GenomeFlowInterface Settings:
shuffle: True
pos_sampling_rate: 0.05
validation_chroms: ['chr22']
holdout_chroms: ['chr1', 'chr8', 'chr21']
optimizer: adam
learning rate: 0.0003
batch size: 256
epoch size: 2500000
max num of epochs: 100
early stopping metrics: auPRC
early stopping patience: 4

Epoch 1:
Metrics across all datasets:
Balanced Accuracy: 86.54%	auROC: 0.984	 auPRC: 0.484
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 0.0% | 21.5% | 49.9%	Num Positives: 5476	 Num Negatives: 968578


New best auPRC. Saving model.


Epoch 2:
Metrics across all datasets:
Balanced Accuracy: 85.19%	auROC: 0.985	 auPRC: 0.499
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 6.3% | 22.7% | 51.1%	Num Positives: 5476	 Num Negatives: 968578


New best auPRC. Saving model.


Epoch 3:
Metrics across all datasets:
Balanced Accuracy: 88.96%	auROC: 0.984	 auPRC: 0.502
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 3.9% | 22.8% | 52.4%	Num Positives: 5476	 Num Negatives: 968578


New best auPRC. Saving model.


Epoch 4:
Metrics across all datasets:
Balanced Accuracy: 85.59%	auROC: 0.983	 auPRC: 0.507
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 3.5% | 24.3% | 53.3%	Num Positives: 5476	 Num Negatives: 968578


New best auPRC. Saving model.


Epoch 5:
Metrics across all datasets:
Balanced Accuracy: 85.83%	auROC: 0.982	 auPRC: 0.501
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 5.5% | 24.4% | 52.9%	Num Positives: 5476	 Num Negatives: 968578



Epoch 6:
Metrics across all datasets:
Balanced Accuracy: 85.23%	auROC: 0.980	 auPRC: 0.496
Recall at 5% | 10% | 25% | 50% FDR: 0.2% | 5.5% | 23.1% | 51.4%	Num Positives: 5476	 Num Negatives: 968578



Epoch 7:
Metrics across all datasets:
Balanced Accuracy: 85.66%	auROC: 0.979	 auPRC: 0.487
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 0.0% | 23.3% | 50.6%	Num Positives: 5476	 Num Negatives: 968578



Epoch 8:
Metrics across all datasets:
Balanced Accuracy: 84.02%	auROC: 0.977	 auPRC: 0.487
Recall at 5% | 10% | 25% | 50% FDR: 0.9% | 5.5% | 22.5% | 50.1%	Num Positives: 5476	 Num Negatives: 968578



Epoch 9:
Metrics across all datasets:
Balanced Accuracy: 83.96%	auROC: 0.976	 auPRC: 0.485
Recall at 5% | 10% | 25% | 50% FDR: 0.7% | 7.2% | 20.1% | 50.8%	Num Positives: 5476	 Num Negatives: 968578


Finished training after 9 epochs.
The best model's architecture and weights (from epoch 4) were saved to /users/marinovg/2018-01-25-histone-mods/tfdragonn-SequenceAndDnaseClassifier-K562-ATAC-E2F6/model.arch.json and /users/marinovg/2018-01-25-histone-mods/tfdragonn-SequenceAndDnaseClassifier-K562-ATAC-E2F6/model.weights.h5