Deep-MAC error: Unexpected Keyword Argument 'box_ind'

Question

I'm encountering an issue while attempting to execute the "deepmac_colab.ipynb" notebook.

Everything runs smoothly until the final line:

masks = prediction_function(tf.convert_to_tensor(image),
                            tf.convert_to_tensor(boxes, dtype=tf.float32))

TypeError: Got an unexpected keyword argument 'box_ind'

This error seems to originate from the function "reframe_box_masks_to_image_masks"

I've tested two TensorFlow versions, specifically 2.14.0 and 2.8.0.

The only change I made is that I didnt annotate anything above and I just use our preannotated boxes.

# boxes_list = [np.array([[0.000, 0.160, 0.362, 0.812],
#                         [0.340, 0.286, 0.472, 0.619],
#                         [0.437, 0.008, 0.650, 0.263],
#                         [0.382, 0.003, 0.538, 0.594],
#                         [0.518, 0.444, 0.625,0.554]], dtype=np.float32)]

Could you please help me make changes to achieve the correct answer? Thanks in advance.

Answer 1

Please note that this function does not take a 'box_ind' argument. It only takes the following arguments:

reframe_box_masks_to_image_masks(box_masks, boxes, image_height, image_width)

To fix this issue, you should update the call to reframe_box_masks_to_image_masks to match the expected arguments.