You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
Thank you for sharing a great work!
I met the error when I ran the script python suds/train.py suds --experiment-name test --pipeline.datamanager.dataparser.metadata_path /mnt/kitti/metadata/kitti-06.json
The text was updated successfully, but these errors were encountered:
That might be due to memory issues, can you monitor that your memory usage isn't getting to the state the OOM killer is terminating your process? There is an --pipeline.datamanager.items_per_chunk parameter that you can reduce is memory is a concern, and if you're using DINO features I'd recommend double-checking that the --pipeline.datamanger.load_on_demand option includes features
Thank you for your reply. I use watch free command to monitor the memory usage. It is true that the code cannot be run due to insufficient memory. But even if I set items_per_chunk to a very small value, the same problem occurs. I also check that load_on_demand includes features.
Hi,
Thank you for sharing a great work!
I met the error when I ran the script
python suds/train.py suds --experiment-name test --pipeline.datamanager.dataparser.metadata_path /mnt/kitti/metadata/kitti-06.json
The text was updated successfully, but these errors were encountered: