Nodes Browser
ComfyDeploy: How ComfyUI-StableAudioOpen works in ComfyUI?
What is ComfyUI-StableAudioOpen?
The implementation of the audio generation model stable-audio-open in ComfyUI enables ComfyUI to achieve text-to-audio functionality.
How to install it in ComfyDeploy?
Head over to the machine page
- Click on the "Create a new machine" button
- Select the
Edit
build steps - Add a new step -> Custom Node
- Search for
ComfyUI-StableAudioOpen
and select it - Close the build step dialig and then click on the "Save" button to rebuild the machine
介绍
音频生成模型 stable-audio-open 在ComfyUI中的实现,让ComfyUI也可以实现文生音频功能。<br>
安装
方法1:
- 进入节点目录,
ComfyUI/custom_nodes/
git clone https://github.com/MoonHugo/ComfyUI-StableAudioOpen.git
cd ComfyUI-StableAudioOpen
pip install -r requirements.txt
- 重启ComfyUI
方法2:
直接下载节点源码包,然后解压到custom_nodes目录下,最后重启ComfyUI
方法3:
通过ComfyUI-Manager安装,搜索“ComfyUI-StableAudioOpen”进行安装
使用说明
参数说明
prompt: 正向提示词,比如:The sound of dog barking.
<br>
negative_prompt: 反向提示词,比如:Low quality.
<br>
seed: 整数类型,设置种子值来确保结果的可重复性,取值范围在0到0xffffffffffffffff之间。<br>
control_after_generate: 种子变化方式,有固定、增加、减少、随机四种方式。<br>
steps: 生成音频步数,比如:250
<br>
cfg_scale: 取值范围是0到10,默认值为6,值越高,生成的内容通常更紧密地符合给定的描述,但可能牺牲一些创造性。<br>
sampler_type: 采样类型,有dpmpp-3m-sde、dpmpp-2m-sde、k-heun、k-dpmpp-2s-ancestral、k-dpm-2、k-dpm-fast六种采样类型。<br>
audio_length: 设置生成的音频长度,单位是秒,最高可以生成47秒的音频。<br>
save_path: 设置保存音频路径,比如:C:\Users\Desktop\
,如果为空,则默认保存在ComfyUI\output\stable-audio-open-1.0
里面。<br>
load_local_model: 加载本地模型,默认值是False。<br>
local_model_path: 加载本地模型的时候需要把load_local_model设置为True,并把local_model_path设置为本地模型所在路径,例如:J:\stable_audio_open,如下所示:
模型下载地址:https://huggingface.co/stabilityai/stable-audio-open-1.0
社交账号
- Bilibili:我的B站主页
感谢
感谢stabilityai/stable-audio-open-1.0仓库的所有作者 stabilityai/stable-audio-open-1.0