Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Scores in log file for training #26

Closed
sanakhamekhem opened this issue Nov 25, 2017 · 2 comments
Closed

Scores in log file for training #26

sanakhamekhem opened this issue Nov 25, 2017 · 2 comments

Comments

@sanakhamekhem
Copy link

Hi,
I would like to understand the meaning of these two lines in the log file:
Is the training going ok??

Learning-rate-control: error key 'dev_score' from {'dev_error': 0.99728335186974681, 'dev_score': 3.6586107690151706}
epoch 1 score: 3.96068657248 elapsed: 3:19:03 dev: score 3.65861076902 error 0.99728335187

@albertz
Copy link
Member

albertz commented Nov 25, 2017

Learning-rate-control is the module which does control the learning rate (e.g. decreasing or Newbob or so), depending on some score information it gets per epoch. It just states what it stored as information for this epoch (that is {'dev_error': 0.99728335186974681, 'dev_score': 3.6586107690151706}). The other line just gives a summary with the same information and some more like how long this epoch took to train. Then the Learning-rate-control also states which key it uses to do its logic (in this case dev_score, which is probably Cross-Entropy or so). dev_error is most likely the frame-error-rate. So this is pretty high. But I don't know exactly what you are doing.

Note that the GitHub issue tracker is about reporting bugs. This doesn't seem like a bug, so I'm closing this now.

@albertz albertz closed this as completed Nov 25, 2017
@sanakhamekhem
Copy link
Author

Thank you Mr Albertz for your response

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants