================================================================================
Model Configuration (CRITICAL FOR INFERENCE)
================================================================================
modelCategory: sota
Task: detection
Genome: [[[3, 0], [6, 1]], [[6, 2], [0, 2]], [[2, 1], [0, 0]], [[0, 4], [8, 0]], [[0, 1], [4, 5]]]
Architecture: AdapterGenotype(ops=[('sep_conv_3x3', 0), ('avg_pool_3x3', 1), ('avg_pool_3x3', 2), ('none', 2), ('conv_1x1', 1), ('none', 0), ('none', 4), ('channel_attention', 0), ('none', 1), ('sep_conv_5x5', 5)], ops_concat=[3, 6])
backbone: tiny_darknet
search_space: micro
num_cells: 3
num_classes: 2
image_size: 640
class_names: ['person', 'safety_shoe']
================================================================================
Performance Metrics
================================================================================
Params: 2.13232 MB
Forward FLOPs: 467.960064 M
Training FLOPs: 1622.861056 M [thop]
--------------------------------------------------------------------------------
Epoch  | Train Loss | Valid mAP  | Error      | FLOPs     
--------------------------------------------------------------------------------
0      | 10.5168    | 0.0000     | 100.00     | 1622.86   
