search

summerlvsong / Aggregation-Cross-Entropy

Aggregation Cross-Entropy for Sequence Recognition. CVPR 2019.
303 stars 60 forks source link

loss decline but accuracy near to 0 #25

Open bjlgcxc opened 4 years ago

bjlgcxc commented 4 years ago

I train a model(CRNN) base on dataset synth90k, through the loss decline step by step, the accuracy is near to 0 all the time. What casue this problem?

bjlgcxc commented 4 years ago

@summerlvsong

dqfx1949 commented 4 years ago

I also find this problem in the training of crnn,have you solved it?

Hubert2102 commented 4 years ago

+1

bjlgcxc commented 4 years ago

I also find this problem in the training of crnn,have you solved it?

not yet

Zhang-O commented 2 years ago

I train a model(CRNN) base on dataset synth90k, through the loss decline step by step, the accuracy is near to 0 all the time. What casue this problem?

make sure your decode method works correctly. you'd better test it by some your customed fake data.