Analysis¶
GQCNNAnalyzer¶
A tool for analyzing trained GQ-CNNs. Calculates statistics such as training/valiation errors and losses. Also plots Precision-Recall Curve and ROC, and saves sample TP/TN/FP/FN training/validation examples.
-
class
gqcnn.
GQCNNAnalyzer
(config, verbose=True, plot_backend='pdf')¶ Bases:
object
Analyzes a trained GQ-CNN model.
-
__init__
(config, verbose=True, plot_backend='pdf')¶ - Parameters
config (dict) – Dictionary of analysis configuration parameters.
verbose (bool) – Whether or not to log analysis output to stdout.
plot_backend (str) – Matplotlib plotting backend to use, default is non-interactive “pdf” backend.
-
analyze
(model_dir, output_dir, dataset_config=None)¶ Run analysis.
- Parameters
model_dir (str) – Path to the GQ-CNN model to analyze.
output_dir (str) – Path to save the analysis.
dataset_config (dict) – Dictionary to configure dataset used for training evaluation if different from one used during training.
- Returns
autolab_core.BinaryClassificationResult
– Result of analysis on training data.autolab_core.BinaryClassificationResult
– Result of analysis on validation data.
-