Simulators, Learners, Evaluators, Comparators

ID	Class Name	Base Class	Type	Part of command	Description
none	Simulator	none	Simulator	seqgra	generates synthetic data
none	Learner	none	Learner	seqgra	trains models on synthetic or experimental data
none	MultiClassClassificationLearner	Learner	Learner	seqgra	learner for multi-class classification tasks
none	MultiLabelClassificationLearner	Learner	Learner	seqgra	learner for multi-label classification tasks
none	DNAMultiClassClassificationLearner	MultiClassClassificationLearner	Learner	seqgra	learner for DNA sequence multi-class classification tasks
none	DNAMultiLabelClassificationLearner	MultiLabelClassificationLearner	Learner	seqgra	learner for DNA sequence multi-label classification tasks
none	KerasDNAMultiClassClassificationLearner	DNAMultiClassClassificationLearner	Learner	seqgra	TensorFlow Keras learner for DNA sequence multi-class classification tasks
none	TorchDNAMultiClassClassificationLearner	DNAMultiClassClassificationLearner	Learner	seqgra	PyTorch learner for DNA sequence multi-class classification tasks
none	BayesOptimalDNAMultiClassClassificationLearner	DNAMultiClassClassificationLearner	Learner	seqgra	BOC learner for DNA sequence multi-class classification tasks
none	KerasDNAMultiLabelClassificationLearner	DNAMultiLabelClassificationLearner	Learner	seqgra	TensorFlow Keras learner for DNA sequence multi-label classification tasks
none	TorchDNAMultiLabelClassificationLearner	DNAMultiLabelClassificationLearner	Learner	seqgra	PyTorch learner for DNA sequence multi-label classification tasks
none	BayesOptimalDNAMultiLabelClassificationLearner	DNAMultiLabelClassificationLearner	Learner	seqgra	BOC learner for DNA sequence multi-label classification tasks
none	Evaluator	none	Evaluator	seqgra	evaluates trained models
none	FeatureImportanceEvaluator	Evaluator	Evaluator	seqgra	evaluates trained models based on input feature importance/attribution/contribution (FI)
none	GradientBasedEvaluator	FeatureImportanceEvaluator	Evaluator	seqgra	evaluates trained models based on input FI, requires access to gradient
roc	ROCEvaluator	Evaluator	Evaluator	seqgra	create ROC curves for trained models
pr	PREvaluator	Evaluator	Evaluator	seqgra	create precision-recall curves for trained models
metrics	MetricsEvaluator	Evaluator	Evaluator	seqgra	saves accuracy and loss of best model to file
prediction	PredictEvaluator	Evaluator	Evaluator	seqgra	saves all example predictions to file
sis	SISEvaluator	FeatureImportanceEvaluator	Evaluator	seqgra	Sufficient Input Subsets (SIS) FI evaluator
gradient	GradientEvaluator	GradientBasedEvaluator	Evaluator	seqgra	raw gradient FI evaluator
saliency	SaliencyEvaluator	GradientBasedEvaluator	Evaluator	seqgra	absolute gradient FI evaluator
gradient-x-input	GradientxInputEvaluator	GradientBasedEvaluator	Evaluator	seqgra	gradient times input FI evaluator
integrated-gradients	IntegratedGradientEvaluator	GradientBasedEvaluator	Evaluator	seqgra	Integrated Gradients FI evaluator
guided-backprop	GuidedBackpropEvaluator	GradientBasedEvaluator	Evaluator	seqgra	Guided Backpropagation FI evaluator
deconv	DeconvEvaluator	GradientBasedEvaluator	Evaluator	seqgra	Deconvolution FI evaluator
none	Comparator	none	Comparator	seqgras	compares properties and performance metrics across data sets, models
table	TableComparator	Comparator	Comparator	seqgras	compiles table of properties from simulation processes, training processes, and evaluators
curve-table	CurveTableComparator	Comparator	Comparator	seqgras	compiles table of ROC and PR curves
fi-eval-table	FIETableComparator	Comparator	Comparator	seqgras	compiles table of metrics from FI evaluators
roc	ROCComparator	Comparator	Comparator	seqgras	creates plot of ROC curves from multiple models
pr	PRComparator	Comparator	Comparator	seqgras	creates plot of precision-recall curves from multiple models