detr-resnet-50-dc5-ordd2024-finetuned

This model is a fine-tuned version of facebook/detr-resnet-50 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.0037
• Map: 0.0003
• Map 50: 0.0011
• Map 75: 0.0001
• Map Small: 0.0125
• Map Medium: 0.0
• Map Large: 0.001
• Mar 1: 0.0
• Mar 10: 0.0132
• Mar 100: 0.1132
• Mar Small: 0.025
• Mar Medium: 0.05
• Mar Large: 0.1435
• Map Longitudinal Crack: 0.0003
• Mar 100 Longitudinal Crack: 0.0526
• Map Transverse Crack: -1.0
• Mar 100 Transverse Crack: -1.0
• Map Aligator Crack: 0.0009
• Mar 100 Aligator Crack: 0.4
• Map Pothole: 0.0
• Mar 100 Pothole: 0.0
• Map Other Corruptions: 0.0
• Mar 100 Other Corruptions: 0.0

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Map	Map 50	Map 75	Map Small	Map Medium	Map Large	Mar 1	Mar 10	Mar 100	Mar Small	Mar Medium	Mar Large	Map Longitudinal Crack	Mar 100 Longitudinal Crack	Map Transverse Crack	Mar 100 Transverse Crack	Map Aligator Crack	Mar 100 Aligator Crack	Map Pothole	Mar 100 Pothole	Map Other Corruptions	Mar 100 Other Corruptions
No log	1.0	174	3.9597	0.0001	0.0006	0.0001	0.0	0.0	0.0007	0.0	0.0161	0.0341	0.0	0.0	0.0864	0.0004	0.0421	-1.0	-1.0	0.0001	0.0722	0.0	0.0222	0.0	0.0
No log	2.0	348	2.4478	0.0004	0.0034	0.0	0.0	0.0	0.0013	0.0	0.0039	0.0762	0.0	0.0	0.092	0.0012	0.0158	-1.0	-1.0	0.0006	0.2889	0.0	0.0	0.0	0.0
3.3803	3.0	522	2.1677	0.0001	0.0004	0.0	0.0	0.0001	0.0005	0.0	0.0055	0.0686	0.0	0.0179	0.0824	0.0001	0.0579	-1.0	-1.0	0.0004	0.2167	0.0	0.0	0.0	0.0
3.3803	4.0	696	2.1413	0.0002	0.0006	0.0001	0.0	0.0	0.0007	0.0	0.0066	0.0842	0.0	0.0071	0.1022	0.0001	0.0368	-1.0	-1.0	0.0006	0.3	0.0	0.0	0.0	0.0
3.3803	5.0	870	2.0419	0.0002	0.0008	0.0	0.0	0.0	0.0015	0.0	0.008	0.0871	0.0	0.0	0.1085	0.0001	0.0263	-1.0	-1.0	0.0008	0.3222	0.0	0.0	0.0	0.0
2.2481	6.0	1044	1.9375	0.0002	0.0006	0.0001	0.0	0.0	0.0008	0.0	0.0107	0.0843	0.0	0.0	0.1074	0.0001	0.0316	-1.0	-1.0	0.0006	0.3056	0.0	0.0	0.0	0.0
2.2481	7.0	1218	1.9988	0.0001	0.0005	0.0001	0.0	0.0	0.0005	0.0	0.0066	0.0885	0.0	0.0	0.11	0.0001	0.0263	-1.0	-1.0	0.0005	0.3278	0.0	0.0	0.0	0.0
2.2481	8.0	1392	2.0228	0.0002	0.0008	0.0001	0.0125	0.0	0.0008	0.0	0.0132	0.098	0.025	0.0	0.1275	0.0002	0.0474	-1.0	-1.0	0.0007	0.3444	0.0	0.0	0.0	0.0
2.1491	9.0	1566	1.9563	0.0002	0.0007	0.0002	0.0	0.0	0.0008	0.0	0.004	0.1012	0.0	0.0	0.1181	0.0002	0.0105	-1.0	-1.0	0.0008	0.3944	0.0	0.0	0.0	0.0
2.1491	10.0	1740	2.0037	0.0003	0.0011	0.0001	0.0125	0.0	0.001	0.0	0.0132	0.1132	0.025	0.05	0.1435	0.0003	0.0526	-1.0	-1.0	0.0009	0.4	0.0	0.0	0.0	0.0

Framework versions

• Transformers 4.52.4
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1