Related issue: #1366
EarlyStopping class uses mAP by default, but mAP only is computet if evaluation dataset is provided. If one tries to train the model without any validation dataset, a warning appears:
RuntimeWarning: Early stopping conditioned on metric mAP which is not available. Available metrics are: loss,regression_loss,classification_loss,lr (self.monitor, ','.join(list(logs.keys()))), RuntimeWarning
hgaiser
commented
4 years ago
Related issue: #1366 EarlyStopping class uses mAP by default, but mAP only is computet if evaluation dataset is provided. If one tries to train the model without any validation dataset, a warning appears:
RuntimeWarning: Early stopping conditioned on metric mAP which is not available. Available metrics are: loss,regression_loss,classification_loss,lr (self.monitor, ','.join(list(logs.keys()))), RuntimeWarning