Update README file

README.md

@@ -6,41 +6,41 @@ tags: []
 ## Original result
 ```
 IoU metric: bbox
- Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.044
- Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.056
- Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.051
- Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.157
- Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.030
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.024
+ Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.032
+ Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.027
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.092
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.070
  Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = -1.000
- Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.070
- Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.202
- Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.466
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.074
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.254
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.464
  Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.389
- Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.557
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.552
  Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = -1.000
 ```
 
 ## After training result
 ```
 IoU metric: bbox
- Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.001
- Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.010
- Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.000
- Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.031
- Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.009
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.452
+ Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.566
+ Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.515
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.461
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.521
  Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = -1.000
- Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.000
- Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.018
- Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.048
- Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.030
- Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.070
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.174
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.636
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.718
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.615
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.839
  Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = -1.000
 ```
 
 ## Config
 - dataset: NIH
 - original model: facebook/detr-resnet-50
-- lr: 5e-05
+- lr: 5e-06
 - dropout_rate: 0.1
 - weight_decay: 0.05
 - max_epochs: 20
@@ -49,27 +49,27 @@ IoU metric: bbox
 ## Logging
 ### Training process
 ```
-{'validation_loss': tensor(2.5684, device='cuda:0'), 'validation_loss_ce': tensor(0.6809, device='cuda:0'), 'validation_loss_bbox': tensor(0.1171, device='cuda:0'), 'validation_loss_giou': tensor(0.6509, device='cuda:0'), 'validation_cardinality_error': tensor(38.2500, device='cuda:0')}
-{'training_loss': tensor(4.4525, device='cuda:0'), 'train_loss_ce': tensor(0.8844, device='cuda:0'), 'train_loss_bbox': tensor(0.3868, device='cuda:0'), 'train_loss_giou': tensor(0.8171, device='cuda:0'), 'train_cardinality_error': tensor(19., device='cuda:0'), 'validation_loss': tensor(1.2852, device='cuda:0'), 'validation_loss_ce': tensor(0.5324, device='cuda:0'), 'validation_loss_bbox': tensor(0.0360, device='cuda:0'), 'validation_loss_giou': tensor(0.2863, device='cuda:0'), 'validation_cardinality_error': tensor(4., device='cuda:0')}
-{'training_loss': tensor(2.3822, device='cuda:0'), 'train_loss_ce': tensor(1.1789, device='cuda:0'), 'train_loss_bbox': tensor(0.1039, device='cuda:0'), 'train_loss_giou': tensor(0.3419, device='cuda:0'), 'train_cardinality_error': tensor(28., device='cuda:0'), 'validation_loss': tensor(1.5449, device='cuda:0'), 'validation_loss_ce': tensor(0.5027, device='cuda:0'), 'validation_loss_bbox': tensor(0.0720, device='cuda:0'), 'validation_loss_giou': tensor(0.3411, device='cuda:0'), 'validation_cardinality_error': tensor(3.5385, device='cuda:0')}
-{'training_loss': tensor(0.6928, device='cuda:0'), 'train_loss_ce': tensor(0.3421, device='cuda:0'), 'train_loss_bbox': tensor(0.0502, device='cuda:0'), 'train_loss_giou': tensor(0.0498, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.7679, device='cuda:0'), 'validation_loss_ce': tensor(0.4698, device='cuda:0'), 'validation_loss_bbox': tensor(0.0951, device='cuda:0'), 'validation_loss_giou': tensor(0.4114, device='cuda:0'), 'validation_cardinality_error': tensor(3.3077, device='cuda:0')}
-{'training_loss': tensor(0.4587, device='cuda:0'), 'train_loss_ce': tensor(0.2548, device='cuda:0'), 'train_loss_bbox': tensor(0.0098, device='cuda:0'), 'train_loss_giou': tensor(0.0775, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.1728, device='cuda:0'), 'validation_loss_ce': tensor(0.4435, device='cuda:0'), 'validation_loss_bbox': tensor(0.0387, device='cuda:0'), 'validation_loss_giou': tensor(0.2678, device='cuda:0'), 'validation_cardinality_error': tensor(5.7692, device='cuda:0')}
-{'training_loss': tensor(0.9540, device='cuda:0'), 'train_loss_ce': tensor(0.4630, device='cuda:0'), 'train_loss_bbox': tensor(0.0313, device='cuda:0'), 'train_loss_giou': tensor(0.1673, device='cuda:0'), 'train_cardinality_error': tensor(19., device='cuda:0'), 'validation_loss': tensor(1.0820, device='cuda:0'), 'validation_loss_ce': tensor(0.4409, device='cuda:0'), 'validation_loss_bbox': tensor(0.0341, device='cuda:0'), 'validation_loss_giou': tensor(0.2353, device='cuda:0'), 'validation_cardinality_error': tensor(7.6154, device='cuda:0')}
-{'training_loss': tensor(0.6391, device='cuda:0'), 'train_loss_ce': tensor(0.3858, device='cuda:0'), 'train_loss_bbox': tensor(0.0238, device='cuda:0'), 'train_loss_giou': tensor(0.0672, device='cuda:0'), 'train_cardinality_error': tensor(5., device='cuda:0'), 'validation_loss': tensor(1.0656, device='cuda:0'), 'validation_loss_ce': tensor(0.4144, device='cuda:0'), 'validation_loss_bbox': tensor(0.0261, device='cuda:0'), 'validation_loss_giou': tensor(0.2604, device='cuda:0'), 'validation_cardinality_error': tensor(10.8462, device='cuda:0')}
-{'training_loss': tensor(1.7558, device='cuda:0'), 'train_loss_ce': tensor(0.3972, device='cuda:0'), 'train_loss_bbox': tensor(0.1433, device='cuda:0'), 'train_loss_giou': tensor(0.3210, device='cuda:0'), 'train_cardinality_error': tensor(5., device='cuda:0'), 'validation_loss': tensor(1.9132, device='cuda:0'), 'validation_loss_ce': tensor(0.4682, device='cuda:0'), 'validation_loss_bbox': tensor(0.0978, device='cuda:0'), 'validation_loss_giou': tensor(0.4779, device='cuda:0'), 'validation_cardinality_error': tensor(12., device='cuda:0')}
-{'training_loss': tensor(3.0507, device='cuda:0'), 'train_loss_ce': tensor(0.7805, device='cuda:0'), 'train_loss_bbox': tensor(0.1745, device='cuda:0'), 'train_loss_giou': tensor(0.6989, device='cuda:0'), 'train_cardinality_error': tensor(16., device='cuda:0'), 'validation_loss': tensor(2.6096, device='cuda:0'), 'validation_loss_ce': tensor(0.4846, device='cuda:0'), 'validation_loss_bbox': tensor(0.1573, device='cuda:0'), 'validation_loss_giou': tensor(0.6693, device='cuda:0'), 'validation_cardinality_error': tensor(5.8462, device='cuda:0')}
-{'training_loss': tensor(1.7916, device='cuda:0'), 'train_loss_ce': tensor(0.4071, device='cuda:0'), 'train_loss_bbox': tensor(0.1937, device='cuda:0'), 'train_loss_giou': tensor(0.2079, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(3.2897, device='cuda:0'), 'validation_loss_ce': tensor(0.5079, device='cuda:0'), 'validation_loss_bbox': tensor(0.2703, device='cuda:0'), 'validation_loss_giou': tensor(0.7151, device='cuda:0'), 'validation_cardinality_error': tensor(3.2308, device='cuda:0')}
-{'training_loss': tensor(3.2579, device='cuda:0'), 'train_loss_ce': tensor(0.7360, device='cuda:0'), 'train_loss_bbox': tensor(0.2364, device='cuda:0'), 'train_loss_giou': tensor(0.6700, device='cuda:0'), 'train_cardinality_error': tensor(4., device='cuda:0'), 'validation_loss': tensor(2.0824, device='cuda:0'), 'validation_loss_ce': tensor(0.5116, device='cuda:0'), 'validation_loss_bbox': tensor(0.1048, device='cuda:0'), 'validation_loss_giou': tensor(0.5234, device='cuda:0'), 'validation_cardinality_error': tensor(6.9231, device='cuda:0')}
-{'training_loss': tensor(3.7396, device='cuda:0'), 'train_loss_ce': tensor(0.4191, device='cuda:0'), 'train_loss_bbox': tensor(0.3937, device='cuda:0'), 'train_loss_giou': tensor(0.6760, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(3.1265, device='cuda:0'), 'validation_loss_ce': tensor(0.5484, device='cuda:0'), 'validation_loss_bbox': tensor(0.2265, device='cuda:0'), 'validation_loss_giou': tensor(0.7228, device='cuda:0'), 'validation_cardinality_error': tensor(4.3077, device='cuda:0')}
-{'training_loss': tensor(2.6312, device='cuda:0'), 'train_loss_ce': tensor(0.3234, device='cuda:0'), 'train_loss_bbox': tensor(0.2922, device='cuda:0'), 'train_loss_giou': tensor(0.4235, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(3.3476, device='cuda:0'), 'validation_loss_ce': tensor(0.5086, device='cuda:0'), 'validation_loss_bbox': tensor(0.2609, device='cuda:0'), 'validation_loss_giou': tensor(0.7671, device='cuda:0'), 'validation_cardinality_error': tensor(3.8462, device='cuda:0')}
-{'training_loss': tensor(1.7406, device='cuda:0'), 'train_loss_ce': tensor(0.5410, device='cuda:0'), 'train_loss_bbox': tensor(0.0662, device='cuda:0'), 'train_loss_giou': tensor(0.4344, device='cuda:0'), 'train_cardinality_error': tensor(3., device='cuda:0'), 'validation_loss': tensor(3.2995, device='cuda:0'), 'validation_loss_ce': tensor(0.5209, device='cuda:0'), 'validation_loss_bbox': tensor(0.2275, device='cuda:0'), 'validation_loss_giou': tensor(0.8205, device='cuda:0'), 'validation_cardinality_error': tensor(6.3846, device='cuda:0')}
-{'training_loss': tensor(2.6756, device='cuda:0'), 'train_loss_ce': tensor(0.4669, device='cuda:0'), 'train_loss_bbox': tensor(0.2744, device='cuda:0'), 'train_loss_giou': tensor(0.4184, device='cuda:0'), 'train_cardinality_error': tensor(2., device='cuda:0'), 'validation_loss': tensor(3.9060, device='cuda:0'), 'validation_loss_ce': tensor(0.5255, device='cuda:0'), 'validation_loss_bbox': tensor(0.3109, device='cuda:0'), 'validation_loss_giou': tensor(0.9129, device='cuda:0'), 'validation_cardinality_error': tensor(7.3846, device='cuda:0')}
-{'training_loss': tensor(2.6693, device='cuda:0'), 'train_loss_ce': tensor(0.5077, device='cuda:0'), 'train_loss_bbox': tensor(0.1733, device='cuda:0'), 'train_loss_giou': tensor(0.6475, device='cuda:0'), 'train_cardinality_error': tensor(2., device='cuda:0'), 'validation_loss': tensor(3.3320, device='cuda:0'), 'validation_loss_ce': tensor(0.5506, device='cuda:0'), 'validation_loss_bbox': tensor(0.2695, device='cuda:0'), 'validation_loss_giou': tensor(0.7171, device='cuda:0'), 'validation_cardinality_error': tensor(6., device='cuda:0')}
-{'training_loss': tensor(3.8867, device='cuda:0'), 'train_loss_ce': tensor(0.6607, device='cuda:0'), 'train_loss_bbox': tensor(0.3213, device='cuda:0'), 'train_loss_giou': tensor(0.8097, device='cuda:0'), 'train_cardinality_error': tensor(4., device='cuda:0'), 'validation_loss': tensor(3.0807, device='cuda:0'), 'validation_loss_ce': tensor(0.5218, device='cuda:0'), 'validation_loss_bbox': tensor(0.2218, device='cuda:0'), 'validation_loss_giou': tensor(0.7249, device='cuda:0'), 'validation_cardinality_error': tensor(3.8462, device='cuda:0')}
-{'training_loss': tensor(2.8666, device='cuda:0'), 'train_loss_ce': tensor(0.4696, device='cuda:0'), 'train_loss_bbox': tensor(0.2954, device='cuda:0'), 'train_loss_giou': tensor(0.4601, device='cuda:0'), 'train_cardinality_error': tensor(2., device='cuda:0'), 'validation_loss': tensor(3.0620, device='cuda:0'), 'validation_loss_ce': tensor(0.5262, device='cuda:0'), 'validation_loss_bbox': tensor(0.2149, device='cuda:0'), 'validation_loss_giou': tensor(0.7306, device='cuda:0'), 'validation_cardinality_error': tensor(3.8462, device='cuda:0')}
-{'training_loss': tensor(2.1570, device='cuda:0'), 'train_loss_ce': tensor(0.6110, device='cuda:0'), 'train_loss_bbox': tensor(0.0905, device='cuda:0'), 'train_loss_giou': tensor(0.5468, device='cuda:0'), 'train_cardinality_error': tensor(7., device='cuda:0'), 'validation_loss': tensor(2.8941, device='cuda:0'), 'validation_loss_ce': tensor(0.5307, device='cuda:0'), 'validation_loss_bbox': tensor(0.2003, device='cuda:0'), 'validation_loss_giou': tensor(0.6809, device='cuda:0'), 'validation_cardinality_error': tensor(3.8462, device='cuda:0')}
-{'training_loss': tensor(4.1725, device='cuda:0'), 'train_loss_ce': tensor(0.9266, device='cuda:0'), 'train_loss_bbox': tensor(0.2818, device='cuda:0'), 'train_loss_giou': tensor(0.9185, device='cuda:0'), 'train_cardinality_error': tensor(11., device='cuda:0'), 'validation_loss': tensor(3.3124, device='cuda:0'), 'validation_loss_ce': tensor(0.5442, device='cuda:0'), 'validation_loss_bbox': tensor(0.2570, device='cuda:0'), 'validation_loss_giou': tensor(0.7416, device='cuda:0'), 'validation_cardinality_error': tensor(3.8462, device='cuda:0')}
-{'training_loss': tensor(1.4272, device='cuda:0'), 'train_loss_ce': tensor(0.3878, device='cuda:0'), 'train_loss_bbox': tensor(0.0514, device='cuda:0'), 'train_loss_giou': tensor(0.3911, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(2.9401, device='cuda:0'), 'validation_loss_ce': tensor(0.5416, device='cuda:0'), 'validation_loss_bbox': tensor(0.2045, device='cuda:0'), 'validation_loss_giou': tensor(0.6879, device='cuda:0'), 'validation_cardinality_error': tensor(3.8462, device='cuda:0')}
+{'validation_loss': tensor(2.7805, device='cuda:0'), 'validation_loss_ce': tensor(0.7506, device='cuda:0'), 'validation_loss_bbox': tensor(0.1834, device='cuda:0'), 'validation_loss_giou': tensor(0.5565, device='cuda:0'), 'validation_cardinality_error': tensor(65.2500, device='cuda:0')}
+{'training_loss': tensor(1.3692, device='cuda:0'), 'train_loss_ce': tensor(0.6425, device='cuda:0'), 'train_loss_bbox': tensor(0.0420, device='cuda:0'), 'train_loss_giou': tensor(0.2583, device='cuda:0'), 'train_cardinality_error': tensor(49., device='cuda:0'), 'validation_loss': tensor(1.9955, device='cuda:0'), 'validation_loss_ce': tensor(0.6964, device='cuda:0'), 'validation_loss_bbox': tensor(0.0949, device='cuda:0'), 'validation_loss_giou': tensor(0.4123, device='cuda:0'), 'validation_cardinality_error': tensor(54.8462, device='cuda:0')}
+{'training_loss': tensor(0.7776, device='cuda:0'), 'train_loss_ce': tensor(0.4950, device='cuda:0'), 'train_loss_bbox': tensor(0.0243, device='cuda:0'), 'train_loss_giou': tensor(0.0805, device='cuda:0'), 'train_cardinality_error': tensor(8., device='cuda:0'), 'validation_loss': tensor(1.6961, device='cuda:0'), 'validation_loss_ce': tensor(0.6763, device='cuda:0'), 'validation_loss_bbox': tensor(0.0584, device='cuda:0'), 'validation_loss_giou': tensor(0.3638, device='cuda:0'), 'validation_cardinality_error': tensor(48.2308, device='cuda:0')}
+{'training_loss': tensor(0.7145, device='cuda:0'), 'train_loss_ce': tensor(0.4630, device='cuda:0'), 'train_loss_bbox': tensor(0.0209, device='cuda:0'), 'train_loss_giou': tensor(0.0735, device='cuda:0'), 'train_cardinality_error': tensor(2., device='cuda:0'), 'validation_loss': tensor(1.5552, device='cuda:0'), 'validation_loss_ce': tensor(0.6569, device='cuda:0'), 'validation_loss_bbox': tensor(0.0531, device='cuda:0'), 'validation_loss_giou': tensor(0.3164, device='cuda:0'), 'validation_cardinality_error': tensor(39.6154, device='cuda:0')}
+{'training_loss': tensor(0.7801, device='cuda:0'), 'train_loss_ce': tensor(0.6306, device='cuda:0'), 'train_loss_bbox': tensor(0.0154, device='cuda:0'), 'train_loss_giou': tensor(0.0363, device='cuda:0'), 'train_cardinality_error': tensor(41., device='cuda:0'), 'validation_loss': tensor(1.4825, device='cuda:0'), 'validation_loss_ce': tensor(0.6332, device='cuda:0'), 'validation_loss_bbox': tensor(0.0563, device='cuda:0'), 'validation_loss_giou': tensor(0.2839, device='cuda:0'), 'validation_cardinality_error': tensor(31.3846, device='cuda:0')}
+{'training_loss': tensor(1.1043, device='cuda:0'), 'train_loss_ce': tensor(0.6572, device='cuda:0'), 'train_loss_bbox': tensor(0.0306, device='cuda:0'), 'train_loss_giou': tensor(0.1471, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.4447, device='cuda:0'), 'validation_loss_ce': tensor(0.6126, device='cuda:0'), 'validation_loss_bbox': tensor(0.0545, device='cuda:0'), 'validation_loss_giou': tensor(0.2798, device='cuda:0'), 'validation_cardinality_error': tensor(25., device='cuda:0')}
+{'training_loss': tensor(0.7144, device='cuda:0'), 'train_loss_ce': tensor(0.5595, device='cuda:0'), 'train_loss_bbox': tensor(0.0076, device='cuda:0'), 'train_loss_giou': tensor(0.0585, device='cuda:0'), 'train_cardinality_error': tensor(10., device='cuda:0'), 'validation_loss': tensor(1.6468, device='cuda:0'), 'validation_loss_ce': tensor(0.5955, device='cuda:0'), 'validation_loss_bbox': tensor(0.0691, device='cuda:0'), 'validation_loss_giou': tensor(0.3530, device='cuda:0'), 'validation_cardinality_error': tensor(20.7692, device='cuda:0')}
+{'training_loss': tensor(0.7055, device='cuda:0'), 'train_loss_ce': tensor(0.5208, device='cuda:0'), 'train_loss_bbox': tensor(0.0134, device='cuda:0'), 'train_loss_giou': tensor(0.0587, device='cuda:0'), 'train_cardinality_error': tensor(4., device='cuda:0'), 'validation_loss': tensor(1.4338, device='cuda:0'), 'validation_loss_ce': tensor(0.5776, device='cuda:0'), 'validation_loss_bbox': tensor(0.0511, device='cuda:0'), 'validation_loss_giou': tensor(0.3002, device='cuda:0'), 'validation_cardinality_error': tensor(15.6154, device='cuda:0')}
+{'training_loss': tensor(0.6106, device='cuda:0'), 'train_loss_ce': tensor(0.3884, device='cuda:0'), 'train_loss_bbox': tensor(0.0299, device='cuda:0'), 'train_loss_giou': tensor(0.0363, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.5870, device='cuda:0'), 'validation_loss_ce': tensor(0.5597, device='cuda:0'), 'validation_loss_bbox': tensor(0.0673, device='cuda:0'), 'validation_loss_giou': tensor(0.3454, device='cuda:0'), 'validation_cardinality_error': tensor(11.7692, device='cuda:0')}
+{'training_loss': tensor(0.7294, device='cuda:0'), 'train_loss_ce': tensor(0.5093, device='cuda:0'), 'train_loss_bbox': tensor(0.0122, device='cuda:0'), 'train_loss_giou': tensor(0.0795, device='cuda:0'), 'train_cardinality_error': tensor(2., device='cuda:0'), 'validation_loss': tensor(1.3645, device='cuda:0'), 'validation_loss_ce': tensor(0.5496, device='cuda:0'), 'validation_loss_bbox': tensor(0.0514, device='cuda:0'), 'validation_loss_giou': tensor(0.2789, device='cuda:0'), 'validation_cardinality_error': tensor(8.9231, device='cuda:0')}
+{'training_loss': tensor(0.7845, device='cuda:0'), 'train_loss_ce': tensor(0.5872, device='cuda:0'), 'train_loss_bbox': tensor(0.0111, device='cuda:0'), 'train_loss_giou': tensor(0.0710, device='cuda:0'), 'train_cardinality_error': tensor(4., device='cuda:0'), 'validation_loss': tensor(1.5335, device='cuda:0'), 'validation_loss_ce': tensor(0.5374, device='cuda:0'), 'validation_loss_bbox': tensor(0.0669, device='cuda:0'), 'validation_loss_giou': tensor(0.3307, device='cuda:0'), 'validation_cardinality_error': tensor(6.6154, device='cuda:0')}
+{'training_loss': tensor(0.3810, device='cuda:0'), 'train_loss_ce': tensor(0.3312, device='cuda:0'), 'train_loss_bbox': tensor(0.0028, device='cuda:0'), 'train_loss_giou': tensor(0.0179, device='cuda:0'), 'train_cardinality_error': tensor(0., device='cuda:0'), 'validation_loss': tensor(1.4720, device='cuda:0'), 'validation_loss_ce': tensor(0.5209, device='cuda:0'), 'validation_loss_bbox': tensor(0.0572, device='cuda:0'), 'validation_loss_giou': tensor(0.3326, device='cuda:0'), 'validation_cardinality_error': tensor(5.7692, device='cuda:0')}
+{'training_loss': tensor(0.9936, device='cuda:0'), 'train_loss_ce': tensor(0.7054, device='cuda:0'), 'train_loss_bbox': tensor(0.0178, device='cuda:0'), 'train_loss_giou': tensor(0.0996, device='cuda:0'), 'train_cardinality_error': tensor(13., device='cuda:0'), 'validation_loss': tensor(1.4543, device='cuda:0'), 'validation_loss_ce': tensor(0.5103, device='cuda:0'), 'validation_loss_bbox': tensor(0.0641, device='cuda:0'), 'validation_loss_giou': tensor(0.3117, device='cuda:0'), 'validation_cardinality_error': tensor(4.7692, device='cuda:0')}
+{'training_loss': tensor(0.5030, device='cuda:0'), 'train_loss_ce': tensor(0.3787, device='cuda:0'), 'train_loss_bbox': tensor(0.0170, device='cuda:0'), 'train_loss_giou': tensor(0.0196, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.6086, device='cuda:0'), 'validation_loss_ce': tensor(0.5086, device='cuda:0'), 'validation_loss_bbox': tensor(0.0739, device='cuda:0'), 'validation_loss_giou': tensor(0.3652, device='cuda:0'), 'validation_cardinality_error': tensor(4.3077, device='cuda:0')}
+{'training_loss': tensor(0.6252, device='cuda:0'), 'train_loss_ce': tensor(0.3750, device='cuda:0'), 'train_loss_bbox': tensor(0.0278, device='cuda:0'), 'train_loss_giou': tensor(0.0556, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.0120, device='cuda:0'), 'validation_loss_ce': tensor(0.4808, device='cuda:0'), 'validation_loss_bbox': tensor(0.0253, device='cuda:0'), 'validation_loss_giou': tensor(0.2022, device='cuda:0'), 'validation_cardinality_error': tensor(3.3077, device='cuda:0')}
+{'training_loss': tensor(1.2132, device='cuda:0'), 'train_loss_ce': tensor(0.7323, device='cuda:0'), 'train_loss_bbox': tensor(0.0422, device='cuda:0'), 'train_loss_giou': tensor(0.1350, device='cuda:0'), 'train_cardinality_error': tensor(11., device='cuda:0'), 'validation_loss': tensor(0.8473, device='cuda:0'), 'validation_loss_ce': tensor(0.4638, device='cuda:0'), 'validation_loss_bbox': tensor(0.0177, device='cuda:0'), 'validation_loss_giou': tensor(0.1475, device='cuda:0'), 'validation_cardinality_error': tensor(2.6923, device='cuda:0')}
+{'training_loss': tensor(0.4347, device='cuda:0'), 'train_loss_ce': tensor(0.3896, device='cuda:0'), 'train_loss_bbox': tensor(0.0037, device='cuda:0'), 'train_loss_giou': tensor(0.0132, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(0.8593, device='cuda:0'), 'validation_loss_ce': tensor(0.4498, device='cuda:0'), 'validation_loss_bbox': tensor(0.0195, device='cuda:0'), 'validation_loss_giou': tensor(0.1560, device='cuda:0'), 'validation_cardinality_error': tensor(2.6154, device='cuda:0')}
+{'training_loss': tensor(0.4821, device='cuda:0'), 'train_loss_ce': tensor(0.4537, device='cuda:0'), 'train_loss_bbox': tensor(0.0014, device='cuda:0'), 'train_loss_giou': tensor(0.0106, device='cuda:0'), 'train_cardinality_error': tensor(2., device='cuda:0'), 'validation_loss': tensor(0.9156, device='cuda:0'), 'validation_loss_ce': tensor(0.4412, device='cuda:0'), 'validation_loss_bbox': tensor(0.0223, device='cuda:0'), 'validation_loss_giou': tensor(0.1814, device='cuda:0'), 'validation_cardinality_error': tensor(2.6154, device='cuda:0')}
+{'training_loss': tensor(0.5579, device='cuda:0'), 'train_loss_ce': tensor(0.3216, device='cuda:0'), 'train_loss_bbox': tensor(0.0260, device='cuda:0'), 'train_loss_giou': tensor(0.0531, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(0.8854, device='cuda:0'), 'validation_loss_ce': tensor(0.4308, device='cuda:0'), 'validation_loss_bbox': tensor(0.0209, device='cuda:0'), 'validation_loss_giou': tensor(0.1750, device='cuda:0'), 'validation_cardinality_error': tensor(2.6154, device='cuda:0')}
+{'training_loss': tensor(0.3421, device='cuda:0'), 'train_loss_ce': tensor(0.3072, device='cuda:0'), 'train_loss_bbox': tensor(0.0029, device='cuda:0'), 'train_loss_giou': tensor(0.0103, device='cuda:0'), 'train_cardinality_error': tensor(0., device='cuda:0'), 'validation_loss': tensor(0.8500, device='cuda:0'), 'validation_loss_ce': tensor(0.4143, device='cuda:0'), 'validation_loss_bbox': tensor(0.0191, device='cuda:0'), 'validation_loss_giou': tensor(0.1700, device='cuda:0'), 'validation_cardinality_error': tensor(2.3846, device='cuda:0')}
+{'training_loss': tensor(0.4960, device='cuda:0'), 'train_loss_ce': tensor(0.3847, device='cuda:0'), 'train_loss_bbox': tensor(0.0085, device='cuda:0'), 'train_loss_giou': tensor(0.0343, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(0.8550, device='cuda:0'), 'validation_loss_ce': tensor(0.4080, device='cuda:0'), 'validation_loss_bbox': tensor(0.0216, device='cuda:0'), 'validation_loss_giou': tensor(0.1695, device='cuda:0'), 'validation_cardinality_error': tensor(2.3077, device='cuda:0')}
 ```
 
 ## Examples