Request a Wan2.1 LoRA on our Discord and we will train and open-source it for free.
Join our Discord to generate videos of the Deflate Effect LoRA for free.
Wan2.1 14B I2V 480p v1.0:
Trained on 1.5 minutes of video comprised of 20 short clips of things being deflated. This was trained on the Wan.21 14B I2V 480p model.
The trigger word is: 'd3d1at3 deflate it'
See below for some example prompts that have worked very well. I believe the prompt structure can be very similar as the examples are; you just need to specify the object being deflated.
Recommended Settings:
LoRA strength = 1.0
Embedded guidance scale = 6.0
Flow shift = 5.0
Here's a 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 is a slight modification to Kijai's version, with the main difference being the addition of the WanVideo Lora Select node, connected to the 'lora' field of the WanVideo Lora Select node. Find Kijai's original workflow here:
https://github.com/kijai/ComfyUI-WanVideoWrapper/blob/main/example_workflows/wanvideo_480p_I2V_example_02.json
Prompt Examples:
Example 1: The video opens with a [object]. As the video progresses, the [object] begins to d3d1at3 deflate it, gradually shrinking and losing shape, eventually flattening completely into a lifeless, deflated mass on the ground.
Let me know if there are any questions, I'll be happy to help!