models_mgt Namespace Reference

Functions
def	ai_runner_invoke (image_processed, ai_runner_interpreter)
def	_get_zoo_model (DictConfig cfg)
tuple	load_model_for_training (DictConfig cfg)

Function Documentation

_get_zoo_model()

def models_mgt._get_zoo_model ( DictConfig  cfg )

private

Returns a Keras model object based on the specified configuration and parameters.

Args:
    cfg (DictConfig): A dictionary containing the configuration for the model.
    num_classes (int): The number of classes for the model.
    dropout (float): The dropout rate for the model.
    section (str): The section of the model to be used.

Returns:
    tf.keras.Model: A Keras model object based on the specified configuration and parameters.

Definition at line 53 of file models_mgt.py.

Referenced by load_model_for_training().

Here is the caller graph for this function:

ai_runner_invoke()

def models_mgt.ai_runner_invoke	(		image_processed,
			ai_runner_interpreter
	)

Definition at line 36 of file models_mgt.py.

Referenced by evaluate.evaluate().

Here is the caller graph for this function:

load_model_for_training()

tuple models_mgt.load_model_for_training

(

DictConfig

cfg

)

"
Loads a model for training.

The model to train can be:
- a model from the Model Zoo
- a user model (BYOM)
- a model previously trained during a training that was interrupted.

When a training is run, the following files are saved in the saved_models
directory:
    base_model.h5:
        Model saved before the training started. Weights are random.
    best_weights.h5:
        Best weights obtained since the beginning of the training.
    last_weights.h5:
        Weights saved at the end of the last epoch.

To resume a training, the last weights are loaded into the base model.

Definition at line 97 of file models_mgt.py.

References _get_zoo_model().

Here is the call graph for this function: