Spaces:
Running
on
A10G
Having an issue running it locally saving the output video.
I was trying to run the gradio version but after processing it would just put error in the output screen. So I made a virtual environment and installed everything:
pip install modelscope==1.8.4
pip install xformers==0.0.20
pip install torch==2.0.1
pip install open_clip_torch>=2.0.2
pip install opencv-python-headless
pip install opencv-python
pip install einops>=0.4
pip install rotary-embedding-torch
pip install fairscale
pip install scipy
pip install imageio
pip install pytorch-lightning
pip install torchsde
I also had to install easydict.
Everything seems to run fine.
my steps include:
from modelscope.pipelines import pipeline
from modelscope.outputs import OutputKeys
pipe = pipeline(task='image-to-video', model='damo/Image-to-Video', model_revision='v1.1.0')
output_video_path = pipe('./1.png', output_video='./output.mp4')[OutputKeys.OUTPUT_VIDEO]
about 13 gigs of vram get used and the GPU is at 100% for about a minute then I get:
output_video_path = pipe('./1.png', output_video='./output.mp4')[OutputKeys.OUTPUT_VIDEO]
2023-11-07 17:05:35,506 - modelscope - WARNING - task image-to-video input definition is missing
2023-11-07 17:10:00,577 - modelscope - ERROR - Save Video Error with
2023-11-07 17:10:00,581 - modelscope - WARNING - task image-to-video output keys are missing
So "save Video Error with" and then just blank. So I'm not sure what the problem is. I have tried different save locations, tried putting full paths to both the input and output file but still the same error. I looked in /tmp just to see if there was anything saved there but there is not. if I run >>> print(output_video_path) after the error it gives the expected output:
print(output_video_path)
./output.mp4
I'm stumped.