Why doesn't the loss function go down？
LYJvincent opened this issue · comments
Hi！Excuse me！
I use this network to detect defects.Use MixSoftmaxCrossEntropyLoss loss function, training to 400 eopch, but loss function has been drop down between 0.5 to 0.7.May I ask what the problem is?
MixSoftmaxCrossEntropyOHEMLoss and MixSoftmaxCrossEntropyLoss have the same effect, The learning rate is not going down
@LYJvincent I meet the same problem with you, I use my dataset to train it , the loss always shakes.
How much is the learning rate?
@Tramac the lr = 1e-4, now I get better results。thanks！
@LYJvincent What the values of miou can achieve? and how long have you trained?
@feitiandemiaomi current mean_iou is 0.97602%， about 3.5 hour
@feitiandemiaomi is just for my database not cityspaces
@feitiandemiaomi current mean_iou is 0.97602%， about 3.5 hour
0.97602% or 97.602%?
@Tramac Oh,sorry, is my fault。not 0.97602%, but 97.602%
@LYJvincent Can you share hyperparameters ?
@LYJvincent The paper's miou is 69%, your result is amazing. Hope to learn it ,Thanks
It depends on the datasets.
@feitiandemiaomi Hi, I meet the same problem as you, the training loss fluctuates between 0.9 and 1.1
Have you solved this problem? Could you get the better result
@feitiandemiaomi Hi, I meet the same problem as you. Have you solve this problem? If you have solve this problem, can you share me with your learning rate and batch_size? Thanks!
@CuttlefishXuan Hi! Have you solve this problem? Can you tell me your learning rate and batch_size?T hank you very much!