xlians555 commited on
Commit
b5aac65
·
verified ·
1 Parent(s): 12d736c

Update org links to studio-dots-ai / dots-studio

Browse files
Files changed (1) hide show
  1. README.md +12 -12
README.md CHANGED
@@ -15,9 +15,9 @@ library_name: dots_tts
15
  # dots.tts-base
16
 
17
  <p align="left">
18
- <a href="https://github.com/rednote-hilab/dots.tts"><img src="https://img.shields.io/badge/GitHub-rednote--hilab%2Fdots.tts-blue?logo=github" alt="GitHub"></a>
19
- <a href="https://huggingface.co/spaces/rednote-hilab/dots.tts"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Spaces-Playground-orange" alt="Playground"></a>
20
- <a href="https://rednote-hilab.github.io/dots.tts-demo/"><img src="https://img.shields.io/badge/Demo%20Page-Live-red" alt="Demo Page"></a>
21
  </p>
22
 
23
  **dots.tts** is a **2B-parameter fully continuous, end-to-end autoregressive (AR) text-to-speech system**. The backbone pairs a semantic encoder, an LLM, and an autoregressive flow-matching acoustic head over a 48 kHz AudioVAE — no discrete codec tokens anywhere in the pipeline.
@@ -26,15 +26,15 @@ This repository hosts **`dots.tts-base`**, the **end-to-end pretrained checkpoin
26
 
27
  <table>
28
  <tr>
29
- <td align="left" valign="middle"><a href="https://huggingface.co/rednote-hilab/dots.tts-base"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-dots.tts--base-yellow" alt="dots.tts-base"></a></td>
30
  <td>← <em>you are here</em> — Pretrain (~1.5M h). Fine-tuning, full CFG / NFE control.</td>
31
  </tr>
32
  <tr>
33
- <td align="left" valign="middle"><a href="https://huggingface.co/rednote-hilab/dots.tts-soar"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-dots.tts--soar-yellow" alt="dots.tts-soar"></a></td>
34
  <td>+ Self-corrective Alignment. Highest zero-shot fidelity and speaker similarity; also recommended for fine-tuning.</td>
35
  </tr>
36
  <tr>
37
- <td align="left" valign="middle"><a href="https://huggingface.co/rednote-hilab/dots.tts-mf"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-dots.tts--mf-yellow" alt="dots.tts-mf"></a></td>
38
  <td>+ MeanFlow distillation. Few-step inference (NFE = 4), low latency.</td>
39
  </tr>
40
  </table>
@@ -50,8 +50,8 @@ conda create -n dots_tts python=3.10 -y
50
  conda activate dots_tts
51
 
52
  python -m pip install --upgrade pip
53
- python -m pip install "git+https://github.com/rednote-hilab/dots.tts.git" \
54
- -c "https://raw.githubusercontent.com/rednote-hilab/dots.tts/main/constraints/recommended.txt"
55
  ```
56
 
57
  ### CLI
@@ -59,7 +59,7 @@ python -m pip install "git+https://github.com/rednote-hilab/dots.tts.git" \
59
  ```bash
60
  # Continuation voice cloning (reference audio + transcript) — recommended
61
  dots.tts \
62
- --model-name-or-path rednote-hilab/dots.tts-base \
63
  --text "Hello, this is a zero-shot voice cloning demonstration." \
64
  --prompt-audio /path/to/reference.wav \
65
  --prompt-text "The exact transcript of the reference audio." \
@@ -73,7 +73,7 @@ from dots_tts.runtime import DotsTtsRuntime
73
  import soundfile as sf
74
 
75
  runtime = DotsTtsRuntime.from_pretrained(
76
- "rednote-hilab/dots.tts-base",
77
  precision="bfloat16",
78
  )
79
 
@@ -97,7 +97,7 @@ sf.write("output.wav", result["audio"].float().cpu().squeeze().numpy(), result["
97
 
98
  ### Fine-tuning
99
 
100
- `dots.tts-base` is the recommended starting point for fine-tuning. See the [training script](https://github.com/rednote-hilab/dots.tts/blob/main/scripts/train_dots_tts.py) and [smoke config](https://github.com/rednote-hilab/dots.tts/blob/main/configs/dots_tts.yaml) in the source repository:
101
 
102
  ```bash
103
  accelerate launch scripts/train_dots_tts.py --config configs/dots_tts.yaml
@@ -135,7 +135,7 @@ A frozen **AudioVAE** encodes 48 kHz mono waveform into a continuous latent and
135
  | VoxCPM 2 | 5.7 | 82.3 |
136
  | **dots.tts-base** | 6.6 | **83.5** |
137
 
138
- See the [project README](https://github.com/rednote-hilab/dots.tts#-performance) for the full per-language breakdown, CV3-Eval and EmergentTTS-Eval results.
139
 
140
  ---
141
 
 
15
  # dots.tts-base
16
 
17
  <p align="left">
18
+ <a href="https://github.com/studio-dots-ai/dots.tts"><img src="https://img.shields.io/badge/GitHub-studio--dots--ai%2Fdots.tts-blue?logo=github" alt="GitHub"></a>
19
+ <a href="https://huggingface.co/spaces/dots-studio/dots.tts"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Spaces-Playground-orange" alt="Playground"></a>
20
+ <a href="https://studio-dots-ai.github.io/dots.tts-demo/"><img src="https://img.shields.io/badge/Demo%20Page-Live-red" alt="Demo Page"></a>
21
  </p>
22
 
23
  **dots.tts** is a **2B-parameter fully continuous, end-to-end autoregressive (AR) text-to-speech system**. The backbone pairs a semantic encoder, an LLM, and an autoregressive flow-matching acoustic head over a 48 kHz AudioVAE — no discrete codec tokens anywhere in the pipeline.
 
26
 
27
  <table>
28
  <tr>
29
+ <td align="left" valign="middle"><a href="https://huggingface.co/dots-studio/dots.tts-base"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-dots.tts--base-yellow" alt="dots.tts-base"></a></td>
30
  <td>← <em>you are here</em> — Pretrain (~1.5M h). Fine-tuning, full CFG / NFE control.</td>
31
  </tr>
32
  <tr>
33
+ <td align="left" valign="middle"><a href="https://huggingface.co/dots-studio/dots.tts-soar"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-dots.tts--soar-yellow" alt="dots.tts-soar"></a></td>
34
  <td>+ Self-corrective Alignment. Highest zero-shot fidelity and speaker similarity; also recommended for fine-tuning.</td>
35
  </tr>
36
  <tr>
37
+ <td align="left" valign="middle"><a href="https://huggingface.co/dots-studio/dots.tts-mf"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-dots.tts--mf-yellow" alt="dots.tts-mf"></a></td>
38
  <td>+ MeanFlow distillation. Few-step inference (NFE = 4), low latency.</td>
39
  </tr>
40
  </table>
 
50
  conda activate dots_tts
51
 
52
  python -m pip install --upgrade pip
53
+ python -m pip install "git+https://github.com/studio-dots-ai/dots.tts.git" \
54
+ -c "https://raw.githubusercontent.com/studio-dots-ai/dots.tts/main/constraints/recommended.txt"
55
  ```
56
 
57
  ### CLI
 
59
  ```bash
60
  # Continuation voice cloning (reference audio + transcript) — recommended
61
  dots.tts \
62
+ --model-name-or-path dots-studio/dots.tts-base \
63
  --text "Hello, this is a zero-shot voice cloning demonstration." \
64
  --prompt-audio /path/to/reference.wav \
65
  --prompt-text "The exact transcript of the reference audio." \
 
73
  import soundfile as sf
74
 
75
  runtime = DotsTtsRuntime.from_pretrained(
76
+ "dots-studio/dots.tts-base",
77
  precision="bfloat16",
78
  )
79
 
 
97
 
98
  ### Fine-tuning
99
 
100
+ `dots.tts-base` is the recommended starting point for fine-tuning. See the [training script](https://github.com/studio-dots-ai/dots.tts/blob/main/scripts/train_dots_tts.py) and [smoke config](https://github.com/studio-dots-ai/dots.tts/blob/main/configs/dots_tts.yaml) in the source repository:
101
 
102
  ```bash
103
  accelerate launch scripts/train_dots_tts.py --config configs/dots_tts.yaml
 
135
  | VoxCPM 2 | 5.7 | 82.3 |
136
  | **dots.tts-base** | 6.6 | **83.5** |
137
 
138
+ See the [project README](https://github.com/studio-dots-ai/dots.tts#-performance) for the full per-language breakdown, CV3-Eval and EmergentTTS-Eval results.
139
 
140
  ---
141