ArtificialGuyBR

Home / Blog / AI video

360 Rotation LoRA for Wan 2.2 I2V: Turn Static Images into Turntable Videos

Create smooth 360-degree product rotations from single images using this Wan 2.2 I2V LoRA. Free, CC0-licensed, works in ComfyUI.

2 sources cited AI video

On this page

  1. What This Model Does
  2. Training Details
  3. How to Use (ComfyUI)
  4. Example Prompts
  5. Recommended Settings
  6. Common Errors and Fixes
  7. FAQ
  8. Sources
  9. Related Models

360 Rotation LoRA for Wan 2.2 I2V: Turn Static Images into Turntable Videos

TL;DR

What This Model Does

The 360 Rotation LoRA specializes in creating smooth turntable animations from a single static image. Feed it a product photo, character render, or any object, and it generates a video that rotates the subject 360 degrees, revealing every angle. This is purpose-built for product showcases, character displays, and any scenario where a rotating view adds value.

The effect is a seamless loop — the rotation completes a full circle without visible jumps or artifacts, making it suitable for direct use in presentations, e-commerce, or social media.

Training Details

How to Use (ComfyUI)

  1. Download both LoRA files:
  1. Place them in ComfyUI/models/loras/
  2. In your Wan I2V workflow, add a Load LoRA node after the base model loader
  3. Connect the model output to the Load LoRA node's model input
  4. Select both LoRA files in the node
  5. Connect the MODEL output to your KSampler
  6. Set strength_model to 0.8-1.0 (start at 1.0, reduce if rotation is too aggressive)
  7. Use the trigger phrase 360 Rotation in your prompt

Example Prompts

360 Rotation

That's it — the trigger phrase alone activates the rotation behavior. You can add descriptive tokens for the subject (e.g., 360 Rotation, ceramic vase, studio lighting) but the rotation itself is driven by the trigger.

Parameter | Value

|---|---

Base model | Wan-AI/Wan2.2-I2V-A14B LoRA strength | 0.8 - 1.0 Trigger phrase | 360 Rotation Resolution | 1080x720 or 664x1000 (match training) Sampler | Any compatible with Wan I2V

Common Errors and Fixes

Error | Cause | Fix | Source

|---|---|---|---

Rotation doesn't appear | Trigger phrase missing or misspelled | Use exact phrase 360 Rotation in prompt | HF README Video has artifacts/jumps | LoRA strength too high | Reduce strength_model to 0.8-0.9 | HF README Only one LoRA loaded | Forgot the second file | Load both low_noise and high_noise variants | HF README Model not found | Wrong base model | Ensure using Wan2.2-I2V-A14B, not T2V or other variants | HF README

FAQ

Q: Does this work with Wan T2V (text-to-video)? A: No — trained specifically for Wan 2.2 I2V (image-to-video). The LoRA expects an input image to rotate.

Q: Can I use this with diffusers instead of ComfyUI? A: Yes, the LoRA files are standard safetensors. Load them with load_lora_weights() in diffusers pipelines configured for Wan I2V.

Q: Is the rotation always 360 degrees? A: The training targets a full 360° turntable loop. The output video length determines rotation speed — longer videos = slower rotation.

Q: What license applies for commercial use? A: CC0-1.0 on HuggingFace (public domain dedication), Apache 2.0 on Civitai. Both permit commercial use.

Sources