search

G-U-N / AnimateLCM

AnimateLCM: Let's Accelerate the Video Generation within 4 Steps!
https://animatelcm.github.io
MIT License
567 stars 41 forks source link

Hi，the motion is smooth, but video is blurry,why? #23

Closed ersanliqiao closed 3 months ago

ersanliqiao commented 3 months ago

image

ersanliqiao commented 3 months ago

报纸墙 可爱粉色长发少女 è 桌面壁纸_彼岸壁纸

https://github.com/G-U-N/AnimateLCM/assets/8411769/b2863f32-df5d-4cc6-b0ca-cb5eceb8bdb9

seems the result is unstable

G-U-N commented 3 months ago

Are you using the AnimateLCM-I2V?

ersanliqiao commented 3 months ago

Yes, SVD I2V

G-U-N commented 3 months ago

Yes, it is not as stable as the original SVD. Generation failure happens sometimes. Try to increase the inference steps when needed.

ersanliqiao commented 3 months ago

OK

ersanliqiao commented 3 months ago

Hi，where to download realistic1.safetensors,realistic2.safetensors?

G-U-N commented 3 months ago

Realistic1. Realistic2. Put them in the Personalized Folder.

ersanliqiao commented 3 months ago

Realistic1. Realistic2. Put them in the Personalized Folder.

Thank you very much!!

截屏2024-05-09 下午3 54 01
G-U-N commented 3 months ago

My pleasure : )

ersanliqiao commented 3 months ago

Hi, i have a question about dataset for svd, what dataset did you to train?

G-U-N commented 3 months ago

WebVid 2M. The implementation details on SVD are not included in the paper currently. See issue-22 for more training details.

ersanliqiao commented 3 months ago

hi，will you release sd1.5 training code of t2v and i2v?

G-U-N commented 3 months ago

yes, but will release them a bit later.