Py-faster-rcnn: Training problem : Value Error

[n3011]

I am trying to train a model in dataset with 61 classes including background and hence i set the num of output in prototxt file to 61 and that of bbox reggression to 244.
But I am getting this error during stage1 fast rcnn training. Unable to figure it out.

self.bbox_stds[:, np.newaxis]
ValueError: operands could not be broadcast together with shapes (244,4096) (240,1)

[gavinmh]

Hi @n3011 . I apologize for adding to a closed issue. I am encountering the same error; my data set has two classes. How did you resolve this?

[ednarb29]

I have a similar problem to this one. Does someone has a solution?

[n3011]

@gavinmh @ednarb29 check ur prototxt files. make sure all fields are correct

[EloiZ]

Make sure you change the prototxt files (train and test) as explained in https://github.com/deboc/py-faster-rcnn/tree/master/help and https://github.com/andrewliao11/py-faster-rcnn
Also do not forget to change the first line of solver.prototxt to take into account the train.prototxt
train_net: "models/<YOURDATASET>/VGG16/faster_rcnn_end2end/train.prototxt"

[ednarb29]

Yes, I forgot to adapt the path in the solver. ;)

Thanks!