search

lizhengwei1992 / Semantic_Human_Matting

Semantic Human Matting
532 stars 146 forks source link

The model does not overfit #18

Closed pourfard closed 5 years ago

pourfard commented 5 years ago

Hi,

I decreased the size of training images to just 20 images, but after training tnet loss remains around 0.3. Is this normal? I thought that it should reach around 0 after a few epochs?

tobechao commented 5 years ago

@pourfard hi, pourfard, I train the t-net overfit, loss also about 0.3, have you solved?