Configuration

Key	Value	Information
experiment_name	my_experiment	Name of your experiment
nb_epochs	30	Number of epochs for the training of the model
seed	1	Random seed to set the experiments

dir

Directories to set

Key	Value	Information
model_path	"./Models"	Directories to set Path where to save the models
logs_path	"./Logs"	Path where to save the trainings logs

config_active_learner

Key	Value	Information
AL_batch_size	500	Number of samples to query at each iteration
budget	5000	Total number of samples to query to the oracle (Opt.)
targeted_accuracy	1	Minimum accuracy to reach for the active learning process to stop (Opt.)
strategy	EntropySampling	Name of the query strategy to use
strategy_args		Parameters specific to the strategy. We advise to you to check the documentation of their respective library for more information.
batch_size	64

trainer

Trainer configuration, for more information see pytorch-lightning

Key	Value	Information
max_epochs	${nb_epochs}	Trainer configuration, for more information see pytorch-lightning
accelerator	gpu
devices	1
log_every_n_steps	4

test

Key	Value	Information
dataloader	``
batch_size	800
num_workers	0

train

Key	Value	Information
dataloader
batch_size	64
num_workers	0