Metric name is hardcoded in scan_object.evaluate_models()

autonomio / talos

Hyperparameter Experiments with TensorFlow and Keras

https://autonom.io

MIT License

1.62k stars 269 forks source link

Metric name is hardcoded in scan_object.evaluate_models() #385

Closed beeb closed 4 years ago

beeb commented 4 years ago

As seen on this line and the next, the evaluate_models method adds the columns eval_f1score_mean and eval_f1score_std in all cases, even when the used metric is different from f1score (e.g. for "continuous" tasks)

mikkokotila commented 4 years ago

Somehow I totally missed this. Thanks. Will try to sort it out today.