TextSnake.pytorch icon indicating copy to clipboard operation
TextSnake.pytorch copied to clipboard

validation loss dont decrease under 0.45

Open johnSmith1990 opened this issue 4 years ago • 1 comments

Hi dear,

thank you for your code. but my both train and validation loss don't decrease under 0.45 with 400 epochs. May you help me for more config such as learning rate and so on?

regards

johnSmith1990 avatar Aug 15 '20 05:08 johnSmith1990

What dataset you are training on? if total-text or icdar2013/15, the configuration in README can reproduce the result in table. If not in two datasets above, please check dataloader you implement is correct first.

princewang1994 avatar Aug 16 '20 12:08 princewang1994