I2V Usage Essentials
Input Image Requirements
- Resolution Matching: Suggested aspect ratio of the original image is close to 544:960
- clarity of content: Subjects are clearly outlined, avoiding blurred or low-contrast images
- Format Support: Standard PNG/JPG format without transparency channel
Adjustment of command parameters
python3 video_generate.py
--task_type i2v
--image ./input/your_image.jpg
--prompt "FPS-24, Action description"
Special attention is required:-promptIt should describe the action you wish to add, not what is already in the image.
Effectiveness Optimization Tips
- For portraits: specify expression changes in the cue (e.g. "slowly smiling").
- For objects: specify the trajectory (e.g. "rotating 360 degrees").
- Debugging parameters:
--embedded_guidance_scaleControls the strength of original image feature retention (default 1.0)
This answer comes from the articleSkyReels-V1: An Open Source Video Model for Generating High Quality Human Action VideoThe
May not be reproduced without permission:AI productivity tools " What are the considerations when using SkyReels-V1 for Image to Video (I2V)?
