Skyreels & Hunyuan video: Img2video

Introduction

The combination of Skyreels and Hunyuan Vidio represents an innovative fusion of video processing and generation technologies. Skyreels provides a solid foundation for workflows with its powerful video processing capabilities, while Hunyuan Vidio injects new vitality into the creation of video content through its intelligent generation function. The uploaded image is pre-processed with the Resize image node, and then the image color is processed through nodes such as Color Match. Finally, the video synthesis is completed through Video Combine, providing an efficient and flexible solution for creative video production.

SkyReels-V1-Hunyuan-I2V

SkyReels V1 is the first and most advanced open-source human-centric video foundation model. By fine-tuning HunyuanVideo on O(10M) high-quality film and television clips, Skyreels V1 offers three key advantages:

1. Open-Source Leadership: Our Text-to-Video model achieves state-of-the-art (SOTA) performance among open-source models, comparable to proprietary models like Kling and Hailuo.
2. Advanced Facial Animation: Captures 33 distinct facial expressions with over 400 natural movement combinations, accurately reflecting human emotions.
3. Cinematic Lighting and Aesthetics: Trained on high-quality Hollywood-level film and television data, each generated frame exhibits cinematic quality in composition, actor positioning, and camera angles.

https://huggingface.co/Skywork/SkyReels-V1-Hunyuan-I2V

https://huggingface.co/Kijai/SkyReels-V1-Hunyuan_comfy

Recommended models：Ultra

Workflow Overview

How to use this workflow

Step 1：Load Image

Step 2：Crop the image

Official restrictions only support ratios below 544*960

Step 3：Describe your video content

Try to keep the content consistent with the original image, and only need to add additional character movements and the direction of the camera.

Step 4：Set video parameters

Step 5：Get the video

TIP

Skyreels still has room for optimization and cannot directly produce ideal results every time. To get satisfactory results, you may need to try the "draw card" operation multiple times