Error in call to target 'sam_pt.point_tracker.pips.tracker.PipsPointTracker'

[remo-help]

Hi! I'm trying to replicate the demo, and I'm running into an issue. I have CUDA 11.8 installed. TensorRT is also installed and can be loaded into python, so I'm not sure what the issue is there.

I tried fresh installs of the environment, CUDA, Tensorflow, Pythorch and TensorRT.

This is the relevant log:

Loading PIPS model from /mnt/sam-pt/models/pips_ckpts/reference_model

reading ckpt from /mnt/sam-pt/models/pips_ckpts/reference_model

Error executing job with overrides: ['frames_path=${hydra:runtime.cwd}/data/demo_data/bees/', 'query_points_path=null', 'longest_side_length=1024', 'frame_stride=1', 'max_frames=-1']

Error in call to target 'sam_pt.point_tracker.pips.tracker.PipsPointTracker':

AssertionError()

full_key: model.point_tracker

[m43]

Hi! Thanks for reporting this issue and sorry for the delay. From the log snippet you've shared, it appears that an AssertionError is being thrown. However, the current details are not comprehensive enough to precisely identify the cause of the failure. To gain a better understanding of the issue, it would be helpful if you could share a more extensive portion of your log output, particularly any error stack trace.

Additionally, I recommend setting the HYDRA_FULL_ERROR environment variable to 1 to get a more detailed error log. You can set this variable in your environment by executing export HYDRA_FULL_ERROR=1 in your terminal before running the demo command.

[remo-help]

Hi, thanks for your reply. Turns out this was an issue on my server. Pip was installing packages for a different python than the one that the virtual environment was using. This issue can be closed