nodes/billwuhao-ComfyUI_DiffRhythm
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Support Linux

Browse Source
This commit is contained in:
billwuhao
2025-03-15 23:15:09 +08:00
parent ea92c64dcf
commit 776a61a6ab
7 changed files with 114 additions and 80 deletions
Download Patch File Download Diff File Expand all files Collapse all files
DiffRhythmNode.py
+5
View File
@@ -19,6 +19,11 @@ import librosa
from mutagen.mp3 import MP3
import torch
from einops import rearrange
import sys
import os
current_dir = os.path.dirname(os.path.abspath(__file__))
sys.path.insert(0, current_dir)
from diffrhythm_utils import (
decode_audio,
README-CN.md
+64
View File
@@ -0,0 +1,64 @@
[中文](README.md) | [English](README-en.md)
# DiffRhythm 的 ComfyUI 节点
快速而简单的端到端全长歌曲生成.
![](https://github.com/billwuhao/ComfyUI_DiffRhythm/blob/master/images/2025-03-12_23-49-32.png)
## 📣 更新
[2025-03-13]⚒️: 发布版本 v1.0.0.
- 所有参数均是可选的, 不提供任何参数随机生成音乐.
## 安装
```
cd ComfyUI/custom_nodes
git clone https://github.com/billwuhao/ComfyUI_DiffRhythm.git
cd ComfyUI_DiffRhythm
pip install -r requirements.txt
# python_embeded
./python_embeded/python.exe -m pip install -r requirements.txt
```
## 模型下载
模型会自动下载到 `ComfyUI\models\TTS\DiffRhythm` 文件夹下.
结构如下:
![](https://github.com/billwuhao/ComfyUI_DiffRhythm/blob/master/images/2025-03-13_00-08-51.png)
手动下载地址:
https://huggingface.co/ASLP-lab/DiffRhythm-base/blob/main/cfm_model.pt
https://huggingface.co/ASLP-lab/DiffRhythm-vae/blob/main/vae_model.pt
https://huggingface.co/OpenMuQ/MuQ-MuLan-large/tree/main
https://huggingface.co/OpenMuQ/MuQ-large-msd-iter/tree/main
https://huggingface.co/FacebookAI/xlm-roberta-base/tree/main
## 环境配置
Windows 系统做如下配置.
下载安装最新版 [espeak-ng](https://github.com/espeak-ng/espeak-ng/releases/tag/1.52.0)
添加环境变量 `PHONEMIZER_ESPEAK_LIBRARY` 到系统中, 值是你安装的 espeak-ng 软件中 `libespeak-ng.dll` 文件的路径, 例如: `C:\Program Files\eSpeak NG\libespeak-ng.dll`.
Linux 系统下, 需要安装 `espeak-ng` 软件包. 执行如下命令安装:
`apt-get -qq -y install espeak-ng > /dev/null 2>&1`
应该支持 Mac, 但尚未测试.
享受音乐吧🎶
## 鸣谢
[DiffRhythm](https://github.com/ASLP-lab/DiffRhythm)
感谢 DiffRhythm 团队的卓越的工作, 目前最强开源 音乐/歌曲 生成模型👍.
README-en.md
-45
View File
@@ -1,45 +0,0 @@
[中文](README.md) | [English](README-en.md)
# DiffRhythm Node for ComfyUI
Blazingly Fast and Embarrassingly Simple End-to-End Full-Length Song Generation.
![](https://github.com/billwuhao/ComfyUI_DiffRhythm/blob/master/images/2025-03-12_23-49-32.png)
## 📣 update
[2025-03-13]⚒️: Release version v1.0.0.
- All parameters are optional; you can generate random music without providing any parameters.
## Model Download
Models will be automatically downloaded to the `ComfyUI\models\TTS\DiffRhythm` folder.
The structure is as follows:
![](https://github.com/billwuhao/ComfyUI_DiffRhythm/blob/master/images/2025-03-13_00-08-51.png)
Manual Download Addresses:
https://huggingface.co/ASLP-lab/DiffRhythm-base/blob/main/cfm_model.pt
https://huggingface.co/ASLP-lab/DiffRhythm-vae/blob/main/vae_model.pt
https://huggingface.co/OpenMuQ/MuQ-MuLan-large/tree/main
https://huggingface.co/OpenMuQ/MuQ-large-msd-iter/tree/main
https://huggingface.co/FacebookAI/xlm-roberta-base/tree/main
## Environment Configuration
Configure the following on Windows systems, other systems have not been tested. Should support Linux and Mac.
Download and install the latest version of [espeak-ng](https://github.com/espeak-ng/espeak-ng/releases/tag/1.52.0)
Add the environment variable `PHONEMIZER_ESPEAK_LIBRARY` to your system. The value should be the path to the `libespeak-ng.dll` file in your espeak-ng installation, for example: `C:\Program Files\eSpeak NG\libespeak-ng.dll`.
Enjoy the music! 🎶
## Acknowledgements
[DiffRhythm](https://github.com/ASLP-lab/DiffRhythm)
Thanks to the DiffRhythm team for their excellent work. Currently the strongest open-source music/song generation model 👍.
README.md
+33 -16
View File
@@ -1,27 +1,38 @@
[中文](README.md) | [English](README-en.md)
# DiffRhythm 的 ComfyUI 节点
# DiffRhythm Node for ComfyUI
快速而简单的端到端全长歌曲生成.
Blazingly Fast and Embarrassingly Simple End-to-End Full-Length Song Generation.
![](https://github.com/billwuhao/ComfyUI_DiffRhythm/blob/master/images/2025-03-12_23-49-32.png)
## 📣 update
## 📣 更新
[2025-03-13]⚒️: Release version v1.0.0.
[2025-03-13]⚒️: 发布版本 v1.0.0.
- All parameters are optional; you can generate random music without providing any parameters.
- 所有参数均是可选的, 不提供任何参数随机生成音乐.
## Installation
## 模型下载
```
cd ComfyUI/custom_nodes
git clone https://github.com/billwuhao/ComfyUI_DiffRhythm.git
cd ComfyUI_DiffRhythm
pip install -r requirements.txt
模型会自动下载到 `ComfyUI\models\TTS\DiffRhythm` 文件夹下.
# python_embeded
./python_embeded/python.exe -m pip install -r requirements.txt
```
结构如下:
## Model Download
Models will be automatically downloaded to the `ComfyUI\models\TTS\DiffRhythm` folder.
The structure is as follows:
![](https://github.com/billwuhao/ComfyUI_DiffRhythm/blob/master/images/2025-03-13_00-08-51.png)
手动下载地址:
Manual Download Addresses:
https://huggingface.co/ASLP-lab/DiffRhythm-base/blob/main/cfm_model.pt
https://huggingface.co/ASLP-lab/DiffRhythm-vae/blob/main/vae_model.pt
@@ -29,18 +40,24 @@ https://huggingface.co/OpenMuQ/MuQ-MuLan-large/tree/main
https://huggingface.co/OpenMuQ/MuQ-large-msd-iter/tree/main
https://huggingface.co/FacebookAI/xlm-roberta-base/tree/main
## 环境配置
## Environment Configuration
Windows 系统做如下配置, 其他系统未测试. 应该支持 Linux, Mac.
- Configure the following on Windows systems:
下载安装最新版 [espeak-ng](https://github.com/espeak-ng/espeak-ng/releases/tag/1.52.0)
Download and install the latest version of [espeak-ng](https://github.com/espeak-ng/espeak-ng/releases/tag/1.52.0)
添加环境变量 `PHONEMIZER_ESPEAK_LIBRARY` 到系统中, 值是你安装的 espeak-ng 软件中 `libespeak-ng.dll` 文件的路径, 例如: `C:\Program Files\eSpeak NG\libespeak-ng.dll`.
Add the environment variable `PHONEMIZER_ESPEAK_LIBRARY` to your system. The value should be the path to the `libespeak-ng.dll` file in your espeak-ng installation, for example: `C:\Program Files\eSpeak NG\libespeak-ng.dll`.
享受音乐吧🎶
- On Linux systems, you need to install the `espeak-ng` package. Execute the following command to install:
## 鸣谢
`apt-get -qq -y install espeak-ng > /dev/null 2>&1`
It should support Mac, but has not been tested.
Enjoy the music! 🎶
## Acknowledgements
[DiffRhythm](https://github.com/ASLP-lab/DiffRhythm)
感谢 DiffRhythm 团队的卓越的工作, 目前最强开源 音乐/歌曲 生成模型👍.
Thanks to the DiffRhythm team for their excellent work. Currently the strongest open-source music/song generation model 👍.
__init__.py
+1 -6
View File
@@ -1,8 +1,3 @@
import sys
import os
from .DiffRhythmNode import NODE_CLASS_MAPPINGS, NODE_DISPLAY_NAME_MAPPINGS
current_dir = os.path.dirname(os.path.abspath(__file__))
sys.path.insert(0, current_dir)
from DiffRhythmNode import NODE_CLASS_MAPPINGS, NODE_DISPLAY_NAME_MAPPINGS
__all__ = ["NODE_CLASS_MAPPINGS", "NODE_DISPLAY_NAME_MAPPINGS"]
pyproject.toml
+1 -1
View File
@@ -1,7 +1,7 @@
[project]
name = "diffrhythm_mw"
description = "Blazingly Fast and Embarrassingly Simple End-to-End Full-Length Song Generation. A node for ComfyUI."
version = "1.0.1"
version = "1.0.2"
license = {file = "LICENSE"}
dependencies = ["# accelerate==1.4.0", "# torchdiffeq==0.2.5", "# torchaudio==2.6.0", "# transformers==4.49.0", "# librosa==0.10.2.post1", "# pyarrow==19.0.1", "# pandas==2.2.3", "# bitsandbytes", "# jieba==0.42.1", "# cn2an==0.5.23", "# pypinyin==0.53.0", "# onnxruntime", "LangSegment", "x-transformers", "pylance", "ema-pytorch", "prefigure", "muq", "mutagen", "pyopenjtalk", "pykakasi", "Unidecode", "phonemizer"]
requirements.txt
+10 -12
View File
@@ -1,15 +1,13 @@
# accelerate==1.4.0
# torchdiffeq==0.2.5
# torchaudio==2.6.0
# transformers==4.49.0
# librosa==0.10.2.post1
# pyarrow==19.0.1
# pandas==2.2.3
# bitsandbytes
# jieba==0.42.1
# cn2an==0.5.23
# pypinyin==0.53.0
# onnxruntime
accelerate
torchdiffeq
librosa>=0.10.2.post1
pyarrow
pandas
bitsandbytes
jieba
cn2an
pypinyin
onnxruntime
LangSegment
x-transformers
pylance