AnimateLCM-I2V

LoRAs and Motion weights for fast Image-to-Video generation.

We support high-quality image conditioned video generation in 4~8 steps.

Important Notes:

1. Use LCMScheduler for sampling.

2. CFG should be kept between 1 and 2. CFG is set to 1 by default.

3. LoRA weights is set to 0.8 by default.

4. Motion scale factor is set to 0.8 by default.

5. 16 frames is preferred. Too long or too short videos tend to cause generation failure.

6. Image size greatly influences the generation quality.

7. Generation failure might occur due to the limited model size and training resources.

For more details, please refer to https://github.com/G-U-N/AnimateLCM

Links

Contact: [email protected]

[Paper]

[Code]

[Demos]