Instructions to use ScreamingTony/AffectF5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- F5-TTS
How to use ScreamingTony/AffectF5 with F5-TTS:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Add AffectF5 model card
Browse files
README.md
ADDED
|
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-4.0
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
library_name: pytorch
|
| 6 |
+
pipeline_tag: text-to-speech
|
| 7 |
+
base_model: SWivid/F5-TTS
|
| 8 |
+
tags:
|
| 9 |
+
- text-to-speech
|
| 10 |
+
- f5-tts
|
| 11 |
+
- emotion
|
| 12 |
+
- arousal
|
| 13 |
+
- voice-cloning
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
# AffectF5 C7
|
| 17 |
+
|
| 18 |
+
AffectF5 为 F5-TTS v1 Base 增加五类 Emotion 与连续 Arousal 控制。该仓库只保存轻量 adapter;模型运行时仍需下载官方 F5-TTS v1 Base。
|
| 19 |
+
|
| 20 |
+
## 控制接口
|
| 21 |
+
|
| 22 |
+
- Emotion:Neutral、Happy、Sad、Angry、Surprise。
|
| 23 |
+
- Arousal:`[-1, 1]`。
|
| 24 |
+
- 推荐推理:32 NFE、global CFG 2、Emotion/A CFG 0、Arousal native scale 0.75、reference edge VAD。
|
| 25 |
+
|
| 26 |
+
## 使用
|
| 27 |
+
|
| 28 |
+
```bash
|
| 29 |
+
git clone https://github.com/ScreamingTony/AffectF5.git
|
| 30 |
+
cd AffectF5
|
| 31 |
+
pip install -e ".[demo]"
|
| 32 |
+
affect-f5 --help
|
| 33 |
+
```
|
| 34 |
+
|
| 35 |
+
## 权重内容
|
| 36 |
+
|
| 37 |
+
`affectf5_c7_adapter.safetensors` 包含:
|
| 38 |
+
|
| 39 |
+
- rank-16 LoRA;
|
| 40 |
+
- Emotion/Arousal conditioners;
|
| 41 |
+
- Layer 11/17 Emotion-FiLM;
|
| 42 |
+
- Layer 14/20 Arousal-residual。
|
| 43 |
+
|
| 44 |
+
不包含官方底模、优化器状态或训练数据。
|
| 45 |
+
|
| 46 |
+
## 许可与限制
|
| 47 |
+
|
| 48 |
+
Adapter 采用 CC BY-NC 4.0,仅供非商业用途。当前主要验证英文;自动情感指标不能替代人工听测。
|
| 49 |
+
|