Open 007invoker opened 1 year ago
Hello. I had the same problem, the cuda version in my docker is 10.0. However, I checked the cog.yaml, the cuda version should be '11.0' and the python version should be '3.7.9'. It seems that I faced the same problem as you, which is the actual version in cuda is lower than it should be. Do you know how to solve this?
too slow, I ran it without the container and used cuda 12.1 ,pytorch 2.1.0. The time of the camel iterations was about 3*40 min using 4090 ...
Use Docker image to install dependencies and successfully run this project container before executing Python run object sketching.py --target file "flamingo.png" --num strokes 8
(Sketching the flamingo with higher level of abstraction, using 8 strokes) Error File "run_oject_sketching. py", line 46
Target=f "{abs_path}/target_images/{args. target_file}". By checking the Python version in the Docker container, version number 2.7.5, it is inferred that the Python version in the Python dependency is incorrect. The correct version should be 3.6 or higher