search

yuantailing / ctw-baseline

Baseline methods for [CTW dataset](https://ctwdataset.github.io/)
MIT License
329 stars 88 forks source link

about the run efficiency #23

Open parquets opened 6 years ago

parquets commented 6 years ago

thank you for your sharing. I haven't run this respository successfully. Actually the fiIes is so complex and I don't know how to start. I want ask about the efficiency. How many times(average time) does the method run per image. And the trained weight is getting from the lib your mention in paper? How can I use it?

yuantailing commented 6 years ago

A1 YOLOv2, TITAN X (Pascal): 0.2 sec / image (1216x1216) for each thread, characters larger than 32x32 is hopeful to be detected. You can run 3 threads on each card. You can improve the code to mini-batch version to speed up it.

A2 Yes, follow the tutorial and you will reproduce the results, despite of random factors.

parquets commented 6 years ago

Sorry，I tried but I could not find the main function of the project. Beg your help. Could you tell me where is the image read in and where is the detection and recognition result output or display. I could not find the code.

yuantailing commented 6 years ago

Please see https://github.com/yuantailing/ctw-baseline/blob/master/tutorial/1-basics.ipynb and other 2 tutorials.