search

mjq11302010044 / RRPN

Arbitrary-Oriented Scene Text Detection via Rotation Proposals (TMM 2018)
Other
436 stars 141 forks source link

Training Speed and Iteration Count #35

Closed ghost closed 6 years ago

ghost commented 6 years ago

Helo @mjq11302010044. I'm trying to create a model on my own just like your trained model that you shared.

But the training speed is very low. The iteration speed is aroud 5 seconds and the iteration count is 490000.

Do you have any trick to speed up the training speed? What can be the least iteration count to get the demo works?

mjq11302010044 commented 6 years ago

@gyaliniz We are using Cuda 8 + Cudnn 5 in a Titan X, and this configuration can have ~1iter / s with input of (1000, 1700) And in fact you can have a pretty good visualization after nearly 50K iters