ComfyUI Node
Float_Animator
Generates speaking portrait video frames from an image and audio using the FLOAT model. / 使用 FLOAT 模型从图像和音频生成说话肖像视频帧。
Float_Animator
- ref_image
- audio
- animated_frames
- audio
- fps
◄seed15►
◄emotionnone►
◄fps25►
◄aud_cfg_scale2.0►
◄ref_cfg_scale1.0►
◄emo_cfg_scale1.0►
◄modelfloat.pth►
◄auto_cropfalse►
CategoryAnimator
Inputs (10)
| Name | Type | Default | Description |
|---|---|---|---|
| ref_image | IMAGE | The still portrait image to animate. / 待动画化的静态肖像图像 | |
| audio | AUDIO | The driving audio for animation. / 驱动动画的音频 | |
| seed | INT | 150–18446744073709550000 | Random seed for reproducibility. / 结果可复现的随机种子 |
| emotion | COMBO | none | Target emotion style. 'none' infers from audio. / 目标情感风格。'none'从音频推断。 |
| fps | FLOAT | 251–60 | Frames per second for output animation. / 输出动画的每秒帧数 |
| aud_cfg_scale | FLOAT | 2.0 | Classifier-free guidance scale for audio control. / 音频控制的无分类器引导尺度 |
| ref_cfg_scale | FLOAT | 1.0 | Classifier-free guidance scale for reference image control. / 参考图像控制的无分类器引导尺度 |
| emo_cfg_scale | FLOAT | 1.0 | Classifier-free guidance scale for emotion control. / 情感控制的无分类器引导尺度 |
| model | COMBO | float.pth | Data will be auto_downloaded for the first time. After completed, refresh the page to reload the list / 首次运行节点会自动下载数据，下载完成后刷新页面以加载列表 |
| auto_crop | BOOLEAN | false | Automatically crop face in reference image. / 自动裁剪参考图像中的人脸 |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| animated_frames | IMAGE | — |
| audio | AUDIO | — |
| fps | FLOAT | — |