Diffutoon example generate video take too long

modelscope / DiffSynth-Studio

Enjoy the magic of Diffusion models!

Apache License 2.0

6.39k stars 574 forks source link

Diffutoon example generate video take too long #211

Open websieu opened 1 week ago

websieu commented 1 week ago

Hi there, great work team! Thank so much!!!

I just ran example of diffutoon Env: cuda 12, python 3.11 Card 80G ram

i just tested a video with 1 minute and it took about 40 minutes for complete cartoonized my video. Is this normal? Or my enviroment didn't use cuda? Thanks!

nitinmukesh commented 1 week ago

You can check if torch is installed with CUDA

activate virtual environment pip list

Processing time depends on other factors like length of input video, width/height, steps, etc

Neethan54 commented 4 days ago

Hii , for me also it is taking 45 minute for 6 seconds video, and i have torch with cuda version, I have 48 GB GPU, width and height 720 and 480 respectively.

What is the reason ?

Im simply running this code , without changing anything https://github.com/modelscope/DiffSynth-Studio/blob/main/examples/video_synthesis/cogvideo_text_to_video.py