
+
"depth": 8,
|
| 424 |
+
"hidden": 2048,
|
| 425 |
+
"hop": 184,
|
| 426 |
+
"frame_rate": 240,
|
| 427 |
+
"trainable_stages": [
|
| 428 |
+
0,
|
| 429 |
+
1,
|
| 430 |
+
2,
|
| 431 |
+
3,
|
| 432 |
+
4,
|
| 433 |
+
5,
|
| 434 |
+
6,
|
| 435 |
+
7,
|
| 436 |
+
8
|
| 437 |
+
],
|
| 438 |
+
"parametrization": "per branch: log-magnitude + phase angle (Vocos), magnitude clipped at 1e2, periodic Hann, centered iSTFT; branches combined by a linear fp32 FIR mixer",
|
| 439 |
+
"recompute_trainable_stages": true,
|
| 440 |
+
"compile_trainable_stages": true,
|
| 441 |
+
"type": "mr_istft",
|
| 442 |
+
"n_ffts": [
|
| 443 |
+
736,
|
| 444 |
+
2208
|
| 445 |
+
],
|
| 446 |
+
"fir_kernel": 127,
|
| 447 |
+
"stage_layers": {
|
| 448 |
+
"0": [
|
| 449 |
+
0,
|
| 450 |
+
31
|
| 451 |
+
],
|
| 452 |
+
"2": [
|
| 453 |
+
0,
|
| 454 |
+
11
|
| 455 |
+
],
|
| 456 |
+
"4": [
|
| 457 |
+
0,
|
| 458 |
+
11
|
| 459 |
+
],
|
| 460 |
+
"6": [
|
| 461 |
+
0,
|
| 462 |
+
11
|
| 463 |
+
],
|
| 464 |
+
"8": [
|
| 465 |
+
0,
|
| 466 |
+
11
|
| 467 |
+
]
|
| 468 |
+
}
|
| 469 |
+
},
|
| 470 |
+
"init": {
|
| 471 |
+
"generator_checkpoint": "hf://Zyphra/ZONOS3-Mini/autoencoder/generator.pt"
|
| 472 |
+
}
|
| 473 |
+
}
|
autoencoder/heads/keep2_plain/step_0050000.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7432e335e717417961267f2242547ce389a796cdc296a8b8d5289658e3c28936
|
| 3 |
+
size 518498361
|
autoencoder/teacher/config.json
ADDED
|
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.0,
|
| 3 |
+
"adapter_act": "relu",
|
| 4 |
+
"adapter_kernel_size": 3,
|
| 5 |
+
"adapter_stride": 2,
|
| 6 |
+
"add_adapter": false,
|
| 7 |
+
"apply_spec_augment": false,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"Wav2Vec2BertModel"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.0,
|
| 12 |
+
"bos_token_id": 1,
|
| 13 |
+
"classifier_proj_size": 768,
|
| 14 |
+
"codevector_dim": 768,
|
| 15 |
+
"conformer_conv_dropout": 0.1,
|
| 16 |
+
"contrastive_logits_temperature": 0.1,
|
| 17 |
+
"conv_depthwise_kernel_size": 31,
|
| 18 |
+
"ctc_loss_reduction": "sum",
|
| 19 |
+
"ctc_zero_infinity": false,
|
| 20 |
+
"diversity_loss_weight": 0.1,
|
| 21 |
+
"dtype": "bfloat16",
|
| 22 |
+
"eos_token_id": 2,
|
| 23 |
+
"feat_proj_dropout": 0.0,
|
| 24 |
+
"feat_quantizer_dropout": 0.0,
|
| 25 |
+
"feature_projection_input_dim": 160,
|
| 26 |
+
"final_dropout": 0.1,
|
| 27 |
+
"hidden_act": "swish",
|
| 28 |
+
"hidden_dropout": 0.0,
|
| 29 |
+
"hidden_size": 1024,
|
| 30 |
+
"initializer_range": 0.02,
|
| 31 |
+
"intermediate_size": 4096,
|
| 32 |
+
"layer_norm_eps": 1e-05,
|
| 33 |
+
"layerdrop": 0.1,
|
| 34 |
+
"left_max_position_embeddings": 64,
|
| 35 |
+
"mask_feature_length": 10,
|
| 36 |
+
"mask_feature_min_masks": 0,
|
| 37 |
+
"mask_feature_prob": 0.0,
|
| 38 |
+
"mask_time_length": 10,
|
| 39 |
+
"mask_time_min_masks": 2,
|
| 40 |
+
"mask_time_prob": 0.05,
|
| 41 |
+
"max_source_positions": 5000,
|
| 42 |
+
"model_type": "wav2vec2-bert",
|
| 43 |
+
"num_adapter_layers": 1,
|
| 44 |
+
"num_attention_heads": 16,
|
| 45 |
+
"num_codevector_groups": 2,
|
| 46 |
+
"num_codevectors_per_group": 320,
|
| 47 |
+
"num_hidden_layers": 16,
|
| 48 |
+
"num_negatives": 100,
|
| 49 |
+
"output_hidden_size": 1024,
|
| 50 |
+
"pad_token_id": 0,
|
| 51 |
+
"position_embeddings_type": "relative_key",
|
| 52 |
+
"proj_codevector_dim": 768,
|
| 53 |
+
"right_max_position_embeddings": 8,
|
| 54 |
+
"rotary_embedding_base": 10000,
|
| 55 |
+
"tdnn_dilation": [
|
| 56 |
+
1,
|
| 57 |
+
2,
|
| 58 |
+
3,
|
| 59 |
+
1,
|
| 60 |
+
1
|
| 61 |
+
],
|
| 62 |
+
"tdnn_dim": [
|
| 63 |
+
512,
|
| 64 |
+
512,
|
| 65 |
+
512,
|
| 66 |
+
512,
|
| 67 |
+
1500
|
| 68 |
+
],
|
| 69 |
+
"tdnn_kernel": [
|
| 70 |
+
5,
|
| 71 |
+
3,
|
| 72 |
+
3,
|
| 73 |
+
1,
|
| 74 |
+
1
|
| 75 |
+
],
|
| 76 |
+
"transformers_version": "5.3.0",
|
| 77 |
+
"use_intermediate_ffn_before_adapter": false,
|
| 78 |
+
"use_weighted_layer_sum": false,
|
| 79 |
+
"vocab_size": null,
|
| 80 |
+
"xvector_output_dim": 512
|
| 81 |
+
}
|
autoencoder/teacher/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a15182741d6ee3cd780b8b830f49fa6b7fcce383655d451970841b24114767a5
|
| 3 |
+
size 774162368
|
gguf/backbone-ptq-edit.gguf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7e23d261e2cef3cbbbaa33f43fbe683acb393a157ac672f2de04858aef21011f
|
| 3 |
+
size 663491840
|
gguf/backbone-ptq.gguf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:66fceb6d73a8c30bde80249e4112bb308380ab7f3384fa11c8e4048115e4ceec
|
| 3 |
+
size 663150624
|
gguf/edit/accents.gguf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ac77fc95e91a6b0c28135f27ee3e4fef1411b8ac3dba969e06f99ce9f84e7376
|
| 3 |
+
size 256592064
|
gguf/spk-encoder.gguf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9539f340b70be246005403b9c254732e180e85e0b3c57fb1973f0d6bc6c59bdc
|
| 3 |
+
size 45385728
|
gguf/ssl-teacher.gguf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:59e4f150e1ccdf54e6349b06605c2274f2262139678688025b6614f1393e415a
|
| 3 |
+
size 1548700000
|
gguf/steer.gguf
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6e8e3f86e89a09801ce083b9286f43263ec696acfaff72e25aca537aa9efa90c
|
| 3 |
+
size 204960
|
gguf/voices/English/Cillian_en-Irish_M_1.codes.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dadae1124c63cda7a0627fd282d54aa255a8800155eae1b9da54b08b345c2f10
|
| 3 |
+
size 5928
|
gguf/voices/English/Cillian_en-Irish_M_1.embedding.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f3125fae2f679fb5c8128d308ddd02986fd7d914975d17cc6dadf9eb130b1061
|
| 3 |
+
size 896
|
gguf/voices/English/Cillian_en-Irish_M_1.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d43eda64c4928dfb30ecf547e4a00100b8f83ed0eb5c1e07116725794da72ea2
|
| 3 |
+
size 852644
|
gguf/voices/English/Imogen_en-British_F_1.codes.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c906f6c405c9078607fe9c54cf1b71ef0e112839490499e34bf8791ed1ad7e86
|
| 3 |
+
size 7728
|
gguf/voices/English/Imogen_en-British_F_1.embedding.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:683424b91faeade48c8a92e05e1c3679366abbba443ef62eccba8dc40de48e9e
|
| 3 |
+
size 896
|
gguf/voices/English/Imogen_en-British_F_1.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dde4691a6ccd95f4928fbbc4e2d49089824b264575bc0bf0ac4986dd2f8082c5
|
| 3 |
+
size 1117244
|
gguf/voices/English/Jett_en-energetic_M_1.codes.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8f32e946509673ae422c37cc9b79957cb569f212360745f9110d343f6f1f9f2c
|
| 3 |
+
size 6608
|
gguf/voices/English/Jett_en-energetic_M_1.embedding.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:624b749d220dac7fdcda479b1964cb8f7de58c838c1357f2f0161dac06e34f48
|
| 3 |
+
size 896
|
gguf/voices/English/Jett_en-energetic_M_1.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:107c8350f191d47401da6844158528d0f493cbf46be4dee6bb81dd71bc052619
|
| 3 |
+
size 952604
|
gguf/voices/English/Morgana_en-witch_F_1.codes.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2f60227a9f7cdc6ce696548f4c28c199d3f5afeadf7317a1bde132575a1ed731
|
| 3 |
+
size 12448
|
gguf/voices/English/Morgana_en-witch_F_1.embedding.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a817387697c43138639aa1e52a9e42d9f3a629655e0faea5ac8565eb8e57c091
|
| 3 |
+
size 896
|
gguf/voices/English/Morgana_en-witch_F_1.mp3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8fec617ca7a4044f930c108ded0daf814ca93e61d97a7df626373aa182240ece
|
| 3 |
+
size 329812
|
gguf/voices/English/Rupert_en-British_M_1.codes.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:abf42171c7402fea5888ec350a50b64fd78777e44d7caf665c04e4e2b629fc79
|
| 3 |
+
size 8048
|
gguf/voices/English/Rupert_en-British_M_1.embedding.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0c68e699376dbafdc1ab0fadc133f48e7df9cb33b3a7b80b05f42c5c0aad8fa7
|
| 3 |
+
size 896
|
gguf/voices/English/Rupert_en-British_M_1.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3743b94cfc35c004fb5c2bcc9c2a537b605e3e15d4677e2b5b811bc18d95151e
|
| 3 |
+
size 1164284
|
gguf/voices/English/Sakura_en-anime_F_1.codes.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4b8c8288b45e1a3ff5aa3adc05a4af386f5889e9ab47902ff36f9daa8900ded2
|
| 3 |
+
size 9168
|
gguf/voices/English/Sakura_en-anime_F_1.embedding.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:daf3e82d2e6f2ce2fe90c73473a312a721295ae7f85763f99c87ee2f861063a9
|
| 3 |
+
size 896
|
gguf/voices/English/Sakura_en-anime_F_1.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:01e8f47241546efd1223bd1cac09ea81f1d9ae2cf3148260f37413f05db40229
|
| 3 |
+
size 1328924
|
gguf/voices/English/Savannah_en-American_F_1.codes.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f6e450fea8fa521c6be32b64fc61f48d6f9fd8301fe5759abd2315ce6ac946fb
|
| 3 |
+
size 6968
|
gguf/voices/English/Savannah_en-American_F_1.embedding.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd81587e66f96fb62da47ef13d5d867de464ca386a83f01b4403cdd19629ca6c
|
| 3 |
+
size 896
|
gguf/voices/English/Savannah_en-American_F_1.mp3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f8d5eaf8f6c18a0a3b452652da62d821f819d4e95f6347aa29986a503e219d14
|
| 3 |
+
size 183527
|
gguf/voices/English/Thaddeus_en-announcervoice_M_1.codes.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5c76de5bf16fd10282fced313c6c00a03fa2150b1df4c2382e19c6cbad65f179
|
| 3 |
+
size 9368
|
gguf/voices/English/Thaddeus_en-announcervoice_M_1.embedding.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fc8e21356730988aa20da70123991404345e0bdbfe3afe31e6371d8d93095461
|
| 3 |
+
size 896
|
gguf/voices/English/Thaddeus_en-announcervoice_M_1.mp3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c7fc96a18645b129b1e55659379ba796f5cfd70a800068b977d37f1254e93c52
|
| 3 |
+
size 247474
|
gguf/voices/English/Willow_en-asrm_F_1.codes.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:29ca4c14c16bb09b2bc48bb93f2d5a56327558d64271e54ebd1f713a4204905b
|
| 3 |
+
size 15048
|
gguf/voices/English/Willow_en-asrm_F_1.embedding.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6ebb84309f21f1dcbe6371e2355dc98f8ee3d0aee5850c305d06a73fe0a7c222
|
| 3 |
+
size 896
|
gguf/voices/English/Willow_en-asrm_F_1.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6a7574e4b91794390867027eb5c2c2a12f6cef557fc7d804b45f985144a50e5e
|
| 3 |
+
size 2193284
|
gguf/voices/English/Wyatt_en-American_M_1.codes.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7a854f9b66eefa9203a9f752c2a1ec808345b4d72ac34f8547ee776c58e4b72f
|
| 3 |
+
size 10248
|
gguf/voices/English/Wyatt_en-American_M_1.embedding.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b39d377b653ebaca6bbd4ecd13c382fe16d81f0cfae27217d9b626e5589a0511
|
| 3 |
+
size 896
|
gguf/voices/English/Wyatt_en-American_M_1.ogg
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:92c17f6eff4c9612ece96f0efae6f74f8e6b381f8c851ce739cdf0854a5a5831
|
| 3 |
+
size 147400
|
gguf/voices/French/Gaspard_fr_M_1.codes.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:34293ea302d3a47c17ab2f8514fb3a3354dc4ad695581d9b2f93fb37476ee1f7
|
| 3 |
+
size 11488
|