Some confusion about Training Model

taigw / brats17

Brain tumor segmentation for MICCAI 2017 BraTS challenge

BSD 3-Clause "New" or "Revised" License

321 stars 130 forks source link

Some confusion about Training Model #11

Closed cb632482135qq closed 6 years ago

cb632482135qq commented 6 years ago

Hello Everyone

  I want to train a model for mysef ,so I run the train.py  with 2015 Brats Data.

 However , the result of loss changes from 0.2 to 0.3 randomly.

 Thus, even I am unable to achieve convergence.

 So, I am confused that whether this is the reason for imperfectly segmentation.

GuojingZhao commented 5 years ago

I have same situation that loss about 0.3 from 500 epoch to 20000 epoch. Is that normal?