File size: 1,832 Bytes
26dc11c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 |
---
license: agpl-3.0
tags:
- rknn
---
# MeloTTS-RKNN2
## (English README see below)
在RK3588上运行MeloTTS文字转语音模型!
- 推理速度(RK3588): 约5倍速
- 内存占用(RK3588): 约0.2GB
## 使用方法
1. 克隆或者下载此仓库到瑞芯微SoC的系统上.
2. 安装依赖
```bash
pip install -r requirements.txt
pip install rknn-toolkit-lite2
```
4. 运行
```bash
python melotts_rknn.py -s "你想要生成的文本"
```
## 模型转换
1. 安装依赖
```bash
pip install -r requirements.txt
pip install rknn-toolkit2==2.3.0
```
2. 转换模型
```bash
python convert_rknn.py
```
## 已知问题
- 和原项目一样，Encoder部分并没有使用NPU加速，但是耗时不大，应该不会对推理速度有太大影响。
## 参考
- [melotts.axera](https://github.com/ml-inory/melotts.axera)
- [MeloTTS](https://github.com/myshell-ai/MeloTTS)
## English README
# MeloTTS-RKNN2
Run the MeloTTS text-to-speech model on RK3588!
- Inference speed (RK3588): about 5x real-time
- Memory usage (RK3588): about 0.2GB
## Usage
1. Clone or download this repository to your Rockchip SoC system.
2. Install dependencies
```bash
pip install -r requirements.txt
pip install rknn-toolkit-lite2
```
3. Run
```bash
python melotts_rknn.py -s "The text you want to generate."
```
## Model Conversion
1. Install dependencies
```bash
pip install -r requirements.txt
pip install rknn-toolkit2==2.3.0
```
2. Convert the model
```bash
python convert_rknn.py
```
## Known Issues
- Same as the original project, the Encoder part is not accelerated by the NPU. However, its processing time is short and should not significantly affect the inference speed.
## References
- [melotts.axera](https://github.com/ml-inory/melotts.axera)
- [MeloTTS](https://github.com/myshell-ai/MeloTTS)
|