Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- audio-to-audio
|
| 4 |
+
- text-to-speech
|
| 5 |
+
- speech-to-text
|
| 6 |
+
license: cc-by-nc-sa-4.0
|
| 7 |
+
language:
|
| 8 |
+
- zh
|
| 9 |
+
- en
|
| 10 |
+
- de
|
| 11 |
+
- ja
|
| 12 |
+
- fr
|
| 13 |
+
- es
|
| 14 |
+
- ko
|
| 15 |
+
- ar
|
| 16 |
+
pipeline_tag: audio-to-audio
|
| 17 |
+
inference: false
|
| 18 |
+
extra_gated_prompt: >-
|
| 19 |
+
You agree to not use the model to generate contents that violate DMCA or local
|
| 20 |
+
laws.
|
| 21 |
+
extra_gated_fields:
|
| 22 |
+
Country: country
|
| 23 |
+
Specific date: date_picker
|
| 24 |
+
I agree to use this model for non-commercial use ONLY: checkbox
|
| 25 |
+
---
|
| 26 |
+
|
| 27 |
+
# Fish Agent V0.1 3B
|
| 28 |
+
|
| 29 |
+
**Fish Agent V0.1 3B** is a groundbreaking Voice-to-Voice model capable of capturing and generating environmental audio information with unprecedented accuracy. What sets it apart is its semantic-token-free architecture, eliminating the need for traditional semantic encoders/decoders like Whisper and CosyVoice.
|
| 30 |
+
|
| 31 |
+
Additionally, it stands as a state-of-the-art text-to-speech (TTS) model, trained on an extensive dataset of 700,000 hours of multilingual audio content.
|
| 32 |
+
|
| 33 |
+
## Supported Languages
|
| 34 |
+
The model supports the following languages with their respective training data sizes:
|
| 35 |
+
- English (en): ~300,000 hours
|
| 36 |
+
- Chinese (zh): ~300,000 hours
|
| 37 |
+
- German (de): ~20,000 hours
|
| 38 |
+
- Japanese (ja): ~20,000 hours
|
| 39 |
+
- French (fr): ~20,000 hours
|
| 40 |
+
- Spanish (es): ~20,000 hours
|
| 41 |
+
- Korean (ko): ~20,000 hours
|
| 42 |
+
- Arabic (ar): ~20,000 hours
|
| 43 |
+
|
| 44 |
+
For detailed information and implementation guidelines, please visit our [Fish Speech GitHub repository](https://github.com/fishaudio/fish-speech).
|
| 45 |
+
|
| 46 |
+
## Citation
|
| 47 |
+
If you find this repository helpful in your work, please consider citing:
|
| 48 |
+
|
| 49 |
+
```bibtex
|
| 50 |
+
@misc{fish-agent-0.1,
|
| 51 |
+
author = {Shijia Liao and Tianyu Li and Rcell and others},
|
| 52 |
+
title = {Fish Agent V0.1 3B},
|
| 53 |
+
year = {2024},
|
| 54 |
+
publisher = {GitHub},
|
| 55 |
+
journal = {GitHub repository},
|
| 56 |
+
howpublished = {\url{https://github.com/fishaudio/fish-speech}}
|
| 57 |
+
}
|
| 58 |
+
```
|
| 59 |
+
|
| 60 |
+
## License
|
| 61 |
+
This model and its associated code are released under the BY-CC-NC-SA-4.0 license, allowing for non-commercial use with appropriate attribution.
|