Closed MartinPlantinga closed 6 years ago
Hi, did you change anything in the prototxt file and/or the code (e.g., number of affordance class)? The top shape is correct for 10 affordance classes Top shape: 1 10 244 244 (595360)
, but the bottom shape is 0, which is not correct.
Yes, I did change the $Affordance-Net/lib/rpn/proposal_target_layer.py
for a script from py-faster-rcnn. This caused the error. The reason I did this is because when using your $Affordance-Net/lib/rpn/proposal_target_layer.py
I got type-error messages. These are probably caused by using a different version of Numpy (v 13.3). I have been able to partially solve these errors, but not all of them.
Therefore I am going to close this Issue and start a new one.
When trying to train with the IIT-AAF dataset I get an error with regards to the Number of labels:
Any idea how to solve this?
Thanks