search

clovaai / deep-text-recognition-benchmark

Text recognition (optical character recognition) with deep learning methods, ICCV 2019
Apache License 2.0
3.72k stars 1.09k forks source link

Can I run train.py ignoring characters that are in the dataset but not in the --character? #393

Open jaewon7963 opened 1 year ago

jaewon7963 commented 1 year ago

Nice to meet you! I am currently training Korean. In the beginning, i was training 1011 characters. I trained up to 81% accuracy, but i thought i'd have to add a character later, so i added a few characters to the --character parameter.

Before adding char, i could train by ignoring characters that were not in the --character parameter. However, after adding char, an error occurs if a char that is not in the parameter is found in dataset.

Like the title of this issue, is there a way to solve it? 00000000

i disable the data_filtering_off option, but the error persists.. 0001

ghost commented 1 year ago

How about try some characters with open txt file? When i 1st try other language, similar error appears. I open character using txt file, it doesn't appear.