search

rizkiarm / LipNet

Keras implementation of 'LipNet: End-to-End Sentence-level Lipreading'
MIT License
628 stars 224 forks source link

ambiguous training validation loss #47

Closed TieLeeyoung closed 6 years ago

TieLeeyoung commented 6 years ago

hello, i'm running the model, but i'm getting a bad loss, in the sixth epoch i got these scores: [Epoch 6] Out of 256 samples: [CER: 25.898 - 0.910] [WER: 7.492 - 1.457] [BLEU: 0.314 - 0.314] Are these scors meaningful? or the model overfit, and what's is the bleu score? is it the accuracy in speech recognition

qslvhw commented 6 years ago

I have the same question. I only got loss and validation loss after every epoch. How to see the accuracy?

rizkiarm commented 6 years ago

The performance of the model is indicated by Word Error Rate (WER), Character Error Rate (CER), and BLEU score. You can find out more about that by reading the paper. Based on the scores that you have got, it seems that your model was not learning. Please check whether you made some mistake in your setup and make sure that you provide enough (and correct) examples for the model.

I will close this issue for now. Please re-open it if you think that you can provide more detail about your problem and have tried your best to fix it, as I think it is not the issue with the program.