So I’ve been running docker odm for 4577 images with a --split of 200,–feature-quality=low, and gpu acceleration. It has been stuck on following step for 3 days now. My current system has 64gb ram. Is there any solution?
1 Like
The sure solution to reduce headaches and guesswork when processing is always more RAM.
Other means to address this would be scaling back image size when creating the Task (Resize…), or reducing quality parameters used to create the Task. You might also need to scale down the split size, as 200 isn’t a firm number, more more a guideline that can vary widely depending upon split-overlap, spatial distribution of the images, etc. So you recommend 200, but you can easily get a submodel with 500 or more images.
issue with taking a 500 split is that it again causes a ram insufficient error. I guess increasing the ram is the only solution possible
1 Like
Making a proper ortophoto really eats ram, so only do that if you need it. I just calculate the pointcloud and if I need an orto I do that from the pointcloud in CloudCompare, it will not look as nice but for my purpose’s it’s great.
1 Like
I meant to direct you to make the split smaller. Try 50 or 75.
1 Like
alright i’ll try out working with only a pointcloud as well. thanks
1 Like
ok got it. will try with a smaller split as well. thanks
2 Likes
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.