utils package¶
Submodules¶
utils.callbacks module¶
-
utils.callbacks.callbacks(save_path: str, depth: int) → List¶ Keras callbacks which include ModelCheckpoint, CSVLogger, TensorBoard, LearningRateScheduler, TerminateOnNaN
- save_path: str
local directory to save model weights
- depth: int
Depth of ResNet model
- Returns
List all callbacks which include ModelCheckpoint, EarlyStopping, CSVLogger, ReduceLROnPlateau, LearningRateScheduler, TerminateOnNaN
utils.misc_utils module¶
-
utils.misc_utils.limit_gpu()¶ - Limit use of GPU by implementing Tensorflow's Memory Graph. Useful in local system where GPU is also used as video card.
-
utils.misc_utils.report(y_true, y_pred, labels: List = None)¶ Logging of report
- y_true: numpy array or pandas series
lables of test data
- y_pred: numpy array or pandas series
labels of predicted data
- labels: list, optional
List of classes’ labels
-
utils.misc_utils.visualize(history: Dict, save_plot: bool = True, save_dir: str = None)¶ Visualize training history of model
- history: Dict
model.fit history
- save_plot: bool, optional
save plot to hard disk, by default True
- save_dir: str, optional
path to save plot, by default None