Walking motion LoRAs for Animatediff v2
Training
This LoRA was trained on RunPod cloud GPUs.
File summary
| Name | Compatible Motion Model | Base AR (w/h) | Frames | input videos |
|---|---|---|---|---|
| 1900_cseti_6931403_walking-frontview-8vid_mv2.safetensors | v2 | (3:2) | 16 | 8 |
| 1600_cseti_1065173_walking-backview-8vid_mv2.safetensors | v2 | (3:2) | 16 | 8 |
| 1100_cseti_6134549_walking-left-sideview-3vid.safetensors | v2 | (3:2) | 16 | 3 |
| 1100_cseti_8836021_walking-right-sideview-3vid.safetensors | v2 | (3:2) | 16 | 3 |
General recommendation
- These are finetuned on a v2 model. They work also with AnimateLCM but don't work with v3 models
- Animatediff v3 adapter lora is recommended regardless they are v2 models
- If you want more motion try incrasing the scale multival (e.g. to 1.2)
- I recommend using the above Base AR (Aspect ratio) for inference
- Try playing with the lora strength and scale multival like increasing the scale multival and lowering the lora strength.
- I recommend using IP-Adapter, however it reduces the effect of the mLORAs, so play with the strength. "Ease in out" weight type is a good starting point.
Details and examples
Walking front view
- Recommended keywords: front-view, walking
- LoRA strength: 0.8-1.0
- scale multival / motion scale: 1.0-1.2
Walking back view
- Recommended keywords: back-view, walking
- LoRA strength: 0.8-1.0
- scale multival / motion scale: 1.0-1.2
Walking right view
- Recommended keywords: side-view, facing right, walking
- LoRA strength: 0.5-1.0
- scale multival / motion scale: 1.0-1.2
Walking left view
- Recommended keywords: side-view, facing left, walking
- LoRA strength: 0.5-1.0
- scale multival / motion scale: 1.0-1.2
Support
Producing and sharing this kind of open-source work requires renting cloud GPUs, which gets expensive quickly. If you find it useful and would like me to keep contributing, your support is very much appreciated: