search

for-ai / CipherGAN

TensorFlow implementation of CipherGAN
https://openreview.net/forum?id=BkeqO7x0-
MIT License
121 stars 24 forks source link

Cannot output accuracy #10

Open kanwangshijie-sudo opened 4 years ago

kanwangshijie-sudo commented 4 years ago 
Hello,I encountered a problem.Following the operation of the README, the training was successful. However, the output is xy_mse, and I want to output the accuracy.Therefore, the xy_mse of the train.py file is changed to acc, But the accuracy cannot be output.

I get the following error: ValueError:MetricSpec without specified prediction_key requires predictions tensor or single element dict, got {'X': <tf.Tensor 'cycle_gan/transforms/G/Reshape_1:0' shape=(64, 100, 202) dtype=float32>, 'Y': <tf.Tensor 'cycle_gan/transforms/F/Reshape_1:0' shape=(64, 100, 202) dtype=float32>}. I look forward to your answer,best wishes,thank you.

aidangomez commented 4 years ago

Which version of TF are you running?

kanwangshijie-sudo commented 4 years ago

Which version of TF are you running?

TF1.4

kanwangshijie-sudo commented 4 years ago

Which version of TF are you running?

I want to ask how you output the accuracy rate at that time. I changed the xy_mse of the train.py file to acc

aidangomez commented 4 years ago

Have you tried looking at the tensorboard for your run? I think we may log more metrics in the tensorboard than what we print to terminal.

kanwangshijie-sudo commented 4 years ago

Have you tried looking at the tensorboard for your run? I think we may log more metrics in the tensorboard than what we print to terminal.

Hello, I read the tensorboard, there are X_groundtruth_acc and Y_groundtruth_acc, but there is no final result acc

aidangomez commented 4 years ago

I think those accuracies are the final accuracies they report whether F and G have found the right encode() and decode() maps

kanwangshijie-sudo commented 4 years ago

I think those accuracies are the final accuracies they report whether F and G have found the right encode() and decode() maps

OK, thank you, but my X_groundtruth_acc and Y_groundtruth_acc are only about 60%, and my data set is vigenere345_brown200，What is going on, the accuracy of your thesis is 75.7%.

aidangomez commented 4 years ago

I would recommend double checking that all the hyperparameters are set as described in paper + appendix and ensure you're using the right timing signal etc.

Mariam-73 commented 4 years ago

Which version of TF are you running?

I want to ask how you output the accuracy rate at that time. I changed the xy_mse of the train.py file to acc

How did you output accuracy?