---
library_name: transformers
tags: []
---

## Original result
```
IoU metric: bbox
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.000
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.000
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.000
 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 ] = 0.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.004
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.004
```

## After training result
```
IoU metric: bbox
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.040
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.071
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.047
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.000
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.044
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.131
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.198
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.227
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = -1.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.018
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.247
```

## Config
- dataset: NIH
- original model: hustvl/yolos-tiny
- lr: 0.0001
- dropout_rate: 0.1
- weight_decay: 0.1
- max_epochs: 30
- train samples: 885

## Logging
### Training process
```
{'validation_loss': tensor(7.0933, device='cuda:0'), 'validation_loss_ce': tensor(2.6325, device='cuda:0'), 'validation_loss_bbox': tensor(0.5049, device='cuda:0'), 'validation_loss_giou': tensor(0.9681, device='cuda:0'), 'validation_cardinality_error': tensor(99., device='cuda:0')}
{'training_loss': tensor(2.2101, device='cuda:0'), 'train_loss_ce': tensor(0.4669, device='cuda:0'), 'train_loss_bbox': tensor(0.1601, device='cuda:0'), 'train_loss_giou': tensor(0.4713, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(2.4272, device='cuda:0'), 'validation_loss_ce': tensor(0.4454, device='cuda:0'), 'validation_loss_bbox': tensor(0.1814, device='cuda:0'), 'validation_loss_giou': tensor(0.5375, device='cuda:0'), 'validation_cardinality_error': tensor(1., device='cuda:0')}
{'training_loss': tensor(1.7165, device='cuda:0'), 'train_loss_ce': tensor(0.4522, device='cuda:0'), 'train_loss_bbox': tensor(0.1315, device='cuda:0'), 'train_loss_giou': tensor(0.3035, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(2.2316, device='cuda:0'), 'validation_loss_ce': tensor(0.4329, device='cuda:0'), 'validation_loss_bbox': tensor(0.1506, device='cuda:0'), 'validation_loss_giou': tensor(0.5228, device='cuda:0'), 'validation_cardinality_error': tensor(1., device='cuda:0')}
{'training_loss': tensor(2.4262, device='cuda:0'), 'train_loss_ce': tensor(0.4559, device='cuda:0'), 'train_loss_bbox': tensor(0.1564, device='cuda:0'), 'train_loss_giou': tensor(0.5942, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(2.1372, device='cuda:0'), 'validation_loss_ce': tensor(0.4360, device='cuda:0'), 'validation_loss_bbox': tensor(0.1478, device='cuda:0'), 'validation_loss_giou': tensor(0.4811, device='cuda:0'), 'validation_cardinality_error': tensor(1., device='cuda:0')}
{'training_loss': tensor(2.1364, device='cuda:0'), 'train_loss_ce': tensor(0.4127, device='cuda:0'), 'train_loss_bbox': tensor(0.1297, device='cuda:0'), 'train_loss_giou': tensor(0.5375, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(2.1923, device='cuda:0'), 'validation_loss_ce': tensor(0.4127, device='cuda:0'), 'validation_loss_bbox': tensor(0.1510, device='cuda:0'), 'validation_loss_giou': tensor(0.5122, device='cuda:0'), 'validation_cardinality_error': tensor(1., device='cuda:0')}
{'training_loss': tensor(2.1083, device='cuda:0'), 'train_loss_ce': tensor(0.3250, device='cuda:0'), 'train_loss_bbox': tensor(0.1613, device='cuda:0'), 'train_loss_giou': tensor(0.4884, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(2.0341, device='cuda:0'), 'validation_loss_ce': tensor(0.4116, device='cuda:0'), 'validation_loss_bbox': tensor(0.1375, device='cuda:0'), 'validation_loss_giou': tensor(0.4674, device='cuda:0'), 'validation_cardinality_error': tensor(1., device='cuda:0')}
{'training_loss': tensor(2.1673, device='cuda:0'), 'train_loss_ce': tensor(0.4102, device='cuda:0'), 'train_loss_bbox': tensor(0.1334, device='cuda:0'), 'train_loss_giou': tensor(0.5452, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(2.0928, device='cuda:0'), 'validation_loss_ce': tensor(0.3856, device='cuda:0'), 'validation_loss_bbox': tensor(0.1463, device='cuda:0'), 'validation_loss_giou': tensor(0.4878, device='cuda:0'), 'validation_cardinality_error': tensor(1., device='cuda:0')}
{'training_loss': tensor(1.7178, device='cuda:0'), 'train_loss_ce': tensor(0.4061, device='cuda:0'), 'train_loss_bbox': tensor(0.0982, device='cuda:0'), 'train_loss_giou': tensor(0.4103, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(2.1328, device='cuda:0'), 'validation_loss_ce': tensor(0.3973, device='cuda:0'), 'validation_loss_bbox': tensor(0.1413, device='cuda:0'), 'validation_loss_giou': tensor(0.5145, device='cuda:0'), 'validation_cardinality_error': tensor(1., device='cuda:0')}
{'training_loss': tensor(2.2961, device='cuda:0'), 'train_loss_ce': tensor(0.3717, device='cuda:0'), 'train_loss_bbox': tensor(0.1444, device='cuda:0'), 'train_loss_giou': tensor(0.6012, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.9445, device='cuda:0'), 'validation_loss_ce': tensor(0.3745, device='cuda:0'), 'validation_loss_bbox': tensor(0.1293, device='cuda:0'), 'validation_loss_giou': tensor(0.4619, device='cuda:0'), 'validation_cardinality_error': tensor(0.8990, device='cuda:0')}
{'training_loss': tensor(2.0082, device='cuda:0'), 'train_loss_ce': tensor(0.4834, device='cuda:0'), 'train_loss_bbox': tensor(0.1454, device='cuda:0'), 'train_loss_giou': tensor(0.3990, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.8980, device='cuda:0'), 'validation_loss_ce': tensor(0.3738, device='cuda:0'), 'validation_loss_bbox': tensor(0.1218, device='cuda:0'), 'validation_loss_giou': tensor(0.4576, device='cuda:0'), 'validation_cardinality_error': tensor(0.8182, device='cuda:0')}
{'training_loss': tensor(2.5432, device='cuda:0'), 'train_loss_ce': tensor(0.4593, device='cuda:0'), 'train_loss_bbox': tensor(0.1676, device='cuda:0'), 'train_loss_giou': tensor(0.6229, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.9865, device='cuda:0'), 'validation_loss_ce': tensor(0.3900, device='cuda:0'), 'validation_loss_bbox': tensor(0.1392, device='cuda:0'), 'validation_loss_giou': tensor(0.4503, device='cuda:0'), 'validation_cardinality_error': tensor(1., device='cuda:0')}
{'training_loss': tensor(1.9886, device='cuda:0'), 'train_loss_ce': tensor(0.3825, device='cuda:0'), 'train_loss_bbox': tensor(0.1169, device='cuda:0'), 'train_loss_giou': tensor(0.5109, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.9185, device='cuda:0'), 'validation_loss_ce': tensor(0.3755, device='cuda:0'), 'validation_loss_bbox': tensor(0.1302, device='cuda:0'), 'validation_loss_giou': tensor(0.4460, device='cuda:0'), 'validation_cardinality_error': tensor(0.9697, device='cuda:0')}
{'training_loss': tensor(1.4652, device='cuda:0'), 'train_loss_ce': tensor(0.3647, device='cuda:0'), 'train_loss_bbox': tensor(0.0751, device='cuda:0'), 'train_loss_giou': tensor(0.3625, device='cuda:0'), 'train_cardinality_error': tensor(0.8000, device='cuda:0'), 'validation_loss': tensor(1.9299, device='cuda:0'), 'validation_loss_ce': tensor(0.3658, device='cuda:0'), 'validation_loss_bbox': tensor(0.1326, device='cuda:0'), 'validation_loss_giou': tensor(0.4505, device='cuda:0'), 'validation_cardinality_error': tensor(0.3434, device='cuda:0')}
{'training_loss': tensor(1.8145, device='cuda:0'), 'train_loss_ce': tensor(0.4718, device='cuda:0'), 'train_loss_bbox': tensor(0.0969, device='cuda:0'), 'train_loss_giou': tensor(0.4291, device='cuda:0'), 'train_cardinality_error': tensor(0.4000, device='cuda:0'), 'validation_loss': tensor(1.9555, device='cuda:0'), 'validation_loss_ce': tensor(0.3583, device='cuda:0'), 'validation_loss_bbox': tensor(0.1334, device='cuda:0'), 'validation_loss_giou': tensor(0.4650, device='cuda:0'), 'validation_cardinality_error': tensor(0.5152, device='cuda:0')}
{'training_loss': tensor(2.0291, device='cuda:0'), 'train_loss_ce': tensor(0.3718, device='cuda:0'), 'train_loss_bbox': tensor(0.1291, device='cuda:0'), 'train_loss_giou': tensor(0.5060, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(2.3917, device='cuda:0'), 'validation_loss_ce': tensor(0.3621, device='cuda:0'), 'validation_loss_bbox': tensor(0.1767, device='cuda:0'), 'validation_loss_giou': tensor(0.5731, device='cuda:0'), 'validation_cardinality_error': tensor(0.7879, device='cuda:0')}
{'training_loss': tensor(2.1299, device='cuda:0'), 'train_loss_ce': tensor(0.4300, device='cuda:0'), 'train_loss_bbox': tensor(0.1054, device='cuda:0'), 'train_loss_giou': tensor(0.5863, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(2.1091, device='cuda:0'), 'validation_loss_ce': tensor(0.3579, device='cuda:0'), 'validation_loss_bbox': tensor(0.1479, device='cuda:0'), 'validation_loss_giou': tensor(0.5058, device='cuda:0'), 'validation_cardinality_error': tensor(0.6061, device='cuda:0')}
{'training_loss': tensor(2.0963, device='cuda:0'), 'train_loss_ce': tensor(0.2759, device='cuda:0'), 'train_loss_bbox': tensor(0.1520, device='cuda:0'), 'train_loss_giou': tensor(0.5301, device='cuda:0'), 'train_cardinality_error': tensor(0.6000, device='cuda:0'), 'validation_loss': tensor(2.0418, device='cuda:0'), 'validation_loss_ce': tensor(0.3464, device='cuda:0'), 'validation_loss_bbox': tensor(0.1385, device='cuda:0'), 'validation_loss_giou': tensor(0.5015, device='cuda:0'), 'validation_cardinality_error': tensor(0.6263, device='cuda:0')}
{'training_loss': tensor(2.1784, device='cuda:0'), 'train_loss_ce': tensor(0.3634, device='cuda:0'), 'train_loss_bbox': tensor(0.1741, device='cuda:0'), 'train_loss_giou': tensor(0.4723, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.9543, device='cuda:0'), 'validation_loss_ce': tensor(0.3518, device='cuda:0'), 'validation_loss_bbox': tensor(0.1322, device='cuda:0'), 'validation_loss_giou': tensor(0.4707, device='cuda:0'), 'validation_cardinality_error': tensor(0.9495, device='cuda:0')}
{'training_loss': tensor(2.0823, device='cuda:0'), 'train_loss_ce': tensor(0.3437, device='cuda:0'), 'train_loss_bbox': tensor(0.1144, device='cuda:0'), 'train_loss_giou': tensor(0.5833, device='cuda:0'), 'train_cardinality_error': tensor(0.8000, device='cuda:0'), 'validation_loss': tensor(1.8429, device='cuda:0'), 'validation_loss_ce': tensor(0.3513, device='cuda:0'), 'validation_loss_bbox': tensor(0.1216, device='cuda:0'), 'validation_loss_giou': tensor(0.4417, device='cuda:0'), 'validation_cardinality_error': tensor(0.8687, device='cuda:0')}
{'training_loss': tensor(1.4899, device='cuda:0'), 'train_loss_ce': tensor(0.3300, device='cuda:0'), 'train_loss_bbox': tensor(0.0826, device='cuda:0'), 'train_loss_giou': tensor(0.3734, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.8750, device='cuda:0'), 'validation_loss_ce': tensor(0.3539, device='cuda:0'), 'validation_loss_bbox': tensor(0.1210, device='cuda:0'), 'validation_loss_giou': tensor(0.4580, device='cuda:0'), 'validation_cardinality_error': tensor(0.7273, device='cuda:0')}
{'training_loss': tensor(2.2234, device='cuda:0'), 'train_loss_ce': tensor(0.3947, device='cuda:0'), 'train_loss_bbox': tensor(0.1587, device='cuda:0'), 'train_loss_giou': tensor(0.5175, device='cuda:0'), 'train_cardinality_error': tensor(0.8000, device='cuda:0'), 'validation_loss': tensor(1.9507, device='cuda:0'), 'validation_loss_ce': tensor(0.3356, device='cuda:0'), 'validation_loss_bbox': tensor(0.1303, device='cuda:0'), 'validation_loss_giou': tensor(0.4817, device='cuda:0'), 'validation_cardinality_error': tensor(0.5657, device='cuda:0')}
{'training_loss': tensor(1.5594, device='cuda:0'), 'train_loss_ce': tensor(0.3898, device='cuda:0'), 'train_loss_bbox': tensor(0.1000, device='cuda:0'), 'train_loss_giou': tensor(0.3347, device='cuda:0'), 'train_cardinality_error': tensor(0.6000, device='cuda:0'), 'validation_loss': tensor(1.9082, device='cuda:0'), 'validation_loss_ce': tensor(0.3350, device='cuda:0'), 'validation_loss_bbox': tensor(0.1296, device='cuda:0'), 'validation_loss_giou': tensor(0.4626, device='cuda:0'), 'validation_cardinality_error': tensor(0.5859, device='cuda:0')}
{'training_loss': tensor(1.3733, device='cuda:0'), 'train_loss_ce': tensor(0.2466, device='cuda:0'), 'train_loss_bbox': tensor(0.0828, device='cuda:0'), 'train_loss_giou': tensor(0.3563, device='cuda:0'), 'train_cardinality_error': tensor(0.8000, device='cuda:0'), 'validation_loss': tensor(1.8808, device='cuda:0'), 'validation_loss_ce': tensor(0.3422, device='cuda:0'), 'validation_loss_bbox': tensor(0.1253, device='cuda:0'), 'validation_loss_giou': tensor(0.4559, device='cuda:0'), 'validation_cardinality_error': tensor(0.7273, device='cuda:0')}
{'training_loss': tensor(1.9375, device='cuda:0'), 'train_loss_ce': tensor(0.3146, device='cuda:0'), 'train_loss_bbox': tensor(0.1273, device='cuda:0'), 'train_loss_giou': tensor(0.4931, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.8543, device='cuda:0'), 'validation_loss_ce': tensor(0.3441, device='cuda:0'), 'validation_loss_bbox': tensor(0.1196, device='cuda:0'), 'validation_loss_giou': tensor(0.4561, device='cuda:0'), 'validation_cardinality_error': tensor(0.7879, device='cuda:0')}
{'training_loss': tensor(1.7360, device='cuda:0'), 'train_loss_ce': tensor(0.2615, device='cuda:0'), 'train_loss_bbox': tensor(0.1172, device='cuda:0'), 'train_loss_giou': tensor(0.4441, device='cuda:0'), 'train_cardinality_error': tensor(0.6000, device='cuda:0'), 'validation_loss': tensor(1.8177, device='cuda:0'), 'validation_loss_ce': tensor(0.3491, device='cuda:0'), 'validation_loss_bbox': tensor(0.1182, device='cuda:0'), 'validation_loss_giou': tensor(0.4387, device='cuda:0'), 'validation_cardinality_error': tensor(0.6465, device='cuda:0')}
{'training_loss': tensor(1.4278, device='cuda:0'), 'train_loss_ce': tensor(0.4191, device='cuda:0'), 'train_loss_bbox': tensor(0.0621, device='cuda:0'), 'train_loss_giou': tensor(0.3491, device='cuda:0'), 'train_cardinality_error': tensor(0.6000, device='cuda:0'), 'validation_loss': tensor(1.7745, device='cuda:0'), 'validation_loss_ce': tensor(0.3331, device='cuda:0'), 'validation_loss_bbox': tensor(0.1184, device='cuda:0'), 'validation_loss_giou': tensor(0.4248, device='cuda:0'), 'validation_cardinality_error': tensor(0.5152, device='cuda:0')}
{'training_loss': tensor(1.5275, device='cuda:0'), 'train_loss_ce': tensor(0.2538, device='cuda:0'), 'train_loss_bbox': tensor(0.1081, device='cuda:0'), 'train_loss_giou': tensor(0.3666, device='cuda:0'), 'train_cardinality_error': tensor(0.4000, device='cuda:0'), 'validation_loss': tensor(1.7723, device='cuda:0'), 'validation_loss_ce': tensor(0.3335, device='cuda:0'), 'validation_loss_bbox': tensor(0.1156, device='cuda:0'), 'validation_loss_giou': tensor(0.4303, device='cuda:0'), 'validation_cardinality_error': tensor(0.5657, device='cuda:0')}
{'training_loss': tensor(1.2553, device='cuda:0'), 'train_loss_ce': tensor(0.2647, device='cuda:0'), 'train_loss_bbox': tensor(0.0725, device='cuda:0'), 'train_loss_giou': tensor(0.3140, device='cuda:0'), 'train_cardinality_error': tensor(0.4000, device='cuda:0'), 'validation_loss': tensor(1.7969, device='cuda:0'), 'validation_loss_ce': tensor(0.3384, device='cuda:0'), 'validation_loss_bbox': tensor(0.1183, device='cuda:0'), 'validation_loss_giou': tensor(0.4334, device='cuda:0'), 'validation_cardinality_error': tensor(0.6667, device='cuda:0')}
{'training_loss': tensor(2.2105, device='cuda:0'), 'train_loss_ce': tensor(0.4101, device='cuda:0'), 'train_loss_bbox': tensor(0.1634, device='cuda:0'), 'train_loss_giou': tensor(0.4917, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.7978, device='cuda:0'), 'validation_loss_ce': tensor(0.3421, device='cuda:0'), 'validation_loss_bbox': tensor(0.1197, device='cuda:0'), 'validation_loss_giou': tensor(0.4286, device='cuda:0'), 'validation_cardinality_error': tensor(0.7677, device='cuda:0')}
{'training_loss': tensor(1.3302, device='cuda:0'), 'train_loss_ce': tensor(0.3538, device='cuda:0'), 'train_loss_bbox': tensor(0.0723, device='cuda:0'), 'train_loss_giou': tensor(0.3074, device='cuda:0'), 'train_cardinality_error': tensor(0.4000, device='cuda:0'), 'validation_loss': tensor(1.8850, device='cuda:0'), 'validation_loss_ce': tensor(0.3273, device='cuda:0'), 'validation_loss_bbox': tensor(0.1313, device='cuda:0'), 'validation_loss_giou': tensor(0.4507, device='cuda:0'), 'validation_cardinality_error': tensor(0.5960, device='cuda:0')}
{'training_loss': tensor(1.5171, device='cuda:0'), 'train_loss_ce': tensor(0.3285, device='cuda:0'), 'train_loss_bbox': tensor(0.0630, device='cuda:0'), 'train_loss_giou': tensor(0.4368, device='cuda:0'), 'train_cardinality_error': tensor(1., device='cuda:0'), 'validation_loss': tensor(1.7658, device='cuda:0'), 'validation_loss_ce': tensor(0.3375, device='cuda:0'), 'validation_loss_bbox': tensor(0.1182, device='cuda:0'), 'validation_loss_giou': tensor(0.4186, device='cuda:0'), 'validation_cardinality_error': tensor(0.7273, device='cuda:0')}
```

## Examples
{'size': tensor([512, 512]), 'image_id': tensor([1]), 'class_labels': tensor([4]), 'boxes': tensor([[0.2622, 0.5729, 0.0847, 0.0773]]), 'area': tensor([1717.9431]), 'iscrowd': tensor([0]), 'orig_size': tensor([1024, 1024])}

![Example](./example.png)