Duplicate from zeroweight-ai/ZeroTTS
Browse filesCo-authored-by: Nguyen Minh Anh <minhanh29@users.noreply.huggingface.co>
This view is limited to 50 files because it contains too many changes. See raw diff
- .gitattributes +8 -0
- .gitignore +1 -0
- README.md +286 -0
- banner.png +3 -0
- config.json +22 -0
- null_voice_emb.npy +3 -0
- onnx/codec/LICENSE-Apache-2.0.txt +9 -0
- onnx/codec/codec_browser_onnx_meta.json +564 -0
- onnx/codec/moss_audio_tokenizer_decode_full.onnx +3 -0
- onnx/codec/moss_audio_tokenizer_decode_shared.data +3 -0
- onnx/codec/moss_audio_tokenizer_decode_step.onnx +3 -0
- onnx/local_frame_decode.onnx +3 -0
- onnx/prefix_step.onnx +3 -0
- onnx/text_encoder.onnx +3 -0
- samples/conversation.mp3 +3 -0
- samples/cross-lingual-english.mp3 +3 -0
- samples/cross-lingual-reference-vi.mp3 +3 -0
- samples/news-code-switch.mp3 +3 -0
- samples/storytelling.mp3 +3 -0
- silence_frame.npy +3 -0
- tokenizer.json +0 -0
- voices/baotrang/meta.json +15 -0
- voices/baotrang/preview.wav +3 -0
- voices/baotrang/voice.bin +3 -0
- voices/baotrang/voice.npz +3 -0
- voices/giahuy/meta.json +15 -0
- voices/giahuy/preview.wav +3 -0
- voices/giahuy/voice.bin +3 -0
- voices/giahuy/voice.npz +3 -0
- voices/hamy/meta.json +15 -0
- voices/hamy/preview.wav +3 -0
- voices/hamy/voice.bin +3 -0
- voices/hamy/voice.npz +3 -0
- voices/huuduc/meta.json +15 -0
- voices/huuduc/preview.wav +3 -0
- voices/huuduc/voice.bin +3 -0
- voices/huuduc/voice.npz +3 -0
- voices/index.json +164 -0
- voices/kimoanh/meta.json +15 -0
- voices/kimoanh/preview.wav +3 -0
- voices/kimoanh/voice.bin +3 -0
- voices/kimoanh/voice.npz +3 -0
- voices/maichi/meta.json +14 -0
- voices/maichi/preview.wav +3 -0
- voices/maichi/voice.bin +3 -0
- voices/maichi/voice.npz +3 -0
- voices/quangminh/meta.json +15 -0
- voices/quangminh/preview.wav +3 -0
- voices/quangminh/voice.bin +3 -0
- voices/quangminh/voice.npz +3 -0
.gitattributes
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.data filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
*.wav filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
*.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 8 |
+
banner.png filter=lfs diff=lfs merge=lfs -text
|
.gitignore
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
*.DS_Store
|
README.md
ADDED
|
@@ -0,0 +1,286 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
language:
|
| 4 |
+
- vi
|
| 5 |
+
library_name: onnx
|
| 6 |
+
pipeline_tag: text-to-speech
|
| 7 |
+
tags:
|
| 8 |
+
- text-to-speech
|
| 9 |
+
- tts
|
| 10 |
+
- vietnamese
|
| 11 |
+
- onnx
|
| 12 |
+
- onnxruntime
|
| 13 |
+
- zero-shot
|
| 14 |
+
- speech-synthesis
|
| 15 |
+
- voice-cloning
|
| 16 |
+
- vietnamese-tts
|
| 17 |
+
- tieng-viet
|
| 18 |
+
metrics:
|
| 19 |
+
- wer
|
| 20 |
+
model-index:
|
| 21 |
+
- name: ZeroTTS
|
| 22 |
+
results:
|
| 23 |
+
- task:
|
| 24 |
+
type: text-to-speech
|
| 25 |
+
name: Zero-Shot Text-to-Speech
|
| 26 |
+
dataset:
|
| 27 |
+
type: zeroweight-ai/ZeroBench-TTS
|
| 28 |
+
name: ZeroBench-TTS
|
| 29 |
+
split: test
|
| 30 |
+
metrics:
|
| 31 |
+
- type: wer
|
| 32 |
+
value: 1.03
|
| 33 |
+
name: WER (%) — raw text
|
| 34 |
+
- type: utmos
|
| 35 |
+
value: 2.91
|
| 36 |
+
name: UTMOSv2 naturalness MOS
|
| 37 |
+
- type: speaker_similarity
|
| 38 |
+
value: 0.936
|
| 39 |
+
name: Speaker similarity (WavLM-SV cosine)
|
| 40 |
+
- type: excess_silence
|
| 41 |
+
value: 0.029
|
| 42 |
+
name: Excess silence (s)
|
| 43 |
+
- task:
|
| 44 |
+
type: text-to-speech
|
| 45 |
+
name: Zero-Shot TTS — monolingual Vietnamese
|
| 46 |
+
dataset:
|
| 47 |
+
type: zeroweight-ai/ZeroBench-TTS
|
| 48 |
+
name: ZeroBench-TTS (vietnamese)
|
| 49 |
+
config: vietnamese
|
| 50 |
+
split: test
|
| 51 |
+
metrics:
|
| 52 |
+
- type: wer
|
| 53 |
+
value: 0.16
|
| 54 |
+
name: WER (%) — raw text
|
| 55 |
+
- task:
|
| 56 |
+
type: text-to-speech
|
| 57 |
+
name: Zero-Shot TTS — Vietnamese/English code-switching
|
| 58 |
+
dataset:
|
| 59 |
+
type: zeroweight-ai/ZeroBench-TTS
|
| 60 |
+
name: ZeroBench-TTS (code_switch)
|
| 61 |
+
config: code_switch
|
| 62 |
+
split: test
|
| 63 |
+
metrics:
|
| 64 |
+
- type: wer
|
| 65 |
+
value: 0.97
|
| 66 |
+
name: WER (%) — raw text
|
| 67 |
+
- task:
|
| 68 |
+
type: text-to-speech
|
| 69 |
+
name: Zero-Shot TTS — cross-lingual voice prompt
|
| 70 |
+
dataset:
|
| 71 |
+
type: zeroweight-ai/ZeroBench-TTS
|
| 72 |
+
name: ZeroBench-TTS (cross_lingual)
|
| 73 |
+
config: cross_lingual
|
| 74 |
+
split: test
|
| 75 |
+
metrics:
|
| 76 |
+
- type: wer
|
| 77 |
+
value: 1.42
|
| 78 |
+
name: WER (%) — raw text
|
| 79 |
+
- task:
|
| 80 |
+
type: text-to-speech
|
| 81 |
+
name: Zero-Shot TTS — acronyms, dates, numbers
|
| 82 |
+
dataset:
|
| 83 |
+
type: zeroweight-ai/ZeroBench-TTS
|
| 84 |
+
name: ZeroBench-TTS (challenging)
|
| 85 |
+
config: challenging
|
| 86 |
+
split: test
|
| 87 |
+
metrics:
|
| 88 |
+
- type: wer
|
| 89 |
+
value: 1.75
|
| 90 |
+
name: WER (%) — raw text
|
| 91 |
+
---
|
| 92 |
+
|
| 93 |
+
<img src="banner.png" alt="ZeroTTS — Vietnamese zero-shot text-to-speech" width="100%">
|
| 94 |
+
|
| 95 |
+
# ZeroTTS
|
| 96 |
+
|
| 97 |
+
### Vietnamese Zero-Shot Text-to-Speech (TTS) with real-time streaming and voice cloning from seconds of audio. Fast, natural, and optimised for CPU inference.
|
| 98 |
+
|
| 99 |
+
**The most accurate open Vietnamese TTS we know of — 4× fewer word errors than
|
| 100 |
+
the next best model**, and it runs faster than real time on a laptop CPU.
|
| 101 |
+
|
| 102 |
+
* 🎯 **Ultra-natural** — 2.91 UTMOS above every other open Vietnamese
|
| 103 |
+
system, with near-zero dead air (0.029 s).
|
| 104 |
+
* 🗣️ **Zero-shot voice cloning** — a voice is a small latent array; drop it in
|
| 105 |
+
and the model speaks in it, cloned from as little as 3 seconds of reference
|
| 106 |
+
audio (up to 30 seconds). No fine-tuning, no per-speaker training.
|
| 107 |
+
* ⚡ **Real-time on CPU, streaming** — ~2× faster than real time (RTF 0.5×),
|
| 108 |
+
first audio chunk in ~70 ms. No GPU required.
|
| 109 |
+
* 🇻🇳 **Built for Vietnamese** — tones, code-switched English, and
|
| 110 |
+
reads `31/12/2025` and `ZeroTTS` without text normalizer.
|
| 111 |
+
|
| 112 |
+
* Code, examples, browser demo: **https://github.com/zeroweight-ai/ZeroTTS**
|
| 113 |
+
* Benchmark dataset: **https://huggingface.co/datasets/zeroweight-ai/ZeroBench-TTS**
|
| 114 |
+
* Blogpost: **https://zeroweight.ai/blog/zero-tts**
|
| 115 |
+
|
| 116 |
+
## Samples
|
| 117 |
+
|
| 118 |
+
**Two-speaker conversation**
|
| 119 |
+
|
| 120 |
+
<audio controls><source src="https://huggingface.co/zeroweight-ai/ZeroTTS/resolve/main/samples/conversation.mp3" type="audio/mpeg"></audio>
|
| 121 |
+
|
| 122 |
+
**Long-form narration**
|
| 123 |
+
|
| 124 |
+
<audio controls><source src="https://huggingface.co/zeroweight-ai/ZeroTTS/resolve/main/samples/storytelling.mp3" type="audio/mpeg"></audio>
|
| 125 |
+
|
| 126 |
+
**News read, code-switched English**
|
| 127 |
+
|
| 128 |
+
<audio controls><source src="https://huggingface.co/zeroweight-ai/ZeroTTS/resolve/main/samples/news-code-switch.mp3" type="audio/mpeg"></audio>
|
| 129 |
+
|
| 130 |
+
**Cross-lingual**
|
| 131 |
+
|
| 132 |
+
Reference audio (Vietnamese)
|
| 133 |
+
|
| 134 |
+
<audio controls><source src="https://huggingface.co/zeroweight-ai/ZeroTTS/resolve/main/samples/cross-lingual-reference-vi.mp3" type="audio/mpeg"></audio>
|
| 135 |
+
|
| 136 |
+
Output (English)
|
| 137 |
+
|
| 138 |
+
<audio controls><source src="https://huggingface.co/zeroweight-ai/ZeroTTS/resolve/main/samples/cross-lingual-english.mp3" type="audio/mpeg"></audio>
|
| 139 |
+
|
| 140 |
+
## Usage
|
| 141 |
+
```python
|
| 142 |
+
pip install zerotts
|
| 143 |
+
```
|
| 144 |
+
|
| 145 |
+
```python
|
| 146 |
+
from zerotts import ZeroTTS
|
| 147 |
+
|
| 148 |
+
tts = ZeroTTS.from_pretrained("zeroweight-ai/ZeroTTS")
|
| 149 |
+
audio = tts.synthesize("Xin chào các bạn, mình là ZeroTTS.", voice="maichi")
|
| 150 |
+
tts.save_audio(audio, "out.wav")
|
| 151 |
+
```
|
| 152 |
+
|
| 153 |
+
Streaming, with first audio in roughly 70 ms:
|
| 154 |
+
|
| 155 |
+
```python
|
| 156 |
+
import queue
|
| 157 |
+
|
| 158 |
+
import numpy as np
|
| 159 |
+
import sounddevice as sd # pip install sounddevice
|
| 160 |
+
|
| 161 |
+
TEXT = ("Đây là chế độ phát trực tuyến. Âm thanh được tạo ra và phát ngay lập tức, "
|
| 162 |
+
"không cần chờ toàn bộ đoạn văn hoàn thành. Nhờ vậy, người nghe chỉ mất "
|
| 163 |
+
"khoảng 70 mili giây là đã nghe thấy câu đầu tiên, ngay cả khi mô "
|
| 164 |
+
"hình đang chạy trên CPU của một chiếc laptop bình thường.")
|
| 165 |
+
|
| 166 |
+
pending, tail = queue.Queue(), np.zeros(0, dtype="float32")
|
| 167 |
+
|
| 168 |
+
def feed(outdata, frames, _time, _status):
|
| 169 |
+
global tail
|
| 170 |
+
while len(tail) < frames and not pending.empty():
|
| 171 |
+
tail = np.concatenate([tail, pending.get_nowait()])
|
| 172 |
+
n = min(frames, len(tail))
|
| 173 |
+
outdata[:n, 0] = tail[:n]
|
| 174 |
+
outdata[n:] = 0
|
| 175 |
+
tail = tail[n:]
|
| 176 |
+
|
| 177 |
+
with sd.OutputStream(samplerate=tts.sample_rate, channels=1,
|
| 178 |
+
dtype="float32", callback=feed):
|
| 179 |
+
for chunk in tts.synthesize_stream(TEXT, voice="maichi"):
|
| 180 |
+
pending.put(chunk.reshape(-1)) # chunk is (1, n) float32 at 48 kHz
|
| 181 |
+
while not pending.empty() or len(tail):
|
| 182 |
+
sd.sleep(50) # let the buffer drain before closing
|
| 183 |
+
```
|
| 184 |
+
|
| 185 |
+
## Benchmarks
|
| 186 |
+
|
| 187 |
+
Measured on **[ZeroBench-TTS](https://huggingface.co/datasets/zeroweight-ai/ZeroBench-TTS)**
|
| 188 |
+
|
| 189 |
+
Every system reads **raw text** — dates, numbers and acronyms verbatim, exactly
|
| 190 |
+
as they appear in the wild, with no text frontend in front of the model.
|
| 191 |
+
|
| 192 |
+
| | **ZeroTTS** | OmniVoice | XTTS-v2-vietnamse | viXTTS |
|
| 193 |
+
|---|:-:|:-:|:-:|:-:|
|
| 194 |
+
| **WER** ↓ | **1.03 %** | 4.13 % | 16.42 % | 18.40 % |
|
| 195 |
+
| **Naturalness** (UTMOS) ↑ | **2.91** | 2.76 | 2.43 | 2.35 |
|
| 196 |
+
| **Voice similarity** (SSIM) ↑ | 0.936 | **0.950** | 0.940 | 0.935 |
|
| 197 |
+
| **Dead air** (excess silence) ↓ | **0.029 s** | 0.340 s | 0.532 s | 0.233 s |
|
| 198 |
+
| **RTF, CPU** ↓ | **0.50×** | 6.12× | 0.71× | 0.73× |
|
| 199 |
+
| **Time to first audio, CPU** ↓ | **~70 ms** | ~34 s | ~6.1 s | ~5.1 s |
|
| 200 |
+
| **Parameters** ↓ | **202 M** | 775 M | 467 M | 467 M |
|
| 201 |
+
|
| 202 |
+
**4× fewer word errors than the next-best system**, and the fastest of the four
|
| 203 |
+
on CPU. The gap is much wider in latency than in throughput: the two XTTS
|
| 204 |
+
fine-tunes also beat real time (0.71×) but need seconds to emit their first
|
| 205 |
+
sample, while OmniVoice is 6× *slower* than real time. All three are sized and
|
| 206 |
+
tuned for a GPU, and it shows.
|
| 207 |
+
|
| 208 |
+
Full comparison tables, per-subset breakdowns, and CPU speed methodology:
|
| 209 |
+
**[docs/BENCHMARKS.md](https://github.com/zeroweight-ai/ZeroTTS/blob/main/docs/BENCHMARKS.md)**
|
| 210 |
+
|
| 211 |
+
### Speed — CPU
|
| 212 |
+
|
| 213 |
+
RTF (realtime factor, wall-clock synthesis time ÷ output audio duration — lower
|
| 214 |
+
is faster; below 1× is faster than real time) and time-to-first-audio, all
|
| 215 |
+
measured **on CPU**, single request, 8 inference threads pinned to a dedicated
|
| 216 |
+
core pool (no other synthesis running concurrently). Three Vietnamese samples —
|
| 217 |
+
short (26 chars), medium (77 chars), long (227 chars) — each run 6 times with
|
| 218 |
+
the first 2 (cold-cache) discarded; figures below are the mean of the
|
| 219 |
+
remaining 4.
|
| 220 |
+
|
| 221 |
+
| | **ZeroTTS** | OmniVoice | XTTS-v2-vietnamse | viXTTS |
|
| 222 |
+
|---|:-:|:-:|:-:|:-:|
|
| 223 |
+
| RTF — short | **0.51×** | 10.87× | 0.70× | 0.71× |
|
| 224 |
+
| RTF — medium | **0.47×** | 4.82× | 0.70× | 0.70× |
|
| 225 |
+
| RTF — long | **0.53×** | 2.67× | 0.71× | 0.78× |
|
| 226 |
+
| TTFA — short | **53 ms** | 21.7 s | 4.02 s | 2.45 s |
|
| 227 |
+
| TTFA — medium | **66 ms** | 28.9 s | 4.02 s | 3.72 s |
|
| 228 |
+
| TTFA — long | **89 ms** | 52.3 s | 10.3 s | 9.22 s |
|
| 229 |
+
|
| 230 |
+
ZeroTTS's time-to-first-audio comes from its real streaming path — first audio frame,
|
| 231 |
+
not first full utterance. The three baselines have no working CPU streaming
|
| 232 |
+
path, so their TTFA is the time to the complete utterance.
|
| 233 |
+
|
| 234 |
+
## Voices, and voice cloning
|
| 235 |
+
|
| 236 |
+
A voice is a small array of speaker latents, `(1, n_voice_queries, d_model)`,
|
| 237 |
+
shipped as a `.npz` under `voices/`. That array is the entire speaker
|
| 238 |
+
conditioning — no reference transcript, no audio prompt.
|
| 239 |
+
|
| 240 |
+
> **Voice cloning is not available in this release.** Those latents come from a
|
| 241 |
+
> voice encoder that reads a reference clip, and that encoder is not published.
|
| 242 |
+
> This repository ships ready-to-use voices; it cannot create new ones from
|
| 243 |
+
> audio.
|
| 244 |
+
>
|
| 245 |
+
> To get latents for your own speaker, see **[zeroweight.ai](https://zeroweight.ai)**
|
| 246 |
+
> or get in touch.
|
| 247 |
+
|
| 248 |
+
Because a voice is just an array, latents obtained that way drop into
|
| 249 |
+
`voices/<name>/voice.npz` and work with no code change.
|
| 250 |
+
|
| 251 |
+
## Intended use and limitations
|
| 252 |
+
|
| 253 |
+
Built for **Vietnamese**. It handles English words embedded in Vietnamese text
|
| 254 |
+
(`code_switch`), but it is not an English TTS system and is not evaluated as one.
|
| 255 |
+
|
| 256 |
+
Do not use it to impersonate a real person, to generate speech attributed to
|
| 257 |
+
someone without their consent, or to produce audio intended to deceive. The
|
| 258 |
+
shipped voices are for evaluation and demos.
|
| 259 |
+
|
| 260 |
+
Synthetic speech should be disclosed as synthetic wherever a listener might
|
| 261 |
+
reasonably assume otherwise.
|
| 262 |
+
|
| 263 |
+
## Credits
|
| 264 |
+
|
| 265 |
+
Speech codec: **MOSS-Audio-Tokenizer-Nano** by the OpenMOSS team, Apache-2.0.
|
| 266 |
+
Its ONNX **decoder** graphs are redistributed under `onnx/codec/` so ZeroTTS has
|
| 267 |
+
no external runtime dependency; the encoder is not included. See
|
| 268 |
+
`onnx/codec/LICENSE-Apache-2.0.txt`.
|
| 269 |
+
|
| 270 |
+
```bibtex
|
| 271 |
+
@misc{gong2026mossaudiotokenizerscalingaudiotokenizers,
|
| 272 |
+
title={MOSS-Audio-Tokenizer: Scaling Audio Tokenizers for Future Audio Foundation Models},
|
| 273 |
+
author={Yitian Gong and Kuangwei Chen and Zhaoye Fei and Xiaogui Yang and Ke Chen
|
| 274 |
+
and Yang Wang and Kexin Huang and Mingshu Chen and Ruixiao Li
|
| 275 |
+
and Qingyuan Cheng and Shimin Li and Xipeng Qiu},
|
| 276 |
+
year={2026}, eprint={2602.10934}, archivePrefix={arXiv}, primaryClass={cs.SD}
|
| 277 |
+
}
|
| 278 |
+
```
|
| 279 |
+
|
| 280 |
+
## License
|
| 281 |
+
|
| 282 |
+
ZeroTTS weights and code: **MIT**.
|
| 283 |
+
|
| 284 |
+
The ZeroBench-TTS *dataset* is CC-BY-NC-4.0 because it redistributes reference
|
| 285 |
+
audio from VIVOS, viVoice, phoaudiobook and Emilia. That license applies to the
|
| 286 |
+
benchmark dataset only — **not** to these weights.
|
banner.png
ADDED
|
Git LFS Details
|
config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"text_format": "bpe",
|
| 3 |
+
"vocab_size": 8192,
|
| 4 |
+
"num_codebooks": 16,
|
| 5 |
+
"codebook_size": 1024,
|
| 6 |
+
"d_model": 768,
|
| 7 |
+
"n_heads": 12,
|
| 8 |
+
"n_layers": 9,
|
| 9 |
+
"n_voice_queries": 10,
|
| 10 |
+
"special_tokens": {
|
| 11 |
+
"<pad>": 0,
|
| 12 |
+
"<bos>": 1,
|
| 13 |
+
"<eot>": 2,
|
| 14 |
+
"<soa>": 3,
|
| 15 |
+
"<slot>": 4,
|
| 16 |
+
"<eoa>": 5,
|
| 17 |
+
"<en>": 6,
|
| 18 |
+
"<vi>": 7
|
| 19 |
+
},
|
| 20 |
+
"sample_rate": 48000,
|
| 21 |
+
"codec_frame_rate": 12.5
|
| 22 |
+
}
|
null_voice_emb.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ec014c14f79e8fc16f4ca6ea557f33fa2cbd7d448b5de5f467aace9ca0321e9f
|
| 3 |
+
size 30848
|
onnx/codec/LICENSE-Apache-2.0.txt
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
MOSS-Audio-Tokenizer-Nano
|
| 2 |
+
Copyright the OpenMOSS Team.
|
| 3 |
+
Licensed under the Apache License, Version 2.0.
|
| 4 |
+
https://www.apache.org/licenses/LICENSE-2.0
|
| 5 |
+
|
| 6 |
+
Source: https://huggingface.co/OpenMOSS-Team/MOSS-Audio-Tokenizer-Nano-ONNX
|
| 7 |
+
https://github.com/OpenMOSS/MOSS-Audio-Tokenizer
|
| 8 |
+
|
| 9 |
+
Only the DECODER graphs are redistributed here; the encoder is not included.
|
onnx/codec/codec_browser_onnx_meta.json
ADDED
|
@@ -0,0 +1,564 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"format_version": 2,
|
| 3 |
+
"checkpoint_path": "MOSS-Audio-Tokenizer-Nano",
|
| 4 |
+
"files": {
|
| 5 |
+
"decode_full": "moss_audio_tokenizer_decode_full.onnx",
|
| 6 |
+
"decode_step": "moss_audio_tokenizer_decode_step.onnx"
|
| 7 |
+
},
|
| 8 |
+
"external_data_files": {
|
| 9 |
+
"moss_audio_tokenizer_decode_full.onnx": [
|
| 10 |
+
"moss_audio_tokenizer_decode_shared.data"
|
| 11 |
+
],
|
| 12 |
+
"moss_audio_tokenizer_decode_step.onnx": [
|
| 13 |
+
"moss_audio_tokenizer_decode_shared.data"
|
| 14 |
+
]
|
| 15 |
+
},
|
| 16 |
+
"codec_config": {
|
| 17 |
+
"sample_rate": 48000,
|
| 18 |
+
"channels": 2,
|
| 19 |
+
"downsample_rate": 3840,
|
| 20 |
+
"num_quantizers": 16
|
| 21 |
+
},
|
| 22 |
+
"onnx": {
|
| 23 |
+
"opset": 17,
|
| 24 |
+
"decode_input_names": [
|
| 25 |
+
"audio_codes",
|
| 26 |
+
"audio_code_lengths"
|
| 27 |
+
],
|
| 28 |
+
"decode_output_names": [
|
| 29 |
+
"audio",
|
| 30 |
+
"audio_lengths"
|
| 31 |
+
],
|
| 32 |
+
"decode_step_input_names": [
|
| 33 |
+
"audio_codes",
|
| 34 |
+
"audio_code_lengths",
|
| 35 |
+
"transformer_offset_0",
|
| 36 |
+
"transformer_offset_1",
|
| 37 |
+
"transformer_offset_2",
|
| 38 |
+
"transformer_offset_3",
|
| 39 |
+
"attn_offset_0",
|
| 40 |
+
"attn_cached_keys_0",
|
| 41 |
+
"attn_cached_values_0",
|
| 42 |
+
"attn_cached_positions_0",
|
| 43 |
+
"attn_offset_1",
|
| 44 |
+
"attn_cached_keys_1",
|
| 45 |
+
"attn_cached_values_1",
|
| 46 |
+
"attn_cached_positions_1",
|
| 47 |
+
"attn_offset_2",
|
| 48 |
+
"attn_cached_keys_2",
|
| 49 |
+
"attn_cached_values_2",
|
| 50 |
+
"attn_cached_positions_2",
|
| 51 |
+
"attn_offset_3",
|
| 52 |
+
"attn_cached_keys_3",
|
| 53 |
+
"attn_cached_values_3",
|
| 54 |
+
"attn_cached_positions_3",
|
| 55 |
+
"attn_offset_4",
|
| 56 |
+
"attn_cached_keys_4",
|
| 57 |
+
"attn_cached_values_4",
|
| 58 |
+
"attn_cached_positions_4",
|
| 59 |
+
"attn_offset_5",
|
| 60 |
+
"attn_cached_keys_5",
|
| 61 |
+
"attn_cached_values_5",
|
| 62 |
+
"attn_cached_positions_5",
|
| 63 |
+
"attn_offset_6",
|
| 64 |
+
"attn_cached_keys_6",
|
| 65 |
+
"attn_cached_values_6",
|
| 66 |
+
"attn_cached_positions_6",
|
| 67 |
+
"attn_offset_7",
|
| 68 |
+
"attn_cached_keys_7",
|
| 69 |
+
"attn_cached_values_7",
|
| 70 |
+
"attn_cached_positions_7",
|
| 71 |
+
"attn_offset_8",
|
| 72 |
+
"attn_cached_keys_8",
|
| 73 |
+
"attn_cached_values_8",
|
| 74 |
+
"attn_cached_positions_8",
|
| 75 |
+
"attn_offset_9",
|
| 76 |
+
"attn_cached_keys_9",
|
| 77 |
+
"attn_cached_values_9",
|
| 78 |
+
"attn_cached_positions_9",
|
| 79 |
+
"attn_offset_10",
|
| 80 |
+
"attn_cached_keys_10",
|
| 81 |
+
"attn_cached_values_10",
|
| 82 |
+
"attn_cached_positions_10",
|
| 83 |
+
"attn_offset_11",
|
| 84 |
+
"attn_cached_keys_11",
|
| 85 |
+
"attn_cached_values_11",
|
| 86 |
+
"attn_cached_positions_11"
|
| 87 |
+
],
|
| 88 |
+
"decode_step_output_names": [
|
| 89 |
+
"audio",
|
| 90 |
+
"audio_lengths",
|
| 91 |
+
"transformer_offset_out_0",
|
| 92 |
+
"transformer_offset_out_1",
|
| 93 |
+
"transformer_offset_out_2",
|
| 94 |
+
"transformer_offset_out_3",
|
| 95 |
+
"attn_offset_out_0",
|
| 96 |
+
"attn_cached_keys_out_0",
|
| 97 |
+
"attn_cached_values_out_0",
|
| 98 |
+
"attn_cached_positions_out_0",
|
| 99 |
+
"attn_offset_out_1",
|
| 100 |
+
"attn_cached_keys_out_1",
|
| 101 |
+
"attn_cached_values_out_1",
|
| 102 |
+
"attn_cached_positions_out_1",
|
| 103 |
+
"attn_offset_out_2",
|
| 104 |
+
"attn_cached_keys_out_2",
|
| 105 |
+
"attn_cached_values_out_2",
|
| 106 |
+
"attn_cached_positions_out_2",
|
| 107 |
+
"attn_offset_out_3",
|
| 108 |
+
"attn_cached_keys_out_3",
|
| 109 |
+
"attn_cached_values_out_3",
|
| 110 |
+
"attn_cached_positions_out_3",
|
| 111 |
+
"attn_offset_out_4",
|
| 112 |
+
"attn_cached_keys_out_4",
|
| 113 |
+
"attn_cached_values_out_4",
|
| 114 |
+
"attn_cached_positions_out_4",
|
| 115 |
+
"attn_offset_out_5",
|
| 116 |
+
"attn_cached_keys_out_5",
|
| 117 |
+
"attn_cached_values_out_5",
|
| 118 |
+
"attn_cached_positions_out_5",
|
| 119 |
+
"attn_offset_out_6",
|
| 120 |
+
"attn_cached_keys_out_6",
|
| 121 |
+
"attn_cached_values_out_6",
|
| 122 |
+
"attn_cached_positions_out_6",
|
| 123 |
+
"attn_offset_out_7",
|
| 124 |
+
"attn_cached_keys_out_7",
|
| 125 |
+
"attn_cached_values_out_7",
|
| 126 |
+
"attn_cached_positions_out_7",
|
| 127 |
+
"attn_offset_out_8",
|
| 128 |
+
"attn_cached_keys_out_8",
|
| 129 |
+
"attn_cached_values_out_8",
|
| 130 |
+
"attn_cached_positions_out_8",
|
| 131 |
+
"attn_offset_out_9",
|
| 132 |
+
"attn_cached_keys_out_9",
|
| 133 |
+
"attn_cached_values_out_9",
|
| 134 |
+
"attn_cached_positions_out_9",
|
| 135 |
+
"attn_offset_out_10",
|
| 136 |
+
"attn_cached_keys_out_10",
|
| 137 |
+
"attn_cached_values_out_10",
|
| 138 |
+
"attn_cached_positions_out_10",
|
| 139 |
+
"attn_offset_out_11",
|
| 140 |
+
"attn_cached_keys_out_11",
|
| 141 |
+
"attn_cached_values_out_11",
|
| 142 |
+
"attn_cached_positions_out_11"
|
| 143 |
+
]
|
| 144 |
+
},
|
| 145 |
+
"streaming_decode": {
|
| 146 |
+
"batch_size": 1,
|
| 147 |
+
"transformer_offsets": [
|
| 148 |
+
{
|
| 149 |
+
"index": 0,
|
| 150 |
+
"decoder_index": 1,
|
| 151 |
+
"input_name": "transformer_offset_0",
|
| 152 |
+
"output_name": "transformer_offset_out_0",
|
| 153 |
+
"shape": [
|
| 154 |
+
1
|
| 155 |
+
],
|
| 156 |
+
"dtype": "int32"
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"index": 1,
|
| 160 |
+
"decoder_index": 3,
|
| 161 |
+
"input_name": "transformer_offset_1",
|
| 162 |
+
"output_name": "transformer_offset_out_1",
|
| 163 |
+
"shape": [
|
| 164 |
+
1
|
| 165 |
+
],
|
| 166 |
+
"dtype": "int32"
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"index": 2,
|
| 170 |
+
"decoder_index": 5,
|
| 171 |
+
"input_name": "transformer_offset_2",
|
| 172 |
+
"output_name": "transformer_offset_out_2",
|
| 173 |
+
"shape": [
|
| 174 |
+
1
|
| 175 |
+
],
|
| 176 |
+
"dtype": "int32"
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"index": 3,
|
| 180 |
+
"decoder_index": 7,
|
| 181 |
+
"input_name": "transformer_offset_3",
|
| 182 |
+
"output_name": "transformer_offset_out_3",
|
| 183 |
+
"shape": [
|
| 184 |
+
1
|
| 185 |
+
],
|
| 186 |
+
"dtype": "int32"
|
| 187 |
+
}
|
| 188 |
+
],
|
| 189 |
+
"attention_caches": [
|
| 190 |
+
{
|
| 191 |
+
"index": 0,
|
| 192 |
+
"decoder_index": 1,
|
| 193 |
+
"layer_index": 0,
|
| 194 |
+
"context": 500,
|
| 195 |
+
"num_heads": 4,
|
| 196 |
+
"head_dim": 64,
|
| 197 |
+
"offset_input_name": "attn_offset_0",
|
| 198 |
+
"offset_output_name": "attn_offset_out_0",
|
| 199 |
+
"cached_keys_input_name": "attn_cached_keys_0",
|
| 200 |
+
"cached_keys_output_name": "attn_cached_keys_out_0",
|
| 201 |
+
"cached_values_input_name": "attn_cached_values_0",
|
| 202 |
+
"cached_values_output_name": "attn_cached_values_out_0",
|
| 203 |
+
"cached_positions_input_name": "attn_cached_positions_0",
|
| 204 |
+
"cached_positions_output_name": "attn_cached_positions_out_0",
|
| 205 |
+
"offset_shape": [
|
| 206 |
+
1
|
| 207 |
+
],
|
| 208 |
+
"cache_shape": [
|
| 209 |
+
1,
|
| 210 |
+
4,
|
| 211 |
+
500,
|
| 212 |
+
64
|
| 213 |
+
],
|
| 214 |
+
"positions_shape": [
|
| 215 |
+
1,
|
| 216 |
+
500
|
| 217 |
+
],
|
| 218 |
+
"cache_dtype": "float32",
|
| 219 |
+
"positions_dtype": "int32"
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"index": 1,
|
| 223 |
+
"decoder_index": 1,
|
| 224 |
+
"layer_index": 1,
|
| 225 |
+
"context": 500,
|
| 226 |
+
"num_heads": 4,
|
| 227 |
+
"head_dim": 64,
|
| 228 |
+
"offset_input_name": "attn_offset_1",
|
| 229 |
+
"offset_output_name": "attn_offset_out_1",
|
| 230 |
+
"cached_keys_input_name": "attn_cached_keys_1",
|
| 231 |
+
"cached_keys_output_name": "attn_cached_keys_out_1",
|
| 232 |
+
"cached_values_input_name": "attn_cached_values_1",
|
| 233 |
+
"cached_values_output_name": "attn_cached_values_out_1",
|
| 234 |
+
"cached_positions_input_name": "attn_cached_positions_1",
|
| 235 |
+
"cached_positions_output_name": "attn_cached_positions_out_1",
|
| 236 |
+
"offset_shape": [
|
| 237 |
+
1
|
| 238 |
+
],
|
| 239 |
+
"cache_shape": [
|
| 240 |
+
1,
|
| 241 |
+
4,
|
| 242 |
+
500,
|
| 243 |
+
64
|
| 244 |
+
],
|
| 245 |
+
"positions_shape": [
|
| 246 |
+
1,
|
| 247 |
+
500
|
| 248 |
+
],
|
| 249 |
+
"cache_dtype": "float32",
|
| 250 |
+
"positions_dtype": "int32"
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"index": 2,
|
| 254 |
+
"decoder_index": 1,
|
| 255 |
+
"layer_index": 2,
|
| 256 |
+
"context": 500,
|
| 257 |
+
"num_heads": 4,
|
| 258 |
+
"head_dim": 64,
|
| 259 |
+
"offset_input_name": "attn_offset_2",
|
| 260 |
+
"offset_output_name": "attn_offset_out_2",
|
| 261 |
+
"cached_keys_input_name": "attn_cached_keys_2",
|
| 262 |
+
"cached_keys_output_name": "attn_cached_keys_out_2",
|
| 263 |
+
"cached_values_input_name": "attn_cached_values_2",
|
| 264 |
+
"cached_values_output_name": "attn_cached_values_out_2",
|
| 265 |
+
"cached_positions_input_name": "attn_cached_positions_2",
|
| 266 |
+
"cached_positions_output_name": "attn_cached_positions_out_2",
|
| 267 |
+
"offset_shape": [
|
| 268 |
+
1
|
| 269 |
+
],
|
| 270 |
+
"cache_shape": [
|
| 271 |
+
1,
|
| 272 |
+
4,
|
| 273 |
+
500,
|
| 274 |
+
64
|
| 275 |
+
],
|
| 276 |
+
"positions_shape": [
|
| 277 |
+
1,
|
| 278 |
+
500
|
| 279 |
+
],
|
| 280 |
+
"cache_dtype": "float32",
|
| 281 |
+
"positions_dtype": "int32"
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"index": 3,
|
| 285 |
+
"decoder_index": 1,
|
| 286 |
+
"layer_index": 3,
|
| 287 |
+
"context": 500,
|
| 288 |
+
"num_heads": 4,
|
| 289 |
+
"head_dim": 64,
|
| 290 |
+
"offset_input_name": "attn_offset_3",
|
| 291 |
+
"offset_output_name": "attn_offset_out_3",
|
| 292 |
+
"cached_keys_input_name": "attn_cached_keys_3",
|
| 293 |
+
"cached_keys_output_name": "attn_cached_keys_out_3",
|
| 294 |
+
"cached_values_input_name": "attn_cached_values_3",
|
| 295 |
+
"cached_values_output_name": "attn_cached_values_out_3",
|
| 296 |
+
"cached_positions_input_name": "attn_cached_positions_3",
|
| 297 |
+
"cached_positions_output_name": "attn_cached_positions_out_3",
|
| 298 |
+
"offset_shape": [
|
| 299 |
+
1
|
| 300 |
+
],
|
| 301 |
+
"cache_shape": [
|
| 302 |
+
1,
|
| 303 |
+
4,
|
| 304 |
+
500,
|
| 305 |
+
64
|
| 306 |
+
],
|
| 307 |
+
"positions_shape": [
|
| 308 |
+
1,
|
| 309 |
+
500
|
| 310 |
+
],
|
| 311 |
+
"cache_dtype": "float32",
|
| 312 |
+
"positions_dtype": "int32"
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"index": 4,
|
| 316 |
+
"decoder_index": 3,
|
| 317 |
+
"layer_index": 0,
|
| 318 |
+
"context": 800,
|
| 319 |
+
"num_heads": 4,
|
| 320 |
+
"head_dim": 64,
|
| 321 |
+
"offset_input_name": "attn_offset_4",
|
| 322 |
+
"offset_output_name": "attn_offset_out_4",
|
| 323 |
+
"cached_keys_input_name": "attn_cached_keys_4",
|
| 324 |
+
"cached_keys_output_name": "attn_cached_keys_out_4",
|
| 325 |
+
"cached_values_input_name": "attn_cached_values_4",
|
| 326 |
+
"cached_values_output_name": "attn_cached_values_out_4",
|
| 327 |
+
"cached_positions_input_name": "attn_cached_positions_4",
|
| 328 |
+
"cached_positions_output_name": "attn_cached_positions_out_4",
|
| 329 |
+
"offset_shape": [
|
| 330 |
+
1
|
| 331 |
+
],
|
| 332 |
+
"cache_shape": [
|
| 333 |
+
1,
|
| 334 |
+
4,
|
| 335 |
+
800,
|
| 336 |
+
64
|
| 337 |
+
],
|
| 338 |
+
"positions_shape": [
|
| 339 |
+
1,
|
| 340 |
+
800
|
| 341 |
+
],
|
| 342 |
+
"cache_dtype": "float32",
|
| 343 |
+
"positions_dtype": "int32"
|
| 344 |
+
},
|
| 345 |
+
{
|
| 346 |
+
"index": 5,
|
| 347 |
+
"decoder_index": 3,
|
| 348 |
+
"layer_index": 1,
|
| 349 |
+
"context": 800,
|
| 350 |
+
"num_heads": 4,
|
| 351 |
+
"head_dim": 64,
|
| 352 |
+
"offset_input_name": "attn_offset_5",
|
| 353 |
+
"offset_output_name": "attn_offset_out_5",
|
| 354 |
+
"cached_keys_input_name": "attn_cached_keys_5",
|
| 355 |
+
"cached_keys_output_name": "attn_cached_keys_out_5",
|
| 356 |
+
"cached_values_input_name": "attn_cached_values_5",
|
| 357 |
+
"cached_values_output_name": "attn_cached_values_out_5",
|
| 358 |
+
"cached_positions_input_name": "attn_cached_positions_5",
|
| 359 |
+
"cached_positions_output_name": "attn_cached_positions_out_5",
|
| 360 |
+
"offset_shape": [
|
| 361 |
+
1
|
| 362 |
+
],
|
| 363 |
+
"cache_shape": [
|
| 364 |
+
1,
|
| 365 |
+
4,
|
| 366 |
+
800,
|
| 367 |
+
64
|
| 368 |
+
],
|
| 369 |
+
"positions_shape": [
|
| 370 |
+
1,
|
| 371 |
+
800
|
| 372 |
+
],
|
| 373 |
+
"cache_dtype": "float32",
|
| 374 |
+
"positions_dtype": "int32"
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"index": 6,
|
| 378 |
+
"decoder_index": 5,
|
| 379 |
+
"layer_index": 0,
|
| 380 |
+
"context": 1200,
|
| 381 |
+
"num_heads": 4,
|
| 382 |
+
"head_dim": 64,
|
| 383 |
+
"offset_input_name": "attn_offset_6",
|
| 384 |
+
"offset_output_name": "attn_offset_out_6",
|
| 385 |
+
"cached_keys_input_name": "attn_cached_keys_6",
|
| 386 |
+
"cached_keys_output_name": "attn_cached_keys_out_6",
|
| 387 |
+
"cached_values_input_name": "attn_cached_values_6",
|
| 388 |
+
"cached_values_output_name": "attn_cached_values_out_6",
|
| 389 |
+
"cached_positions_input_name": "attn_cached_positions_6",
|
| 390 |
+
"cached_positions_output_name": "attn_cached_positions_out_6",
|
| 391 |
+
"offset_shape": [
|
| 392 |
+
1
|
| 393 |
+
],
|
| 394 |
+
"cache_shape": [
|
| 395 |
+
1,
|
| 396 |
+
4,
|
| 397 |
+
1200,
|
| 398 |
+
64
|
| 399 |
+
],
|
| 400 |
+
"positions_shape": [
|
| 401 |
+
1,
|
| 402 |
+
1200
|
| 403 |
+
],
|
| 404 |
+
"cache_dtype": "float32",
|
| 405 |
+
"positions_dtype": "int32"
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"index": 7,
|
| 409 |
+
"decoder_index": 5,
|
| 410 |
+
"layer_index": 1,
|
| 411 |
+
"context": 1200,
|
| 412 |
+
"num_heads": 4,
|
| 413 |
+
"head_dim": 64,
|
| 414 |
+
"offset_input_name": "attn_offset_7",
|
| 415 |
+
"offset_output_name": "attn_offset_out_7",
|
| 416 |
+
"cached_keys_input_name": "attn_cached_keys_7",
|
| 417 |
+
"cached_keys_output_name": "attn_cached_keys_out_7",
|
| 418 |
+
"cached_values_input_name": "attn_cached_values_7",
|
| 419 |
+
"cached_values_output_name": "attn_cached_values_out_7",
|
| 420 |
+
"cached_positions_input_name": "attn_cached_positions_7",
|
| 421 |
+
"cached_positions_output_name": "attn_cached_positions_out_7",
|
| 422 |
+
"offset_shape": [
|
| 423 |
+
1
|
| 424 |
+
],
|
| 425 |
+
"cache_shape": [
|
| 426 |
+
1,
|
| 427 |
+
4,
|
| 428 |
+
1200,
|
| 429 |
+
64
|
| 430 |
+
],
|
| 431 |
+
"positions_shape": [
|
| 432 |
+
1,
|
| 433 |
+
1200
|
| 434 |
+
],
|
| 435 |
+
"cache_dtype": "float32",
|
| 436 |
+
"positions_dtype": "int32"
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"index": 8,
|
| 440 |
+
"decoder_index": 7,
|
| 441 |
+
"layer_index": 0,
|
| 442 |
+
"context": 1600,
|
| 443 |
+
"num_heads": 4,
|
| 444 |
+
"head_dim": 64,
|
| 445 |
+
"offset_input_name": "attn_offset_8",
|
| 446 |
+
"offset_output_name": "attn_offset_out_8",
|
| 447 |
+
"cached_keys_input_name": "attn_cached_keys_8",
|
| 448 |
+
"cached_keys_output_name": "attn_cached_keys_out_8",
|
| 449 |
+
"cached_values_input_name": "attn_cached_values_8",
|
| 450 |
+
"cached_values_output_name": "attn_cached_values_out_8",
|
| 451 |
+
"cached_positions_input_name": "attn_cached_positions_8",
|
| 452 |
+
"cached_positions_output_name": "attn_cached_positions_out_8",
|
| 453 |
+
"offset_shape": [
|
| 454 |
+
1
|
| 455 |
+
],
|
| 456 |
+
"cache_shape": [
|
| 457 |
+
1,
|
| 458 |
+
4,
|
| 459 |
+
1600,
|
| 460 |
+
64
|
| 461 |
+
],
|
| 462 |
+
"positions_shape": [
|
| 463 |
+
1,
|
| 464 |
+
1600
|
| 465 |
+
],
|
| 466 |
+
"cache_dtype": "float32",
|
| 467 |
+
"positions_dtype": "int32"
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"index": 9,
|
| 471 |
+
"decoder_index": 7,
|
| 472 |
+
"layer_index": 1,
|
| 473 |
+
"context": 1600,
|
| 474 |
+
"num_heads": 4,
|
| 475 |
+
"head_dim": 64,
|
| 476 |
+
"offset_input_name": "attn_offset_9",
|
| 477 |
+
"offset_output_name": "attn_offset_out_9",
|
| 478 |
+
"cached_keys_input_name": "attn_cached_keys_9",
|
| 479 |
+
"cached_keys_output_name": "attn_cached_keys_out_9",
|
| 480 |
+
"cached_values_input_name": "attn_cached_values_9",
|
| 481 |
+
"cached_values_output_name": "attn_cached_values_out_9",
|
| 482 |
+
"cached_positions_input_name": "attn_cached_positions_9",
|
| 483 |
+
"cached_positions_output_name": "attn_cached_positions_out_9",
|
| 484 |
+
"offset_shape": [
|
| 485 |
+
1
|
| 486 |
+
],
|
| 487 |
+
"cache_shape": [
|
| 488 |
+
1,
|
| 489 |
+
4,
|
| 490 |
+
1600,
|
| 491 |
+
64
|
| 492 |
+
],
|
| 493 |
+
"positions_shape": [
|
| 494 |
+
1,
|
| 495 |
+
1600
|
| 496 |
+
],
|
| 497 |
+
"cache_dtype": "float32",
|
| 498 |
+
"positions_dtype": "int32"
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"index": 10,
|
| 502 |
+
"decoder_index": 7,
|
| 503 |
+
"layer_index": 2,
|
| 504 |
+
"context": 1600,
|
| 505 |
+
"num_heads": 4,
|
| 506 |
+
"head_dim": 64,
|
| 507 |
+
"offset_input_name": "attn_offset_10",
|
| 508 |
+
"offset_output_name": "attn_offset_out_10",
|
| 509 |
+
"cached_keys_input_name": "attn_cached_keys_10",
|
| 510 |
+
"cached_keys_output_name": "attn_cached_keys_out_10",
|
| 511 |
+
"cached_values_input_name": "attn_cached_values_10",
|
| 512 |
+
"cached_values_output_name": "attn_cached_values_out_10",
|
| 513 |
+
"cached_positions_input_name": "attn_cached_positions_10",
|
| 514 |
+
"cached_positions_output_name": "attn_cached_positions_out_10",
|
| 515 |
+
"offset_shape": [
|
| 516 |
+
1
|
| 517 |
+
],
|
| 518 |
+
"cache_shape": [
|
| 519 |
+
1,
|
| 520 |
+
4,
|
| 521 |
+
1600,
|
| 522 |
+
64
|
| 523 |
+
],
|
| 524 |
+
"positions_shape": [
|
| 525 |
+
1,
|
| 526 |
+
1600
|
| 527 |
+
],
|
| 528 |
+
"cache_dtype": "float32",
|
| 529 |
+
"positions_dtype": "int32"
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"index": 11,
|
| 533 |
+
"decoder_index": 7,
|
| 534 |
+
"layer_index": 3,
|
| 535 |
+
"context": 1600,
|
| 536 |
+
"num_heads": 4,
|
| 537 |
+
"head_dim": 64,
|
| 538 |
+
"offset_input_name": "attn_offset_11",
|
| 539 |
+
"offset_output_name": "attn_offset_out_11",
|
| 540 |
+
"cached_keys_input_name": "attn_cached_keys_11",
|
| 541 |
+
"cached_keys_output_name": "attn_cached_keys_out_11",
|
| 542 |
+
"cached_values_input_name": "attn_cached_values_11",
|
| 543 |
+
"cached_values_output_name": "attn_cached_values_out_11",
|
| 544 |
+
"cached_positions_input_name": "attn_cached_positions_11",
|
| 545 |
+
"cached_positions_output_name": "attn_cached_positions_out_11",
|
| 546 |
+
"offset_shape": [
|
| 547 |
+
1
|
| 548 |
+
],
|
| 549 |
+
"cache_shape": [
|
| 550 |
+
1,
|
| 551 |
+
4,
|
| 552 |
+
1600,
|
| 553 |
+
64
|
| 554 |
+
],
|
| 555 |
+
"positions_shape": [
|
| 556 |
+
1,
|
| 557 |
+
1600
|
| 558 |
+
],
|
| 559 |
+
"cache_dtype": "float32",
|
| 560 |
+
"positions_dtype": "int32"
|
| 561 |
+
}
|
| 562 |
+
]
|
| 563 |
+
}
|
| 564 |
+
}
|
onnx/codec/moss_audio_tokenizer_decode_full.onnx
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0fbbafe3fd4afa2a019af5c5ced204af6e2d1db044fa40f021525d2aee95b4ac
|
| 3 |
+
size 681902
|
onnx/codec/moss_audio_tokenizer_decode_shared.data
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e69d52e0f4e84ca27850557ee54face46632d3a5a16c89bd246c7c408466dcad
|
| 3 |
+
size 44198912
|
onnx/codec/moss_audio_tokenizer_decode_step.onnx
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9527c86a29e1837edec1f74db57d5eeaadb3a715af3382703566460afed25855
|
| 3 |
+
size 351400
|
onnx/local_frame_decode.onnx
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3c4540ef4e69dcf604dc6f60f6f99f119189694ff5f31113def20406e274f025
|
| 3 |
+
size 186743011
|
onnx/prefix_step.onnx
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b7544c9fdd3535b21fe8dea855407180ad2084f9da298f8a153893522496565b
|
| 3 |
+
size 348303894
|
onnx/text_encoder.onnx
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d37557a4abe07953a02686a176664486e73838389a14ff4a3f36b47ac2bccd52
|
| 3 |
+
size 323096725
|
samples/conversation.mp3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:692f94d47b6d411478a82f0de2e8d41f0348383f946e5b8f2343baa6b18c84c5
|
| 3 |
+
size 130412
|
samples/cross-lingual-english.mp3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3a318f85ed126ab59f17f22ac76552a42c5240f5c445270892630513a2db62c7
|
| 3 |
+
size 263468
|
samples/cross-lingual-reference-vi.mp3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b019bb424b186aaf26c721eb2a7ed68eeb3ca9cff8f3c3fe062151fbaaea003
|
| 3 |
+
size 34220
|
samples/news-code-switch.mp3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be58f3701e6169d33ee23502b0b4674be269f03c6ae113035b671a3d20f9783a
|
| 3 |
+
size 544172
|
samples/storytelling.mp3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f0ab526814b46d0bd6c9fdf1026338dc84089cf7e7d9da5552d334ba972d7775
|
| 3 |
+
size 971564
|
silence_frame.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bb2628efa99ae02819b04ff7509ef43ddb3213a9fb0916e4e88b8d1a7c3a1810
|
| 3 |
+
size 256
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
voices/baotrang/meta.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "baotrang",
|
| 3 |
+
"display_name": "Bảo Trang",
|
| 4 |
+
"language": "vi",
|
| 5 |
+
"gender": "nữ",
|
| 6 |
+
"description": "nữ, trưởng thành, tin tức, rõ ràng, trung tính",
|
| 7 |
+
"tags": [
|
| 8 |
+
"nữ",
|
| 9 |
+
"trưởng thành",
|
| 10 |
+
"tin tức",
|
| 11 |
+
"rõ ràng",
|
| 12 |
+
"trung tính"
|
| 13 |
+
],
|
| 14 |
+
"n_voice_queries": 10
|
| 15 |
+
}
|
voices/baotrang/preview.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cfe47bfeabb0177c9f346bae41a42c655ee4531184938d536317470be6ab5cfd
|
| 3 |
+
size 437804
|
voices/baotrang/voice.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:db21fd1342da2d7d213f6b81d4795eee1be13f65410da99cfef4c92f3a80e29f
|
| 3 |
+
size 30720
|
voices/baotrang/voice.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:abbc5807cc4767e7c62cc59f7b7cd04b4b7e0dba176086447a0c114578cf695a
|
| 3 |
+
size 31262
|
voices/giahuy/meta.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "giahuy",
|
| 3 |
+
"display_name": "Gia Huy",
|
| 4 |
+
"language": "vi",
|
| 5 |
+
"gender": "nam",
|
| 6 |
+
"description": "nam, trẻ, kể chuyện, trầm ấm, tâm tình",
|
| 7 |
+
"tags": [
|
| 8 |
+
"nam",
|
| 9 |
+
"trẻ",
|
| 10 |
+
"kể chuyện",
|
| 11 |
+
"trầm ấm",
|
| 12 |
+
"tâm tình"
|
| 13 |
+
],
|
| 14 |
+
"n_voice_queries": 10
|
| 15 |
+
}
|
voices/giahuy/preview.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:16feeedc4bc930bbe5aad4e657d1df78e60abc11fcac179a42f75c0d268b25d1
|
| 3 |
+
size 629804
|
voices/giahuy/voice.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1367686f7b5b154dc75b83d108487794c781812c6695a302f1b71afc821050ad
|
| 3 |
+
size 30720
|
voices/giahuy/voice.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3979a8369000eaaee1130801c1cacdadddc12e8d9a4a604be7430c4bcc6df32e
|
| 3 |
+
size 31262
|
voices/hamy/meta.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "hamy",
|
| 3 |
+
"display_name": "Hà My",
|
| 4 |
+
"language": "vi",
|
| 5 |
+
"gender": "nữ",
|
| 6 |
+
"description": "nữ, trẻ, hoạt hình, cao, biểu cảm",
|
| 7 |
+
"tags": [
|
| 8 |
+
"nữ",
|
| 9 |
+
"trẻ",
|
| 10 |
+
"hoạt hình",
|
| 11 |
+
"cao",
|
| 12 |
+
"biểu cảm"
|
| 13 |
+
],
|
| 14 |
+
"n_voice_queries": 10
|
| 15 |
+
}
|
voices/hamy/preview.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4bc33835023d7786bc75cb6c5c2681b1aec1ca3fb01300044b30d2161b29a171
|
| 3 |
+
size 460844
|
voices/hamy/voice.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e137f9544c20c491b3d139aeebb1d8549f8bd6f84f035a2df78830d3baec025d
|
| 3 |
+
size 30720
|
voices/hamy/voice.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a864c127386f74976154c840875f6ad615d74c377768bc6edcb75e3278f228b7
|
| 3 |
+
size 31262
|
voices/huuduc/meta.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "huuduc",
|
| 3 |
+
"display_name": "Hữu Đức",
|
| 4 |
+
"language": "vi",
|
| 5 |
+
"gender": "nam",
|
| 6 |
+
"description": "nam, lớn tuổi, kể chuyện, trầm, điềm đạm",
|
| 7 |
+
"tags": [
|
| 8 |
+
"nam",
|
| 9 |
+
"lớn tuổi",
|
| 10 |
+
"kể chuyện",
|
| 11 |
+
"trầm",
|
| 12 |
+
"điềm đạm"
|
| 13 |
+
],
|
| 14 |
+
"n_voice_queries": 10
|
| 15 |
+
}
|
voices/huuduc/preview.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8241e4cefd7bc526e101f0d0db7a793b9e2159ffae9f2f2e3c1cc94edebec184
|
| 3 |
+
size 491564
|
voices/huuduc/voice.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:82db4e47cea838d8678962b174ba5e141bf772f943e87f9115b303b1f8c3afe2
|
| 3 |
+
size 30720
|
voices/huuduc/voice.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7a6370180263093eb9dd009661ba24b5309f0ba94ec18d8ff2323c80616a7eb
|
| 3 |
+
size 31262
|
voices/index.json
ADDED
|
@@ -0,0 +1,164 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"voices": [
|
| 3 |
+
{
|
| 4 |
+
"name": "maichi",
|
| 5 |
+
"display_name": "Mai Chi",
|
| 6 |
+
"language": "vi",
|
| 7 |
+
"gender": "nữ",
|
| 8 |
+
"description": "nữ, trẻ, kể chuyện, nhẹ nhàng, thân thiện",
|
| 9 |
+
"tags": [
|
| 10 |
+
"nữ",
|
| 11 |
+
"trẻ",
|
| 12 |
+
"kể chuyện",
|
| 13 |
+
"nhẹ nhàng",
|
| 14 |
+
"thân thiện"
|
| 15 |
+
],
|
| 16 |
+
"n_voice_queries": 10,
|
| 17 |
+
"shape": [
|
| 18 |
+
1,
|
| 19 |
+
10,
|
| 20 |
+
768
|
| 21 |
+
]
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"name": "baotrang",
|
| 25 |
+
"display_name": "Bảo Trang",
|
| 26 |
+
"language": "vi",
|
| 27 |
+
"gender": "nữ",
|
| 28 |
+
"description": "nữ, trưởng thành, tin tức, rõ ràng, trung tính",
|
| 29 |
+
"tags": [
|
| 30 |
+
"nữ",
|
| 31 |
+
"trưởng thành",
|
| 32 |
+
"tin tức",
|
| 33 |
+
"rõ ràng",
|
| 34 |
+
"trung tính"
|
| 35 |
+
],
|
| 36 |
+
"n_voice_queries": 10,
|
| 37 |
+
"shape": [
|
| 38 |
+
1,
|
| 39 |
+
10,
|
| 40 |
+
768
|
| 41 |
+
]
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"name": "kimoanh",
|
| 45 |
+
"display_name": "Kim Oanh",
|
| 46 |
+
"language": "vi",
|
| 47 |
+
"gender": "nữ",
|
| 48 |
+
"description": "nữ, trung niên, kể chuyện, ấm áp, truyền cảm",
|
| 49 |
+
"tags": [
|
| 50 |
+
"nữ",
|
| 51 |
+
"trung niên",
|
| 52 |
+
"kể chuyện",
|
| 53 |
+
"ấm áp",
|
| 54 |
+
"truyền cảm"
|
| 55 |
+
],
|
| 56 |
+
"n_voice_queries": 10,
|
| 57 |
+
"shape": [
|
| 58 |
+
1,
|
| 59 |
+
10,
|
| 60 |
+
768
|
| 61 |
+
]
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"name": "giahuy",
|
| 65 |
+
"display_name": "Gia Huy",
|
| 66 |
+
"language": "vi",
|
| 67 |
+
"gender": "nam",
|
| 68 |
+
"description": "nam, trẻ, kể chuyện, trầm ấm, tâm tình",
|
| 69 |
+
"tags": [
|
| 70 |
+
"nam",
|
| 71 |
+
"trẻ",
|
| 72 |
+
"kể chuyện",
|
| 73 |
+
"trầm ấm",
|
| 74 |
+
"tâm tình"
|
| 75 |
+
],
|
| 76 |
+
"n_voice_queries": 10,
|
| 77 |
+
"shape": [
|
| 78 |
+
1,
|
| 79 |
+
10,
|
| 80 |
+
768
|
| 81 |
+
]
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"name": "huuduc",
|
| 85 |
+
"display_name": "Hữu Đức",
|
| 86 |
+
"language": "vi",
|
| 87 |
+
"gender": "nam",
|
| 88 |
+
"description": "nam, lớn tuổi, kể chuyện, trầm, điềm đạm",
|
| 89 |
+
"tags": [
|
| 90 |
+
"nam",
|
| 91 |
+
"lớn tuổi",
|
| 92 |
+
"kể chuyện",
|
| 93 |
+
"trầm",
|
| 94 |
+
"điềm đạm"
|
| 95 |
+
],
|
| 96 |
+
"n_voice_queries": 10,
|
| 97 |
+
"shape": [
|
| 98 |
+
1,
|
| 99 |
+
10,
|
| 100 |
+
768
|
| 101 |
+
]
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"name": "quangminh",
|
| 105 |
+
"display_name": "Quang Minh",
|
| 106 |
+
"language": "vi",
|
| 107 |
+
"gender": "nam",
|
| 108 |
+
"description": "nam, trẻ, tin tức, rõ ràng, dứt khoát",
|
| 109 |
+
"tags": [
|
| 110 |
+
"nam",
|
| 111 |
+
"trẻ",
|
| 112 |
+
"tin tức",
|
| 113 |
+
"rõ ràng",
|
| 114 |
+
"dứt khoát"
|
| 115 |
+
],
|
| 116 |
+
"n_voice_queries": 10,
|
| 117 |
+
"shape": [
|
| 118 |
+
1,
|
| 119 |
+
10,
|
| 120 |
+
768
|
| 121 |
+
]
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"name": "tiendat",
|
| 125 |
+
"display_name": "Tiến Đạt",
|
| 126 |
+
"language": "vi",
|
| 127 |
+
"gender": "nam",
|
| 128 |
+
"description": "nam, trẻ, bình luận, sôi nổi, năng lượng cao",
|
| 129 |
+
"tags": [
|
| 130 |
+
"nam",
|
| 131 |
+
"trẻ",
|
| 132 |
+
"bình luận",
|
| 133 |
+
"sôi nổi",
|
| 134 |
+
"năng lượng cao"
|
| 135 |
+
],
|
| 136 |
+
"n_voice_queries": 10,
|
| 137 |
+
"shape": [
|
| 138 |
+
1,
|
| 139 |
+
10,
|
| 140 |
+
768
|
| 141 |
+
]
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"name": "hamy",
|
| 145 |
+
"display_name": "Hà My",
|
| 146 |
+
"language": "vi",
|
| 147 |
+
"gender": "nữ",
|
| 148 |
+
"description": "nữ, trẻ, hoạt hình, cao, biểu cảm",
|
| 149 |
+
"tags": [
|
| 150 |
+
"nữ",
|
| 151 |
+
"trẻ",
|
| 152 |
+
"hoạt hình",
|
| 153 |
+
"cao",
|
| 154 |
+
"biểu cảm"
|
| 155 |
+
],
|
| 156 |
+
"n_voice_queries": 10,
|
| 157 |
+
"shape": [
|
| 158 |
+
1,
|
| 159 |
+
10,
|
| 160 |
+
768
|
| 161 |
+
]
|
| 162 |
+
}
|
| 163 |
+
]
|
| 164 |
+
}
|
voices/kimoanh/meta.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "kimoanh",
|
| 3 |
+
"display_name": "Kim Oanh",
|
| 4 |
+
"language": "vi",
|
| 5 |
+
"gender": "nữ",
|
| 6 |
+
"description": "nữ, trung niên, kể chuyện, ấm áp, truyền cảm",
|
| 7 |
+
"tags": [
|
| 8 |
+
"nữ",
|
| 9 |
+
"trung niên",
|
| 10 |
+
"kể chuyện",
|
| 11 |
+
"ấm áp",
|
| 12 |
+
"truyền cảm"
|
| 13 |
+
],
|
| 14 |
+
"n_voice_queries": 10
|
| 15 |
+
}
|
voices/kimoanh/preview.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:235bbe6aac7abbac191b2302ff52e51422076bf8117cd4142de1da7fe837a806
|
| 3 |
+
size 721964
|
voices/kimoanh/voice.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:125a9dd8580ee96902868ea75ace8f572950a80a03b49e1cf4583332668844a0
|
| 3 |
+
size 30720
|
voices/kimoanh/voice.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0aee413f25eae30123f930bd525d31a3b5a20c0803b4899305c2ff4a69466946
|
| 3 |
+
size 31262
|
voices/maichi/meta.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "maichi",
|
| 3 |
+
"display_name": "Mai Chi",
|
| 4 |
+
"language": "vi",
|
| 5 |
+
"gender": "nữ",
|
| 6 |
+
"description": "nữ, trẻ, nhẹ nhàng, thân thiện",
|
| 7 |
+
"tags": [
|
| 8 |
+
"nữ",
|
| 9 |
+
"trẻ",
|
| 10 |
+
"nhẹ nhàng",
|
| 11 |
+
"thân thiện"
|
| 12 |
+
],
|
| 13 |
+
"n_voice_queries": 10
|
| 14 |
+
}
|
voices/maichi/preview.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:978a62c9d93cd2a955005c5c9ffac6f3b8a37cc60fb85656e950c119d97c1dff
|
| 3 |
+
size 646210
|
voices/maichi/voice.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:487e082d0287d7f18414fe87839a9861a564757bf7932b413e9fb25ce2624f2b
|
| 3 |
+
size 30720
|
voices/maichi/voice.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d49e0a406d3a005e95f0a19476426973946c4a97e0087eefd638de222913ae3c
|
| 3 |
+
size 31222
|
voices/quangminh/meta.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"name": "quangminh",
|
| 3 |
+
"display_name": "Quang Minh",
|
| 4 |
+
"language": "vi",
|
| 5 |
+
"gender": "nam",
|
| 6 |
+
"description": "nam, trẻ, tin tức, rõ ràng, dứt khoát",
|
| 7 |
+
"tags": [
|
| 8 |
+
"nam",
|
| 9 |
+
"trẻ",
|
| 10 |
+
"tin tức",
|
| 11 |
+
"rõ ràng",
|
| 12 |
+
"dứt khoát"
|
| 13 |
+
],
|
| 14 |
+
"n_voice_queries": 10
|
| 15 |
+
}
|
voices/quangminh/preview.wav
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0257b6868a5ee37b2ef0f54d35d06c9d193c0358b2e66edc2ea7e1248608cb39
|
| 3 |
+
size 391724
|
voices/quangminh/voice.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5e1073679c7b82ce817f01ea5269307bbb9bb5c22717d218a419ca4185be390c
|
| 3 |
+
size 30720
|
voices/quangminh/voice.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d2acb18f831ade23ed2d02e7b749fa745c95e9ee305795c5ddeb75f93d35fdb
|
| 3 |
+
size 31262
|