Upload folder using huggingface_hub
Browse files- README.md +197 -36
- infer/modules/vc/__init__.py +0 -5
- infer/modules/vc/modules.py +60 -35
- infer/modules/vc/pipeline.py +96 -83
- infer/modules/vc/utils.py +4 -5
- requirements.txt +8 -7
README.md
CHANGED
@@ -1,52 +1,213 @@
|
|
1 |
-
|
2 |
-
license: mit
|
3 |
-
---
|
4 |
|
|
|
|
|
5 |
|
6 |
-
|
|
|
7 |
|
8 |
-
|
9 |
|
10 |
-
|
|
|
|
|
11 |
|
12 |
-
|
13 |
-
- Python 3.8 or higher
|
14 |
|
15 |
-
|
16 |
-
1. **Install Pytorch**:
|
17 |
-
```bash
|
18 |
-
pip install torch torchvision torchaudio
|
19 |
-
```
|
20 |
-
|
21 |
|
22 |
-
|
23 |
-
```bash
|
24 |
-
pip install -r requirements.txt
|
25 |
-
```
|
26 |
-
|
27 |
|
28 |
-
|
29 |
|
30 |
-
|
31 |
-
- **Download Assets**:
|
32 |
-
Download necessary models and files using the scripts in the `tools` directory.
|
33 |
-
- **Install FFmpeg**:
|
34 |
-
```bash
|
35 |
-
sudo apt install ffmpeg
|
36 |
-
```
|
37 |
|
38 |
-
|
39 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
```bash
|
41 |
-
python
|
42 |
```
|
43 |
|
|
|
|
|
|
|
|
|
44 |
|
45 |
-
|
46 |
-
-
|
47 |
-
|
48 |
-
-
|
49 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
50 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
51 |
|
52 |
-
|
|
|
|
|
|
|
|
1 |
+
<div align="center">
|
|
|
|
|
2 |
|
3 |
+
<h1>Retrieval-based-Voice-Conversion-WebUI</h1>
|
4 |
+
一个基于VITS的简单易用的变声框架<br><br>
|
5 |
|
6 |
+
[![madewithlove](https://img.shields.io/badge/made_with-%E2%9D%A4-red?style=for-the-badge&labelColor=orange
|
7 |
+
)](https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI)
|
8 |
|
9 |
+
<img src="https://counter.seku.su/cmoe?name=rvc&theme=r34" /><br>
|
10 |
|
11 |
+
[![Open In Colab](https://img.shields.io/badge/Colab-F9AB00?style=for-the-badge&logo=googlecolab&color=525252)](https://colab.research.google.com/github/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/blob/main/Retrieval_based_Voice_Conversion_WebUI.ipynb)
|
12 |
+
[![Licence](https://img.shields.io/badge/LICENSE-MIT-green.svg?style=for-the-badge)](https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/blob/main/LICENSE)
|
13 |
+
[![Huggingface](https://img.shields.io/badge/🤗%20-Spaces-yellow.svg?style=for-the-badge)](https://huggingface.co/lj1995/VoiceConversionWebUI/tree/main/)
|
14 |
|
15 |
+
[![Discord](https://img.shields.io/badge/RVC%20Developers-Discord-7289DA?style=for-the-badge&logo=discord&logoColor=white)](https://discord.gg/HcsmBBGyVk)
|
|
|
16 |
|
17 |
+
[**更新日志**](https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/blob/main/docs/Changelog_CN.md) | [**常见问题解答**](https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/wiki/%E5%B8%B8%E8%A7%81%E9%97%AE%E9%A2%98%E8%A7%A3%E7%AD%94) | [**AutoDL·5毛钱训练AI歌手**](https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/wiki/Autodl%E8%AE%AD%E7%BB%83RVC%C2%B7AI%E6%AD%8C%E6%89%8B%E6%95%99%E7%A8%8B) | [**对照实验记录**](https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/wiki/Autodl%E8%AE%AD%E7%BB%83RVC%C2%B7AI%E6%AD%8C%E6%89%8B%E6%95%99%E7%A8%8B](https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/wiki/%E5%AF%B9%E7%85%A7%E5%AE%9E%E9%AA%8C%C2%B7%E5%AE%9E%E9%AA%8C%E8%AE%B0%E5%BD%95)) | [**在线演示**](https://modelscope.cn/studios/FlowerCry/RVCv2demo)
|
|
|
|
|
|
|
|
|
|
|
18 |
|
19 |
+
[**English**](./docs/en/README.en.md) | [**中文简体**](./README.md) | [**日本語**](./docs/jp/README.ja.md) | [**한국어**](./docs/kr/README.ko.md) ([**韓國語**](./docs/kr/README.ko.han.md)) | [**Français**](./docs/fr/README.fr.md) | [**Türkçe**](./docs/tr/README.tr.md) | [**Português**](./docs/pt/README.pt.md)
|
|
|
|
|
|
|
|
|
20 |
|
21 |
+
</div>
|
22 |
|
23 |
+
> 底模使用接近50小时的开源高质量VCTK训练集训练,无版权方面的顾虑,请大家放心使用
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
|
25 |
+
> 请期待RVCv3的底模,参数更大,数据更大,效果更好,基本持平的推理速度,需要训练数据量更少。
|
26 |
+
|
27 |
+
<table>
|
28 |
+
<tr>
|
29 |
+
<td align="center">训练推理界面</td>
|
30 |
+
<td align="center">实时变声界面</td>
|
31 |
+
</tr>
|
32 |
+
<tr>
|
33 |
+
<td align="center"><img src="https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/assets/129054828/092e5c12-0d49-4168-a590-0b0ef6a4f630"></td>
|
34 |
+
<td align="center"><img src="https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/assets/129054828/730b4114-8805-44a1-ab1a-04668f3c30a6"></td>
|
35 |
+
</tr>
|
36 |
+
<tr>
|
37 |
+
<td align="center">go-web.bat</td>
|
38 |
+
<td align="center">go-realtime-gui.bat</td>
|
39 |
+
</tr>
|
40 |
+
<tr>
|
41 |
+
<td align="center">可以自由选择想要执行的操作。</td>
|
42 |
+
<td align="center">我们已经实现端到端170ms延迟。如使用ASIO输入输出设备,已能实现端到端90ms延迟,但非常依赖硬件驱动支持。</td>
|
43 |
+
</tr>
|
44 |
+
</table>
|
45 |
+
|
46 |
+
## 简介
|
47 |
+
本仓库具有以下特点
|
48 |
+
+ 使用top1检索替换输入源特征为训练集特征来杜绝音色泄漏
|
49 |
+
+ 即便在相对较差的显卡上也能快速训练
|
50 |
+
+ 使用少量数据进行训练也能得到较好结果(推荐至少收集10分钟低底噪语音数据)
|
51 |
+
+ 可以通过模型融合来改变音色(借助ckpt处理选项卡中的ckpt-merge)
|
52 |
+
+ 简单易用的网页界面
|
53 |
+
+ 可调用UVR5模型来快速分离人声和伴奏
|
54 |
+
+ 使用最先进的[人声音高提取算法InterSpeech2023-RMVPE](#参考项目)根绝哑音问题。效果最好(显著地)但比crepe_full更快、资源占用更小
|
55 |
+
+ A卡I卡加速支持
|
56 |
+
|
57 |
+
点此查看我们的[演示��频](https://www.bilibili.com/video/BV1pm4y1z7Gm/) !
|
58 |
+
|
59 |
+
## 环境配置
|
60 |
+
以下指令需在 Python 版本大于3.8的环境中执行。
|
61 |
+
|
62 |
+
### Windows/Linux/MacOS等平台通用方法
|
63 |
+
下列方法任选其一。
|
64 |
+
#### 1. 通过 pip 安装依赖
|
65 |
+
1. 安装Pytorch及其核心依赖,若已安装则跳过。参考自: https://pytorch.org/get-started/locally/
|
66 |
+
```bash
|
67 |
+
pip install torch torchvision torchaudio
|
68 |
+
```
|
69 |
+
2. 如果是 win 系统 + Nvidia Ampere 架构(RTX30xx),根据 #21 的经验,需要指定 pytorch 对应的 cuda 版本
|
70 |
+
```bash
|
71 |
+
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu117
|
72 |
+
```
|
73 |
+
3. 根据自己的显卡安装对应依赖
|
74 |
+
- N卡
|
75 |
+
```bash
|
76 |
+
pip install -r requirements.txt
|
77 |
+
```
|
78 |
+
- A卡/I卡
|
79 |
+
```bash
|
80 |
+
pip install -r requirements-dml.txt
|
81 |
+
```
|
82 |
+
- A卡ROCM(Linux)
|
83 |
+
```bash
|
84 |
+
pip install -r requirements-amd.txt
|
85 |
+
```
|
86 |
+
- I卡IPEX(Linux)
|
87 |
+
```bash
|
88 |
+
pip install -r requirements-ipex.txt
|
89 |
+
```
|
90 |
+
|
91 |
+
#### 2. 通过 poetry 来安装依赖
|
92 |
+
安装 Poetry 依赖管理工具,若已安装则跳过。参考自: https://python-poetry.org/docs/#installation
|
93 |
+
```bash
|
94 |
+
curl -sSL https://install.python-poetry.org | python3 -
|
95 |
+
```
|
96 |
+
|
97 |
+
通过 Poetry 安装依赖时,python 建议使用 3.7-3.10 版本,其余版本在安装 llvmlite==0.39.0 时会出现冲突
|
98 |
+
```bash
|
99 |
+
poetry init -n
|
100 |
+
poetry env use "path to your python.exe"
|
101 |
+
poetry run pip install -r requirments.txt
|
102 |
+
```
|
103 |
+
|
104 |
+
### MacOS
|
105 |
+
可以通过 `run.sh` 来安装依赖
|
106 |
+
```bash
|
107 |
+
sh ./run.sh
|
108 |
+
```
|
109 |
+
|
110 |
+
## 其他预模型准备
|
111 |
+
RVC需要其他一些预模型来推理和训练。
|
112 |
+
|
113 |
+
你可以从我们的[Hugging Face space](https://huggingface.co/lj1995/VoiceConversionWebUI/tree/main/)下载到这些模型。
|
114 |
+
|
115 |
+
### 1. 下载 assets
|
116 |
+
以下是一份清单,包括了所有RVC所需的预模型和其他文件的名称。你可以在`tools`文件夹找到下载它们的脚本。
|
117 |
+
|
118 |
+
- ./assets/hubert/hubert_base.pt
|
119 |
+
|
120 |
+
- ./assets/pretrained
|
121 |
+
|
122 |
+
- ./assets/uvr5_weights
|
123 |
+
|
124 |
+
想使用v2版本模型的话,需要额外下载
|
125 |
+
|
126 |
+
- ./assets/pretrained_v2
|
127 |
+
|
128 |
+
### 2. 安装 ffmpeg
|
129 |
+
若ffmpeg和ffprobe已安装则跳过。
|
130 |
+
|
131 |
+
#### Ubuntu/Debian 用户
|
132 |
+
```bash
|
133 |
+
sudo apt install ffmpeg
|
134 |
+
```
|
135 |
+
#### MacOS 用户
|
136 |
+
```bash
|
137 |
+
brew install ffmpeg
|
138 |
+
```
|
139 |
+
#### Windows 用户
|
140 |
+
下载后放置在根目录。
|
141 |
+
- 下载[ffmpeg.exe](https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/ffmpeg.exe)
|
142 |
+
|
143 |
+
- 下载[ffprobe.exe](https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/ffprobe.exe)
|
144 |
+
|
145 |
+
### 3. 下载 rmvpe 人声音高提取算法所需文件
|
146 |
+
|
147 |
+
如果你想使用最新的RMVPE人声音高提取算法,则你需要下载音高提取模型参数并放置于RVC根目录。
|
148 |
+
|
149 |
+
- 下载[rmvpe.pt](https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/rmvpe.pt)
|
150 |
+
|
151 |
+
#### 下载 rmvpe 的 dml 环境(可选, A卡/I卡用户)
|
152 |
+
|
153 |
+
- 下载[rmvpe.onnx](https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/rmvpe.onnx)
|
154 |
+
|
155 |
+
### 4. AMD显卡Rocm(可选, 仅Linux)
|
156 |
+
|
157 |
+
如果你想基于AMD的Rocm技术在Linux系统上运行RVC,请先在[这里](https://rocm.docs.amd.com/en/latest/deploy/linux/os-native/install.html)安装所需的驱动。
|
158 |
+
|
159 |
+
若你使用的是Arch Linux,可以使用pacman来安装所需驱动:
|
160 |
+
````
|
161 |
+
pacman -S rocm-hip-sdk rocm-opencl-sdk
|
162 |
+
````
|
163 |
+
对于某些型号的显卡,你可能需要额外配置如下的环境变量(如:RX6700XT):
|
164 |
+
````
|
165 |
+
export ROCM_PATH=/opt/rocm
|
166 |
+
export HSA_OVERRIDE_GFX_VERSION=10.3.0
|
167 |
+
````
|
168 |
+
同时确保你的当前用户处于`render`与`video`用户组内:
|
169 |
+
````
|
170 |
+
sudo usermod -aG render $USERNAME
|
171 |
+
sudo usermod -aG video $USERNAME
|
172 |
+
````
|
173 |
+
|
174 |
+
## 开始使用
|
175 |
+
### 直接启动
|
176 |
+
使用以下指令来启动 WebUI
|
177 |
```bash
|
178 |
+
python infer-web.py
|
179 |
```
|
180 |
|
181 |
+
若先前使用 Poetry 安装依赖,则可以通过以下方式启动WebUI
|
182 |
+
```bash
|
183 |
+
poetry run python infer-web.py
|
184 |
+
```
|
185 |
|
186 |
+
### 使用整合包
|
187 |
+
下载并解压`RVC-beta.7z`
|
188 |
+
#### Windows 用户
|
189 |
+
双击`go-web.bat`
|
190 |
+
#### MacOS 用户
|
191 |
+
```bash
|
192 |
+
sh ./run.sh
|
193 |
+
```
|
194 |
+
### 对于需要使用IPEX技术的I卡用户(仅Linux)
|
195 |
+
```bash
|
196 |
+
source /opt/intel/oneapi/setvars.sh
|
197 |
+
```
|
198 |
|
199 |
+
## 参考项目
|
200 |
+
+ [ContentVec](https://github.com/auspicious3000/contentvec/)
|
201 |
+
+ [VITS](https://github.com/jaywalnut310/vits)
|
202 |
+
+ [HIFIGAN](https://github.com/jik876/hifi-gan)
|
203 |
+
+ [Gradio](https://github.com/gradio-app/gradio)
|
204 |
+
+ [FFmpeg](https://github.com/FFmpeg/FFmpeg)
|
205 |
+
+ [Ultimate Vocal Remover](https://github.com/Anjok07/ultimatevocalremovergui)
|
206 |
+
+ [audio-slicer](https://github.com/openvpi/audio-slicer)
|
207 |
+
+ [Vocal pitch extraction:RMVPE](https://github.com/Dream-High/RMVPE)
|
208 |
+
+ The pretrained model is trained and tested by [yxlllc](https://github.com/yxlllc/RMVPE) and [RVC-Boss](https://github.com/RVC-Boss).
|
209 |
|
210 |
+
## 感谢所有贡献者作出的努力
|
211 |
+
<a href="https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI/graphs/contributors" target="_blank">
|
212 |
+
<img src="https://contrib.rocks/image?repo=RVC-Project/Retrieval-based-Voice-Conversion-WebUI" />
|
213 |
+
</a>
|
infer/modules/vc/__init__.py
CHANGED
@@ -1,5 +0,0 @@
|
|
1 |
-
from .pipeline import Pipeline
|
2 |
-
from .modules import VC
|
3 |
-
from .utils import get_index_path_from_model, load_hubert
|
4 |
-
from .info import show_info
|
5 |
-
from .hash import model_hash_ckpt, hash_id, hash_similarity
|
|
|
|
|
|
|
|
|
|
|
|
infer/modules/vc/modules.py
CHANGED
@@ -1,6 +1,5 @@
|
|
1 |
import traceback
|
2 |
import logging
|
3 |
-
import os
|
4 |
|
5 |
logger = logging.getLogger(__name__)
|
6 |
|
@@ -10,10 +9,14 @@ import torch
|
|
10 |
from io import BytesIO
|
11 |
|
12 |
from infer.lib.audio import load_audio, wav2
|
13 |
-
from
|
14 |
-
|
15 |
-
|
16 |
-
|
|
|
|
|
|
|
|
|
17 |
|
18 |
|
19 |
class VC:
|
@@ -59,45 +62,71 @@ class VC:
|
|
59 |
) = None
|
60 |
if torch.cuda.is_available():
|
61 |
torch.cuda.empty_cache()
|
62 |
-
elif torch.backends.mps.is_available():
|
63 |
-
torch.mps.empty_cache()
|
64 |
###楼下不这么折腾清理不干净
|
65 |
-
self.net_g, self.cpt = get_synthesizer(self.cpt, self.config.device)
|
66 |
self.if_f0 = self.cpt.get("f0", 1)
|
67 |
self.version = self.cpt.get("version", "v1")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
68 |
del self.net_g, self.cpt
|
69 |
if torch.cuda.is_available():
|
70 |
torch.cuda.empty_cache()
|
71 |
-
elif torch.backends.mps.is_available():
|
72 |
-
torch.mps.empty_cache()
|
73 |
return (
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
|
|
|
|
|
|
|
84 |
)
|
85 |
-
|
86 |
person = f'{os.getenv("weight_root")}/{sid}'
|
87 |
logger.info(f"Loading: {person}")
|
88 |
|
89 |
-
self.
|
90 |
self.tgt_sr = self.cpt["config"][-1]
|
91 |
self.cpt["config"][-3] = self.cpt["weight"]["emb_g.weight"].shape[0] # n_spk
|
92 |
self.if_f0 = self.cpt.get("f0", 1)
|
93 |
self.version = self.cpt.get("version", "v1")
|
94 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
95 |
if self.config.is_half:
|
96 |
self.net_g = self.net_g.half()
|
97 |
else:
|
98 |
self.net_g = self.net_g.float()
|
99 |
-
self.pipeline = Pipeline(self.tgt_sr, self.config)
|
100 |
|
|
|
101 |
n_spk = self.cpt["config"][-3]
|
102 |
index = {"value": get_index_path_from_model(sid), "__type__": "update"}
|
103 |
logger.info("Select index: " + index["value"])
|
@@ -109,7 +138,6 @@ class VC:
|
|
109 |
to_return_protect1,
|
110 |
index,
|
111 |
index,
|
112 |
-
show_model_info(self.cpt),
|
113 |
)
|
114 |
if to_return_protect
|
115 |
else {"visible": True, "maximum": n_spk, "__type__": "update"}
|
@@ -132,22 +160,18 @@ class VC:
|
|
132 |
):
|
133 |
if input_audio_path is None:
|
134 |
return "You need to upload an audio", None
|
135 |
-
elif hasattr(input_audio_path, "name"):
|
136 |
-
input_audio_path = str(input_audio_path.name)
|
137 |
f0_up_key = int(f0_up_key)
|
138 |
try:
|
139 |
audio = load_audio(input_audio_path, 16000)
|
140 |
audio_max = np.abs(audio).max() / 0.95
|
141 |
if audio_max > 1:
|
142 |
-
|
143 |
times = [0, 0, 0]
|
144 |
|
145 |
if self.hubert_model is None:
|
146 |
-
self.hubert_model = load_hubert(self.config
|
147 |
|
148 |
if file_index:
|
149 |
-
if hasattr(file_index, "name"):
|
150 |
-
file_index = str(file_index.name)
|
151 |
file_index = (
|
152 |
file_index.strip(" ")
|
153 |
.strip('"')
|
@@ -166,6 +190,7 @@ class VC:
|
|
166 |
self.net_g,
|
167 |
sid,
|
168 |
audio,
|
|
|
169 |
times,
|
170 |
f0_up_key,
|
171 |
f0_method,
|
@@ -179,25 +204,25 @@ class VC:
|
|
179 |
self.version,
|
180 |
protect,
|
181 |
f0_file,
|
182 |
-
)
|
183 |
if self.tgt_sr != resample_sr >= 16000:
|
184 |
tgt_sr = resample_sr
|
185 |
else:
|
186 |
tgt_sr = self.tgt_sr
|
187 |
index_info = (
|
188 |
-
"Index
|
189 |
if os.path.exists(file_index)
|
190 |
else "Index not used."
|
191 |
)
|
192 |
return (
|
193 |
-
"Success.\n%s\nTime:
|
194 |
% (index_info, *times),
|
195 |
(tgt_sr, audio_opt),
|
196 |
)
|
197 |
-
except
|
198 |
info = traceback.format_exc()
|
199 |
logger.warning(info)
|
200 |
-
return
|
201 |
|
202 |
def vc_multi(
|
203 |
self,
|
|
|
1 |
import traceback
|
2 |
import logging
|
|
|
3 |
|
4 |
logger = logging.getLogger(__name__)
|
5 |
|
|
|
9 |
from io import BytesIO
|
10 |
|
11 |
from infer.lib.audio import load_audio, wav2
|
12 |
+
from infer.lib.infer_pack.models import (
|
13 |
+
SynthesizerTrnMs256NSFsid,
|
14 |
+
SynthesizerTrnMs256NSFsid_nono,
|
15 |
+
SynthesizerTrnMs768NSFsid,
|
16 |
+
SynthesizerTrnMs768NSFsid_nono,
|
17 |
+
)
|
18 |
+
from infer.modules.vc.pipeline import Pipeline
|
19 |
+
from infer.modules.vc.utils import *
|
20 |
|
21 |
|
22 |
class VC:
|
|
|
62 |
) = None
|
63 |
if torch.cuda.is_available():
|
64 |
torch.cuda.empty_cache()
|
|
|
|
|
65 |
###楼下不这么折腾清理不干净
|
|
|
66 |
self.if_f0 = self.cpt.get("f0", 1)
|
67 |
self.version = self.cpt.get("version", "v1")
|
68 |
+
if self.version == "v1":
|
69 |
+
if self.if_f0 == 1:
|
70 |
+
self.net_g = SynthesizerTrnMs256NSFsid(
|
71 |
+
*self.cpt["config"], is_half=self.config.is_half
|
72 |
+
)
|
73 |
+
else:
|
74 |
+
self.net_g = SynthesizerTrnMs256NSFsid_nono(*self.cpt["config"])
|
75 |
+
elif self.version == "v2":
|
76 |
+
if self.if_f0 == 1:
|
77 |
+
self.net_g = SynthesizerTrnMs768NSFsid(
|
78 |
+
*self.cpt["config"], is_half=self.config.is_half
|
79 |
+
)
|
80 |
+
else:
|
81 |
+
self.net_g = SynthesizerTrnMs768NSFsid_nono(*self.cpt["config"])
|
82 |
del self.net_g, self.cpt
|
83 |
if torch.cuda.is_available():
|
84 |
torch.cuda.empty_cache()
|
|
|
|
|
85 |
return (
|
86 |
+
{"visible": False, "__type__": "update"},
|
87 |
+
{
|
88 |
+
"visible": True,
|
89 |
+
"value": to_return_protect0,
|
90 |
+
"__type__": "update",
|
91 |
+
},
|
92 |
+
{
|
93 |
+
"visible": True,
|
94 |
+
"value": to_return_protect1,
|
95 |
+
"__type__": "update",
|
96 |
+
},
|
97 |
+
"",
|
98 |
+
"",
|
99 |
)
|
|
|
100 |
person = f'{os.getenv("weight_root")}/{sid}'
|
101 |
logger.info(f"Loading: {person}")
|
102 |
|
103 |
+
self.cpt = torch.load(person, map_location="cpu")
|
104 |
self.tgt_sr = self.cpt["config"][-1]
|
105 |
self.cpt["config"][-3] = self.cpt["weight"]["emb_g.weight"].shape[0] # n_spk
|
106 |
self.if_f0 = self.cpt.get("f0", 1)
|
107 |
self.version = self.cpt.get("version", "v1")
|
108 |
|
109 |
+
synthesizer_class = {
|
110 |
+
("v1", 1): SynthesizerTrnMs256NSFsid,
|
111 |
+
("v1", 0): SynthesizerTrnMs256NSFsid_nono,
|
112 |
+
("v2", 1): SynthesizerTrnMs768NSFsid,
|
113 |
+
("v2", 0): SynthesizerTrnMs768NSFsid_nono,
|
114 |
+
}
|
115 |
+
|
116 |
+
self.net_g = synthesizer_class.get(
|
117 |
+
(self.version, self.if_f0), SynthesizerTrnMs256NSFsid
|
118 |
+
)(*self.cpt["config"], is_half=self.config.is_half)
|
119 |
+
|
120 |
+
del self.net_g.enc_q
|
121 |
+
|
122 |
+
self.net_g.load_state_dict(self.cpt["weight"], strict=False)
|
123 |
+
self.net_g.eval().to(self.config.device)
|
124 |
if self.config.is_half:
|
125 |
self.net_g = self.net_g.half()
|
126 |
else:
|
127 |
self.net_g = self.net_g.float()
|
|
|
128 |
|
129 |
+
self.pipeline = Pipeline(self.tgt_sr, self.config)
|
130 |
n_spk = self.cpt["config"][-3]
|
131 |
index = {"value": get_index_path_from_model(sid), "__type__": "update"}
|
132 |
logger.info("Select index: " + index["value"])
|
|
|
138 |
to_return_protect1,
|
139 |
index,
|
140 |
index,
|
|
|
141 |
)
|
142 |
if to_return_protect
|
143 |
else {"visible": True, "maximum": n_spk, "__type__": "update"}
|
|
|
160 |
):
|
161 |
if input_audio_path is None:
|
162 |
return "You need to upload an audio", None
|
|
|
|
|
163 |
f0_up_key = int(f0_up_key)
|
164 |
try:
|
165 |
audio = load_audio(input_audio_path, 16000)
|
166 |
audio_max = np.abs(audio).max() / 0.95
|
167 |
if audio_max > 1:
|
168 |
+
audio /= audio_max
|
169 |
times = [0, 0, 0]
|
170 |
|
171 |
if self.hubert_model is None:
|
172 |
+
self.hubert_model = load_hubert(self.config)
|
173 |
|
174 |
if file_index:
|
|
|
|
|
175 |
file_index = (
|
176 |
file_index.strip(" ")
|
177 |
.strip('"')
|
|
|
190 |
self.net_g,
|
191 |
sid,
|
192 |
audio,
|
193 |
+
input_audio_path,
|
194 |
times,
|
195 |
f0_up_key,
|
196 |
f0_method,
|
|
|
204 |
self.version,
|
205 |
protect,
|
206 |
f0_file,
|
207 |
+
)
|
208 |
if self.tgt_sr != resample_sr >= 16000:
|
209 |
tgt_sr = resample_sr
|
210 |
else:
|
211 |
tgt_sr = self.tgt_sr
|
212 |
index_info = (
|
213 |
+
"Index:\n%s." % file_index
|
214 |
if os.path.exists(file_index)
|
215 |
else "Index not used."
|
216 |
)
|
217 |
return (
|
218 |
+
"Success.\n%s\nTime:\nnpy: %.2fs, f0: %.2fs, infer: %.2fs."
|
219 |
% (index_info, *times),
|
220 |
(tgt_sr, audio_opt),
|
221 |
)
|
222 |
+
except:
|
223 |
info = traceback.format_exc()
|
224 |
logger.warning(info)
|
225 |
+
return info, (None, None)
|
226 |
|
227 |
def vc_multi(
|
228 |
self,
|
infer/modules/vc/pipeline.py
CHANGED
@@ -5,22 +5,40 @@ import logging
|
|
5 |
|
6 |
logger = logging.getLogger(__name__)
|
7 |
|
8 |
-
from
|
|
|
9 |
|
10 |
import faiss
|
11 |
import librosa
|
12 |
import numpy as np
|
|
|
|
|
13 |
import torch
|
14 |
import torch.nn.functional as F
|
|
|
15 |
from scipy import signal
|
16 |
|
17 |
-
from rvc.f0 import PM, Harvest, RMVPE, CRePE, Dio, FCPE
|
18 |
-
|
19 |
now_dir = os.getcwd()
|
20 |
sys.path.append(now_dir)
|
21 |
|
22 |
bh, ah = signal.butter(N=5, Wn=48, btype="high", fs=16000)
|
23 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
|
25 |
def change_rms(data1, sr1, data2, sr2, rate): # 1是输入音频,2是输出音频,rate是2的占比
|
26 |
# print(data1.max(),data2.max())
|
@@ -65,6 +83,7 @@ class Pipeline(object):
|
|
65 |
|
66 |
def get_f0(
|
67 |
self,
|
|
|
68 |
x,
|
69 |
p_len,
|
70 |
f0_up_key,
|
@@ -72,62 +91,73 @@ class Pipeline(object):
|
|
72 |
filter_radius,
|
73 |
inp_f0=None,
|
74 |
):
|
|
|
|
|
75 |
f0_min = 50
|
76 |
f0_max = 1100
|
77 |
f0_mel_min = 1127 * np.log(1 + f0_min / 700)
|
78 |
f0_mel_max = 1127 * np.log(1 + f0_max / 700)
|
79 |
if f0_method == "pm":
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
87 |
elif f0_method == "harvest":
|
88 |
-
|
89 |
-
|
90 |
-
|
|
|
91 |
elif f0_method == "crepe":
|
92 |
-
|
93 |
-
|
94 |
-
|
95 |
-
|
96 |
-
|
97 |
-
|
98 |
-
|
99 |
-
|
100 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
101 |
elif f0_method == "rmvpe":
|
102 |
-
if not hasattr(self, "
|
|
|
|
|
103 |
logger.info(
|
104 |
-
"Loading rmvpe model
|
105 |
)
|
106 |
-
self.
|
107 |
"%s/rmvpe.pt" % os.environ["rmvpe_root"],
|
108 |
is_half=self.is_half,
|
109 |
device=self.device,
|
110 |
-
# use_jit=self.config.use_jit,
|
111 |
)
|
112 |
-
f0 = self.
|
113 |
|
114 |
if "privateuseone" in str(self.device): # clean ortruntime memory
|
115 |
-
del self.
|
116 |
-
del self.
|
117 |
logger.info("Cleaning ortruntime memory")
|
118 |
|
119 |
-
elif f0_method == "fcpe":
|
120 |
-
if not hasattr(self, "model_fcpe"):
|
121 |
-
logger.info("Loading fcpe model")
|
122 |
-
self.model_fcpe = FCPE(
|
123 |
-
self.window,
|
124 |
-
f0_min,
|
125 |
-
f0_max,
|
126 |
-
self.sr,
|
127 |
-
self.device,
|
128 |
-
)
|
129 |
-
f0 = self.model_fcpe.compute_f0(x, p_len=p_len)
|
130 |
-
|
131 |
f0 *= pow(2, f0_up_key / 12)
|
132 |
# with open("test.txt","w")as f:f.write("\n".join([str(i)for i in f0.tolist()]))
|
133 |
tf0 = self.sr // self.window # 每秒f0点数
|
@@ -184,7 +214,7 @@ class Pipeline(object):
|
|
184 |
"padding_mask": padding_mask,
|
185 |
"output_layer": 9 if version == "v1" else 12,
|
186 |
}
|
187 |
-
t0 =
|
188 |
with torch.no_grad():
|
189 |
logits = model.extract_features(**inputs)
|
190 |
feats = model.final_proj(logits[0]) if version == "v1" else logits[0]
|
@@ -202,10 +232,7 @@ class Pipeline(object):
|
|
202 |
# _, I = index.search(npy, 1)
|
203 |
# npy = big_npy[I.squeeze()]
|
204 |
|
205 |
-
|
206 |
-
score, ix = index.search(npy, k=8)
|
207 |
-
except:
|
208 |
-
raise Exception("index mistatch")
|
209 |
weight = np.square(1 / score)
|
210 |
weight /= weight.sum(axis=1, keepdims=True)
|
211 |
npy = np.sum(big_npy[ix] * np.expand_dims(weight, axis=2), axis=1)
|
@@ -222,7 +249,7 @@ class Pipeline(object):
|
|
222 |
feats0 = F.interpolate(feats0.permute(0, 2, 1), scale_factor=2).permute(
|
223 |
0, 2, 1
|
224 |
)
|
225 |
-
t1 =
|
226 |
p_len = audio0.shape[0] // self.window
|
227 |
if feats.shape[1] < p_len:
|
228 |
p_len = feats.shape[1]
|
@@ -239,26 +266,14 @@ class Pipeline(object):
|
|
239 |
feats = feats.to(feats0.dtype)
|
240 |
p_len = torch.tensor([p_len], device=self.device).long()
|
241 |
with torch.no_grad():
|
242 |
-
|
243 |
-
|
244 |
-
|
245 |
-
|
246 |
-
p_len,
|
247 |
-
sid,
|
248 |
-
pitch=pitch,
|
249 |
-
pitchf=pitchf,
|
250 |
-
)[0, 0]
|
251 |
-
)
|
252 |
-
.data.cpu()
|
253 |
-
.float()
|
254 |
-
.numpy()
|
255 |
-
)
|
256 |
del feats, p_len, padding_mask
|
257 |
if torch.cuda.is_available():
|
258 |
torch.cuda.empty_cache()
|
259 |
-
|
260 |
-
torch.mps.empty_cache()
|
261 |
-
t2 = time()
|
262 |
times[0] += t1 - t0
|
263 |
times[2] += t2 - t1
|
264 |
return audio1
|
@@ -269,6 +284,7 @@ class Pipeline(object):
|
|
269 |
net_g,
|
270 |
sid,
|
271 |
audio,
|
|
|
272 |
times,
|
273 |
f0_up_key,
|
274 |
f0_method,
|
@@ -292,6 +308,7 @@ class Pipeline(object):
|
|
292 |
):
|
293 |
try:
|
294 |
index = faiss.read_index(file_index)
|
|
|
295 |
big_npy = index.reconstruct_n(0, index.ntotal)
|
296 |
except:
|
297 |
traceback.print_exc()
|
@@ -317,7 +334,7 @@ class Pipeline(object):
|
|
317 |
s = 0
|
318 |
audio_opt = []
|
319 |
t = None
|
320 |
-
t1 =
|
321 |
audio_pad = np.pad(audio, (self.t_pad, self.t_pad), mode="reflect")
|
322 |
p_len = audio_pad.shape[0] // self.window
|
323 |
inp_f0 = None
|
@@ -333,29 +350,27 @@ class Pipeline(object):
|
|
333 |
traceback.print_exc()
|
334 |
sid = torch.tensor(sid, device=self.device).unsqueeze(0).long()
|
335 |
pitch, pitchf = None, None
|
336 |
-
if if_f0:
|
337 |
-
|
338 |
-
|
339 |
-
|
340 |
-
|
341 |
-
|
342 |
-
|
343 |
-
|
344 |
-
|
345 |
-
|
346 |
-
elif if_f0 == 2:
|
347 |
-
pitch, pitchf = f0_method
|
348 |
pitch = pitch[:p_len]
|
349 |
pitchf = pitchf[:p_len]
|
350 |
if "mps" not in str(self.device) or "xpu" not in str(self.device):
|
351 |
pitchf = pitchf.astype(np.float32)
|
352 |
pitch = torch.tensor(pitch, device=self.device).unsqueeze(0).long()
|
353 |
pitchf = torch.tensor(pitchf, device=self.device).unsqueeze(0).float()
|
354 |
-
t2 =
|
355 |
times[1] += t2 - t1
|
356 |
for t in opt_ts:
|
357 |
t = t // self.window * self.window
|
358 |
-
if if_f0:
|
359 |
audio_opt.append(
|
360 |
self.vc(
|
361 |
model,
|
@@ -390,7 +405,7 @@ class Pipeline(object):
|
|
390 |
)[self.t_pad_tgt : -self.t_pad_tgt]
|
391 |
)
|
392 |
s = t
|
393 |
-
if if_f0:
|
394 |
audio_opt.append(
|
395 |
self.vc(
|
396 |
model,
|
@@ -435,10 +450,8 @@ class Pipeline(object):
|
|
435 |
max_int16 = 32768
|
436 |
if audio_max > 1:
|
437 |
max_int16 /= audio_max
|
438 |
-
|
439 |
del pitch, pitchf, sid
|
440 |
if torch.cuda.is_available():
|
441 |
torch.cuda.empty_cache()
|
442 |
-
elif torch.backends.mps.is_available():
|
443 |
-
torch.mps.empty_cache()
|
444 |
return audio_opt
|
|
|
5 |
|
6 |
logger = logging.getLogger(__name__)
|
7 |
|
8 |
+
from functools import lru_cache
|
9 |
+
from time import time as ttime
|
10 |
|
11 |
import faiss
|
12 |
import librosa
|
13 |
import numpy as np
|
14 |
+
import parselmouth
|
15 |
+
import pyworld
|
16 |
import torch
|
17 |
import torch.nn.functional as F
|
18 |
+
import torchcrepe
|
19 |
from scipy import signal
|
20 |
|
|
|
|
|
21 |
now_dir = os.getcwd()
|
22 |
sys.path.append(now_dir)
|
23 |
|
24 |
bh, ah = signal.butter(N=5, Wn=48, btype="high", fs=16000)
|
25 |
|
26 |
+
input_audio_path2wav = {}
|
27 |
+
|
28 |
+
|
29 |
+
@lru_cache
|
30 |
+
def cache_harvest_f0(input_audio_path, fs, f0max, f0min, frame_period):
|
31 |
+
audio = input_audio_path2wav[input_audio_path]
|
32 |
+
f0, t = pyworld.harvest(
|
33 |
+
audio,
|
34 |
+
fs=fs,
|
35 |
+
f0_ceil=f0max,
|
36 |
+
f0_floor=f0min,
|
37 |
+
frame_period=frame_period,
|
38 |
+
)
|
39 |
+
f0 = pyworld.stonemask(audio, f0, t, fs)
|
40 |
+
return f0
|
41 |
+
|
42 |
|
43 |
def change_rms(data1, sr1, data2, sr2, rate): # 1是输入音频,2是输出音频,rate是2的占比
|
44 |
# print(data1.max(),data2.max())
|
|
|
83 |
|
84 |
def get_f0(
|
85 |
self,
|
86 |
+
input_audio_path,
|
87 |
x,
|
88 |
p_len,
|
89 |
f0_up_key,
|
|
|
91 |
filter_radius,
|
92 |
inp_f0=None,
|
93 |
):
|
94 |
+
global input_audio_path2wav
|
95 |
+
time_step = self.window / self.sr * 1000
|
96 |
f0_min = 50
|
97 |
f0_max = 1100
|
98 |
f0_mel_min = 1127 * np.log(1 + f0_min / 700)
|
99 |
f0_mel_max = 1127 * np.log(1 + f0_max / 700)
|
100 |
if f0_method == "pm":
|
101 |
+
f0 = (
|
102 |
+
parselmouth.Sound(x, self.sr)
|
103 |
+
.to_pitch_ac(
|
104 |
+
time_step=time_step / 1000,
|
105 |
+
voicing_threshold=0.6,
|
106 |
+
pitch_floor=f0_min,
|
107 |
+
pitch_ceiling=f0_max,
|
108 |
+
)
|
109 |
+
.selected_array["frequency"]
|
110 |
+
)
|
111 |
+
pad_size = (p_len - len(f0) + 1) // 2
|
112 |
+
if pad_size > 0 or p_len - len(f0) - pad_size > 0:
|
113 |
+
f0 = np.pad(
|
114 |
+
f0, [[pad_size, p_len - len(f0) - pad_size]], mode="constant"
|
115 |
+
)
|
116 |
elif f0_method == "harvest":
|
117 |
+
input_audio_path2wav[input_audio_path] = x.astype(np.double)
|
118 |
+
f0 = cache_harvest_f0(input_audio_path, self.sr, f0_max, f0_min, 10)
|
119 |
+
if filter_radius > 2:
|
120 |
+
f0 = signal.medfilt(f0, 3)
|
121 |
elif f0_method == "crepe":
|
122 |
+
model = "full"
|
123 |
+
# Pick a batch size that doesn't cause memory errors on your gpu
|
124 |
+
batch_size = 512
|
125 |
+
# Compute pitch using first gpu
|
126 |
+
audio = torch.tensor(np.copy(x))[None].float()
|
127 |
+
f0, pd = torchcrepe.predict(
|
128 |
+
audio,
|
129 |
+
self.sr,
|
130 |
+
self.window,
|
131 |
+
f0_min,
|
132 |
+
f0_max,
|
133 |
+
model,
|
134 |
+
batch_size=batch_size,
|
135 |
+
device=self.device,
|
136 |
+
return_periodicity=True,
|
137 |
+
)
|
138 |
+
pd = torchcrepe.filter.median(pd, 3)
|
139 |
+
f0 = torchcrepe.filter.mean(f0, 3)
|
140 |
+
f0[pd < 0.1] = 0
|
141 |
+
f0 = f0[0].cpu().numpy()
|
142 |
elif f0_method == "rmvpe":
|
143 |
+
if not hasattr(self, "model_rmvpe"):
|
144 |
+
from infer.lib.rmvpe import RMVPE
|
145 |
+
|
146 |
logger.info(
|
147 |
+
"Loading rmvpe model,%s" % "%s/rmvpe.pt" % os.environ["rmvpe_root"]
|
148 |
)
|
149 |
+
self.model_rmvpe = RMVPE(
|
150 |
"%s/rmvpe.pt" % os.environ["rmvpe_root"],
|
151 |
is_half=self.is_half,
|
152 |
device=self.device,
|
|
|
153 |
)
|
154 |
+
f0 = self.model_rmvpe.infer_from_audio(x, thred=0.03)
|
155 |
|
156 |
if "privateuseone" in str(self.device): # clean ortruntime memory
|
157 |
+
del self.model_rmvpe.model
|
158 |
+
del self.model_rmvpe
|
159 |
logger.info("Cleaning ortruntime memory")
|
160 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
161 |
f0 *= pow(2, f0_up_key / 12)
|
162 |
# with open("test.txt","w")as f:f.write("\n".join([str(i)for i in f0.tolist()]))
|
163 |
tf0 = self.sr // self.window # 每秒f0点数
|
|
|
214 |
"padding_mask": padding_mask,
|
215 |
"output_layer": 9 if version == "v1" else 12,
|
216 |
}
|
217 |
+
t0 = ttime()
|
218 |
with torch.no_grad():
|
219 |
logits = model.extract_features(**inputs)
|
220 |
feats = model.final_proj(logits[0]) if version == "v1" else logits[0]
|
|
|
232 |
# _, I = index.search(npy, 1)
|
233 |
# npy = big_npy[I.squeeze()]
|
234 |
|
235 |
+
score, ix = index.search(npy, k=8)
|
|
|
|
|
|
|
236 |
weight = np.square(1 / score)
|
237 |
weight /= weight.sum(axis=1, keepdims=True)
|
238 |
npy = np.sum(big_npy[ix] * np.expand_dims(weight, axis=2), axis=1)
|
|
|
249 |
feats0 = F.interpolate(feats0.permute(0, 2, 1), scale_factor=2).permute(
|
250 |
0, 2, 1
|
251 |
)
|
252 |
+
t1 = ttime()
|
253 |
p_len = audio0.shape[0] // self.window
|
254 |
if feats.shape[1] < p_len:
|
255 |
p_len = feats.shape[1]
|
|
|
266 |
feats = feats.to(feats0.dtype)
|
267 |
p_len = torch.tensor([p_len], device=self.device).long()
|
268 |
with torch.no_grad():
|
269 |
+
hasp = pitch is not None and pitchf is not None
|
270 |
+
arg = (feats, p_len, pitch, pitchf, sid) if hasp else (feats, p_len, sid)
|
271 |
+
audio1 = (net_g.infer(*arg)[0][0, 0]).data.cpu().float().numpy()
|
272 |
+
del hasp, arg
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
273 |
del feats, p_len, padding_mask
|
274 |
if torch.cuda.is_available():
|
275 |
torch.cuda.empty_cache()
|
276 |
+
t2 = ttime()
|
|
|
|
|
277 |
times[0] += t1 - t0
|
278 |
times[2] += t2 - t1
|
279 |
return audio1
|
|
|
284 |
net_g,
|
285 |
sid,
|
286 |
audio,
|
287 |
+
input_audio_path,
|
288 |
times,
|
289 |
f0_up_key,
|
290 |
f0_method,
|
|
|
308 |
):
|
309 |
try:
|
310 |
index = faiss.read_index(file_index)
|
311 |
+
# big_npy = np.load(file_big_npy)
|
312 |
big_npy = index.reconstruct_n(0, index.ntotal)
|
313 |
except:
|
314 |
traceback.print_exc()
|
|
|
334 |
s = 0
|
335 |
audio_opt = []
|
336 |
t = None
|
337 |
+
t1 = ttime()
|
338 |
audio_pad = np.pad(audio, (self.t_pad, self.t_pad), mode="reflect")
|
339 |
p_len = audio_pad.shape[0] // self.window
|
340 |
inp_f0 = None
|
|
|
350 |
traceback.print_exc()
|
351 |
sid = torch.tensor(sid, device=self.device).unsqueeze(0).long()
|
352 |
pitch, pitchf = None, None
|
353 |
+
if if_f0 == 1:
|
354 |
+
pitch, pitchf = self.get_f0(
|
355 |
+
input_audio_path,
|
356 |
+
audio_pad,
|
357 |
+
p_len,
|
358 |
+
f0_up_key,
|
359 |
+
f0_method,
|
360 |
+
filter_radius,
|
361 |
+
inp_f0,
|
362 |
+
)
|
|
|
|
|
363 |
pitch = pitch[:p_len]
|
364 |
pitchf = pitchf[:p_len]
|
365 |
if "mps" not in str(self.device) or "xpu" not in str(self.device):
|
366 |
pitchf = pitchf.astype(np.float32)
|
367 |
pitch = torch.tensor(pitch, device=self.device).unsqueeze(0).long()
|
368 |
pitchf = torch.tensor(pitchf, device=self.device).unsqueeze(0).float()
|
369 |
+
t2 = ttime()
|
370 |
times[1] += t2 - t1
|
371 |
for t in opt_ts:
|
372 |
t = t // self.window * self.window
|
373 |
+
if if_f0 == 1:
|
374 |
audio_opt.append(
|
375 |
self.vc(
|
376 |
model,
|
|
|
405 |
)[self.t_pad_tgt : -self.t_pad_tgt]
|
406 |
)
|
407 |
s = t
|
408 |
+
if if_f0 == 1:
|
409 |
audio_opt.append(
|
410 |
self.vc(
|
411 |
model,
|
|
|
450 |
max_int16 = 32768
|
451 |
if audio_max > 1:
|
452 |
max_int16 /= audio_max
|
453 |
+
audio_opt = (audio_opt * max_int16).astype(np.int16)
|
454 |
del pitch, pitchf, sid
|
455 |
if torch.cuda.is_available():
|
456 |
torch.cuda.empty_cache()
|
|
|
|
|
457 |
return audio_opt
|
infer/modules/vc/utils.py
CHANGED
@@ -9,8 +9,7 @@ def get_index_path_from_model(sid):
|
|
9 |
f
|
10 |
for f in [
|
11 |
os.path.join(root, name)
|
12 |
-
for
|
13 |
-
for root, _, files in os.walk(path, topdown=False)
|
14 |
for name in files
|
15 |
if name.endswith(".index") and "trained" not in name
|
16 |
]
|
@@ -20,14 +19,14 @@ def get_index_path_from_model(sid):
|
|
20 |
)
|
21 |
|
22 |
|
23 |
-
def load_hubert(
|
24 |
models, _, _ = checkpoint_utils.load_model_ensemble_and_task(
|
25 |
["assets/hubert/hubert_base.pt"],
|
26 |
suffix="",
|
27 |
)
|
28 |
hubert_model = models[0]
|
29 |
-
hubert_model = hubert_model.to(device)
|
30 |
-
if is_half:
|
31 |
hubert_model = hubert_model.half()
|
32 |
else:
|
33 |
hubert_model = hubert_model.float()
|
|
|
9 |
f
|
10 |
for f in [
|
11 |
os.path.join(root, name)
|
12 |
+
for root, _, files in os.walk(os.getenv("index_root"), topdown=False)
|
|
|
13 |
for name in files
|
14 |
if name.endswith(".index") and "trained" not in name
|
15 |
]
|
|
|
19 |
)
|
20 |
|
21 |
|
22 |
+
def load_hubert(config):
|
23 |
models, _, _ = checkpoint_utils.load_model_ensemble_and_task(
|
24 |
["assets/hubert/hubert_base.pt"],
|
25 |
suffix="",
|
26 |
)
|
27 |
hubert_model = models[0]
|
28 |
+
hubert_model = hubert_model.to(config.device)
|
29 |
+
if config.is_half:
|
30 |
hubert_model = hubert_model.half()
|
31 |
else:
|
32 |
hubert_model = hubert_model.float()
|
requirements.txt
CHANGED
@@ -1,15 +1,16 @@
|
|
1 |
joblib>=1.1.0
|
2 |
-
numba
|
3 |
numpy==1.23.5
|
4 |
scipy
|
5 |
librosa==0.9.1
|
6 |
-
llvmlite
|
7 |
-
fairseq
|
8 |
-
faiss-cpu
|
9 |
-
gradio
|
10 |
Cython
|
11 |
pydub>=0.25.1
|
12 |
soundfile>=0.12.1
|
|
|
13 |
tensorboardX
|
14 |
Jinja2>=3.1.2
|
15 |
json5
|
@@ -40,8 +41,8 @@ httpx
|
|
40 |
onnxruntime; sys_platform == 'darwin'
|
41 |
onnxruntime-gpu; sys_platform != 'darwin'
|
42 |
torchcrepe==0.0.20
|
43 |
-
fastapi
|
44 |
torchfcpe
|
|
|
45 |
python-dotenv>=1.0.0
|
46 |
av
|
47 |
-
pybase16384
|
|
|
1 |
joblib>=1.1.0
|
2 |
+
numba==0.56.4
|
3 |
numpy==1.23.5
|
4 |
scipy
|
5 |
librosa==0.9.1
|
6 |
+
llvmlite==0.39.0
|
7 |
+
fairseq==0.12.2
|
8 |
+
faiss-cpu==1.7.3
|
9 |
+
gradio==3.34.0
|
10 |
Cython
|
11 |
pydub>=0.25.1
|
12 |
soundfile>=0.12.1
|
13 |
+
ffmpeg-python>=0.2.0
|
14 |
tensorboardX
|
15 |
Jinja2>=3.1.2
|
16 |
json5
|
|
|
41 |
onnxruntime; sys_platform == 'darwin'
|
42 |
onnxruntime-gpu; sys_platform != 'darwin'
|
43 |
torchcrepe==0.0.20
|
44 |
+
fastapi==0.88
|
45 |
torchfcpe
|
46 |
+
ffmpy==0.3.1
|
47 |
python-dotenv>=1.0.0
|
48 |
av
|
|