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: 92.97%	auROC: 0.995	 auPRC: 0.494
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 0.0% | 31.0% | 54.1%	Num Positives: 1415	 Num Negatives: 996701


New best auPRC. Saving model.


Epoch 2:
Metrics across all datasets:
Balanced Accuracy: 92.19%	auROC: 0.994	 auPRC: 0.536
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 0.0% | 39.4% | 59.2%	Num Positives: 1415	 Num Negatives: 996701


New best auPRC. Saving model.


Epoch 3:
Metrics across all datasets:
Balanced Accuracy: 91.04%	auROC: 0.994	 auPRC: 0.539
Recall at 5% | 10% | 25% | 50% FDR: 0.1% | 0.1% | 39.1% | 60.8%	Num Positives: 1415	 Num Negatives: 996701


New best auPRC. Saving model.


Epoch 4:
Metrics across all datasets:
Balanced Accuracy: 88.20%	auROC: 0.989	 auPRC: 0.541
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 0.0% | 40.1% | 60.1%	Num Positives: 1415	 Num Negatives: 996701


New best auPRC. Saving model.


Epoch 5:
Metrics across all datasets:
Balanced Accuracy: 87.00%	auROC: 0.987	 auPRC: 0.525
Recall at 5% | 10% | 25% | 50% FDR: 0.1% | 0.1% | 36.8% | 58.5%	Num Positives: 1415	 Num Negatives: 996701



Epoch 6:
Metrics across all datasets:
Balanced Accuracy: 87.55%	auROC: 0.992	 auPRC: 0.518
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 0.0% | 33.9% | 59.4%	Num Positives: 1415	 Num Negatives: 996701



Epoch 7:
Metrics across all datasets:
Balanced Accuracy: 87.91%	auROC: 0.990	 auPRC: 0.534
Recall at 5% | 10% | 25% | 50% FDR: 0.1% | 0.1% | 37.0% | 60.2%	Num Positives: 1415	 Num Negatives: 996701



Epoch 8:
Metrics across all datasets:
Balanced Accuracy: 87.31%	auROC: 0.992	 auPRC: 0.531
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 0.0% | 38.7% | 59.8%	Num Positives: 1415	 Num Negatives: 996701



Epoch 9:
Metrics across all datasets:
Balanced Accuracy: 90.07%	auROC: 0.991	 auPRC: 0.535
Recall at 5% | 10% | 25% | 50% FDR: 0.0% | 0.0% | 35.9% | 60.5%	Num Positives: 1415	 Num Negatives: 996701


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-H3K27ac-Broad-CTCFL/model.arch.json and /users/marinovg/2018-01-25-histone-mods/tfdragonn-SequenceAndDnaseClassifier-K562-H3K27ac-Broad-CTCFL/model.weights.h5