Hugging Face's logo

guoyww
/
animatediff-motion-lora-v1-5-3

Text-to-Video
Diffusers
Safetensors
animatediff
Model card Files Community
animatediff-motion-lora-v1-5-3
File size: 899 Bytes 
8b5ff27
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
 
---
library_name: diffusers
pipeline_tag: text-to-video
tags:
  - animatediff
---

# Motion LoRAs

Motion LoRAs allow adding specific types of motion to your animations. 

![animatediff-zoom-out-lora.gif](https://cdn-uploads.huggingface.co/production/uploads/6126e46848005fa9ca5c578c/13B2HSVUuZ1t9UseffdHp.gif)

Currently, the following types of loras are available for models using the `guoyww/animatediff-motion-adapter-v1-5-2` checkpoint.

- Zoom In/Out
- Pan Left/Right
- Tilt Up/Down
- Rolling Clockwise/Anticlockwise

This lora improves the generation quality and is meant to be used with AnimateDiff v3 `guoyww/animatediff-motion-adapter-v1-5-3` checkpoint and [SparseCtrl](https://arxiv.org/abs/2311.16933) checkpoints.

Please refer to the [AnimateDiff documentation](https://huggingface.co/docs/diffusers/main/en/api/pipelines/animatediff) for information on how to use these Motion LoRAs.