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

NN_test_pairs_trained_nw.txt is absent #2

Open
canorbal opened this issue Jul 6, 2018 · 3 comments
Open

NN_test_pairs_trained_nw.txt is absent #2

canorbal opened this issue Jul 6, 2018 · 3 comments

Comments

@canorbal
Copy link

canorbal commented Jul 6, 2018

Hi!

There is some unknown .txt file "NN_test_pairs_trained_nw.txt" which you used in cnn_part/data_provider.lua. But there is only a NN_7scenes.txt file in cnn_part/data/. Could you please explain me how you generated NN_test_pairs_trained_nw.txt ?

@canorbal
Copy link
Author

canorbal commented Jul 6, 2018

I would be very appreciated if you also could put "NN_test_pairs_trained_nw.txt" in this repo.

@imelekhov
Copy link
Collaborator

Hi,

Thanks to point me out! This is definitely a bug coming from some initial experiments. I have added two txt-files with the GT training and validation data (db_all_med_hard_train.txtand db_all_med_hard_valid.txt) and accordingly modified data_provider.lua as well. The list of image pairs is generated by calculating an overlap between camera views. Then, some threshold is applied to this overlap to get 'hard' and 'medium' camera pairs.

@canorbal
Copy link
Author

canorbal commented Jul 7, 2018

Thanks a lot! Could you please tell me which loss did you scored while training? And maybe you have any plot epochs vs. mse_loss on training? I mean loss on pairs in db_all_med_hard_train.txt and loss on db_all_med_hard_valid.txt vs. epochs.

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