Closed beckstev closed 5 years ago
If this difference https://github.com/beckstev/MachineLearningSeminar/blob/19679ec84fd7443e54320935bba6d7a51196b060/dog_classifier/evaluate/evaluate_training.py#L412 is equal to zero the array Y_test will be initialiize as an empty array: https://github.com/beckstev/MachineLearningSeminar/blob/19679ec84fd7443e54320935bba6d7a51196b060/dog_classifier/evaluate/evaluate_training.py#L417
Y_test
I added a if-condition to catch this bug. Furthermore, I used the dataframe saved inside the dataloader to get the race labels. The reason for this is that the datloader converts the initial race labels automatically into a hot vector:
https://github.com/beckstev/MachineLearningSeminar/blob/19679ec84fd7443e54320935bba6d7a51196b060/dog_classifier/net/dataloader.py#L58
I hope you can verify my changes.
If this difference https://github.com/beckstev/MachineLearningSeminar/blob/19679ec84fd7443e54320935bba6d7a51196b060/dog_classifier/evaluate/evaluate_training.py#L412 is equal to zero the array
Y_testwill be initialiize as an empty array: https://github.com/beckstev/MachineLearningSeminar/blob/19679ec84fd7443e54320935bba6d7a51196b060/dog_classifier/evaluate/evaluate_training.py#L417I added a if-condition to catch this bug. Furthermore, I used the dataframe saved inside the dataloader to get the race labels. The reason for this is that the datloader converts the initial race labels automatically into a hot vector:
https://github.com/beckstev/MachineLearningSeminar/blob/19679ec84fd7443e54320935bba6d7a51196b060/dog_classifier/net/dataloader.py#L58
I hope you can verify my changes.