Open rishabhramteke opened 2 years ago
Maybe your machine ran out of memory? Can you enable verbose with verbose=True
?
Thank you for your response. I will try debugging. Do you think there is any memory leakage in the trangulation code?
triangulation.main(..., verbose=True)
. No afaik there is no memory leak.
Traceback (most recent call last):
File "pipe_test.py", line 75, in
I checked by using dmesg -T and got this output : Out of memory: Killed process 86403 (colmap) total-vm:18077248kB, anon-rss:15917864kB, file-rss:2812kB, shmem-rss:0kB, UID:1000 pgtables:31392kB oom_score_adj:0
Do you think increasing my RAM would help? Or is there some other issue (maybe in the features extracted or matches)
The code gets stuck during bundle adjustment. Maybe the issue is with that. Should I make the check = False in "Hierarchical-Localization/hloc/triangulation.py", line 138, in run_triangulation subprocess.run(cmd, check=True)"?
Got this error while running robotcar pipeline : Traceback (most recent call last): File "pipe_test.py", line 78, in
triangulation.main(
File "/home/ubuntu/Hierarchical-Localization/hloc/triangulation.py", line 181, in main
stats = run_triangulation(
File "/home/ubuntu/Hierarchical-Localization/hloc/triangulation.py", line 138, in run_triangulation
subprocess.run(cmd, check=True)
File "/usr/lib/python3.8/subprocess.py", line 516, in run
raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command '['colmap', 'point_triangulator', '--database_path', 'outputs/robotcar_test/sfm_superpoint+superglue/database.db', '--image_path', 'datasets/robotcar_test/images', '--input_path', 'outputs/robotcar_test/sfm_superpoint+superglue/empty', '--output_path', 'outputs/robotcar_test/sfm_superpoint+superglue', '--Mapper.ba_refine_focal_length', '0', '--Mapper.ba_refine_principal_point', '0', '--Mapper.ba_refine_extra_params', '0']' died with <Signals.SIGKILL: 9>.