================================================================================
Model Configuration (CRITICAL FOR INFERENCE)
================================================================================
modelCategory: SOTA
Task: detection
Genome: [[[5, 1], [4, 0]], [[9, 2], [9, 0]], [[9, 3], [6, 2]], [[3, 4], [4, 1]], [[7, 2], [4, 4]]]
Architecture: AdapterGenotype(ops=[('dil_conv_3x3', 1), ('sep_conv_5x5', 0), ('spatial_attention', 2), ('spatial_attention', 0), ('spatial_attention', 3), ('avg_pool_3x3', 2), ('sep_conv_3x3', 4), ('sep_conv_5x5', 1), ('max_pool_3x3', 2), ('sep_conv_5x5', 4)], ops_concat=[5, 6])
backbone: yolov8n
search_space: micro
num_cells: 3
num_classes: 6
image_size: 640
class_names: ['Sorter1', 'Sorter2', 'Interlayer', 'Gripper', 'Item', 'Robot Operating Area']
================================================================================
Performance Metrics
================================================================================
Params: 5.79223 MB
Forward FLOPs: 3423.589152 M
Training FLOPs: 10892.844960 M [thop]
--------------------------------------------------------------------------------
Epoch  | Train Loss | Valid mAP  | FLOPs     
--------------------------------------------------------------------------------
0      | 4.7419     | 56.4106    | 10892.84  
