Trying to replicate the training in SSD_training.ipynb with my own dataset
I am getting
AttributeError: 'float' object has no attribute 'log'
In ssd_utils.py; line 114: encoded_box[:, 2:4][assign_mask] = np.log(box_wh / assigned_priors_wh)
called from the generator...
encoded_boxes = np.apply_along_axis(self.encode_box, 1, boxes[:, :4])
Trying to replicate the training in SSD_training.ipynb with my own dataset I am getting
AttributeError: 'float' object has no attribute 'log'
In ssd_utils.py; line 114:encoded_box[:, 2:4][assign_mask] = np.log(box_wh / assigned_priors_wh)
called from the generator...
encoded_boxes = np.apply_along_axis(self.encode_box, 1, boxes[:, :4])
What is this error?