UnicodeDecodeError: 'utf-8' codec can't decode byte 0x80 in position 0: invalid start byte

[Moushab]

Hi
am getting the following error when running measure_map.py
UnicodeDecodeError: 'utf-8' codec can't decode byte 0x80 in position 0: invalid start byte

Using TensorFlow backend.
Traceback (most recent call last):
File "measure_map.py", line 106, in
C = pickle.load(f_in)
File "/usr/lib/python3.6/codecs.py", line 321, in decode
(result, consumed) = self._buffer_decode(data, self.errors, final)
UnicodeDecodeError: 'utf-8' codec can't decode byte 0x80 in position 0: invalid start byte

[bostankhan6]

I have having the same issue. Can anyone help us with this?

[bostankhan6]

I have resolved the the above error in the measure_map.py file. The error is when the file is read:

with open(config_output_filename, 'r') as f_in:
C = pickle.load(f_in)

above instead of writing 'r' write 'rb' It will resolve the first issue.

There is a second issue and it is that it does not recognize the keyword 'difficult' in a certain "if" condition part of the code. I don't know from where this comes from but it is unable to find the keyword in the specified dictionary. When i commented this out the code ran fine. However, I am not sure if commenting this condition out would affect the mAP evaluation for the dataset.

I used the following command to run the measure_map.py code:
python measure_map.py -o simple -p test.csv

The link of the edited version of the measure_map is given below:
https://github.com/bostankhan6/Object-Detection-YoloV3-RetinaNet-FasterRCNN/blob/master/Faster%20RCNN/measure_map.py