binascii.Error: Incorrect padding

[cswhjiang]

Got binascii.Error: Incorrect padding when reading image 300104 from test2014/test2014_resnet101_faster_rcnn_genome.tsv.1 with tools/read_tsv.py. Anything wrong?

Traceback (most recent call last):
  File "read_tsv.py", line 64, in <module>
    read_and_save(os.path.join(in_dir, in_file), out_dir)
  File "read_tsv.py", line 45, in read_and_save
    item['features'] = np.frombuffer(base64.decodestring(item['features']), dtype=np.float32).reshape((item['num_boxes'], -1))
  File "/usr/lib64/python2.7/base64.py", line 321, in decodestring
    return binascii.a2b_base64(s)
binascii.Error: Incorrect padding

[cswhjiang]

Solved. Features for image 300104, 147295, 321486 are duplicated. Just ignoring the binascii.Error will work.

[peteanderson80]

Thanks @cswhjiang for investigating this. Several tsvs were combined and there must have been a couple of duplicates.

[jamiechoi1995]

what does ''duplicated'' mean? I view the three images, they are totally different, and why their features will be duplicated?

[kangkang59812]

Solved. Features for image 300104, 147295, 321486 are duplicated. Just ignoring the binascii.Error will work.

The features are duplicated between them?

[wtliao]

@jamiechoi1995 have you solved this issue? thanks

[nnakamura3]

changing tools/read_tsv.py line 31
from

item[field] = np.frombuffer(base64.decodestring(item[field]), 
                      dtype=np.float32).reshape((item['num_boxes'],-1))

to

item[field] = np.frombuffer(base64.decodestring(item[field][2:-1]), 
                      dtype=np.float32).reshape((item['num_boxes'],-1))

worked for me (python3)