360 Degree Rotation (Microwave Rotation Wan2.1 I2V LoRA)

Apply for the Wan2.1 LoRA model on our Discord, and we will train and open-source it for free.

Join our Discord to generate 360-degree rotation LoRA videos for free.

Wan2.1 14B I2V 480p Version 1.0 Description:

This model is trained based on 12 short videos of 360-degree rotating objects (total duration of 30 seconds), with the training foundation being the Wan2.1 14B I2V 480p model.

Trigger word: 'r0t4tion 360 degrees rotation'

Below are examples of prompt words that have been tested to perform well. You can also view the videos I posted here and the annotation texts used during generation. Each video's input image only uses the first frame.

Recommended parameter settings:

LoRA strength = 1.0
Embedded guidance scale = 6.0
Flow displacement = 5.0

Here is the link to the Wan2.1 I2V LoRA inference workflow I used to generate these videos:
https://huggingface.co/Remade/Squish/blob/main/workflow/wan_img2video_lora_workflow.json

This workflow is a fine-tuned version of Kijai's, with the main difference being the addition of the WanVideo Lora Select node, which is connected to the 'lora' field of the 'WanVideo Lora Select node' in the original workflow. For Kijai's original workflow, see:
https://github.com/kijai/ComfyUI WanVideoWrapper/blob/main/example_workflows/wanvideo_480p_I2V_example_02.json

Prompt word examples:

The video shows a man sitting on a chair, and the man and the chair complete the r0t4tion 360 degrees rotation action together.

The video presents a Pomeranian puppy sitting on a gravel ground, performing the r0t4tion 360 degrees rotation.

If you have any questions, feel free to ask, and I will be happy to assist you!