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

best model at epoch None step None #3

Open
cguetot opened this issue Feb 11, 2021 · 2 comments
Open

best model at epoch None step None #3

cguetot opened this issue Feb 11, 2021 · 2 comments

Comments

@cguetot
Copy link

cguetot commented Feb 11, 2021

Hi there again,

I am trying to create an small model. However I got this message


2021-02-11 09:06:40,625 [INFO] __main__: training mode
2021-02-11 09:06:40,626 [INFO] data_reader: input spectrum file: /home/cguetot/projects/epigg/deepnovo/species/models/A3/input_files/spectrum.mgf
2021-02-11 09:06:40,626 [INFO] data_reader: input feature file: /home/cguetot/projects/epigg/deepnovo/species/models/A3/input_files/feature.train.csv
2021-02-11 09:06:40,626 [INFO] data_reader: build spectrum location from scratch
2021-02-11 09:06:43,778 [INFO] data_reader: read 4737 features, 0 skipped by mass, 0 skipped by unknown modification, 0 skipped by length
2021-02-11 09:06:43,778 [INFO] train_func: 148 steps per epoch
2021-02-11 09:06:43,778 [INFO] data_reader: input spectrum file: /home/cguetot/projects/epigg/deepnovo/species/models/A3/input_files/spectrum.mgf
2021-02-11 09:06:43,778 [INFO] data_reader: input feature file: /home/cguetot/projects/epigg/deepnovo/species/models/A3/input_files/feature.valid.csv
2021-02-11 09:06:43,778 [INFO] data_reader: read cached spectrum locations
2021-02-11 09:06:43,785 [INFO] data_reader: read 876 features, 0 skipped by mass, 0 skipped by unknown modification, 0 skipped by length
2021-02-11 09:06:43,811 [INFO] train_func: initialize a set of new parameters
2021-02-11 09:24:43,300 [INFO] train_func: best model at epoch None step None

when I try with more data everything seems to be ok.

  • What is the error?
  • What can I do to fix it?
    best,

Carlos

@volpato30
Copy link
Owner

Please change steps_per_validation in config file to 50. Your training set is very small and contains only has 148 batchs per epoch. The default value of steps_per_validation is 300 which means the validation step will not be run and validation loss won't be evaluated. That's why it prompted "best model at epoch None step None".

@cguetot
Copy link
Author

cguetot commented Feb 11, 2021 via email

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