The Cerebros package is an ultra-precise Neural Architecture Search (NAS) / AutoML that is intended to much more closely mimic biological neurons than conventional neural network architecture strategies.
Other
27
stars
4
forks
source link
Added an image classifciation task template for EfficientNetV2L embed… #90
Performance:
val-top-1-categorical-accuracy: 0.91657
Optima:
activation: relu predecessor_level_connection_affinity_factor_first: 1.19289 predecessor_level_connection_affinity_factor_main: 46.039 max_consecutive_lateral_connections: 11 p_lateral_connection: 24.7868 num_lateral_connection_tries_per_unit: 12 learning_rate: 2.30375e-5 epochs: 5 batch_size: 21 maximum_levels: 7 maximum_units_per_level: 10 maximum_neurons_per_unit: 5