the problem about input_size

MichalBusta / E2E-MLT

E2E-MLT - an Unconstrained End-to-End Method for Multi-Language Scene Text

MIT License

291 stars 84 forks source link

the problem about input_size #69

Closed duxiangcheng closed 4 years ago

duxiangcheng commented 4 years ago

Hi, Thanks for sharing the amazing code. I notice that the input_size is 256 which is used in data_gen.py. I know the function of the input_size is the size of the cut images. But why is the size is 256?

MichalBusta commented 4 years ago

Hi, it is just debug setting (to fit GPU on my notebook) see: https://github.com/MichalBusta/E2E-MLT#training I would recommend to train with input_size 512