conditional-detr-resnet-50_til-2023-cv-9

This model is a fine-tuned version of microsoft/conditional-detr-resnet-50 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2502
• Loss Ce: 0.0010
• Loss Bbox: 0.0160
• Loss Giou: 0.0842
• Cardinality Error: 2.1237
• Map: 0.8063
• Map 50: 0.9901
• Map 75: 0.9609
• Map Small: 0.8063
• Map Medium: -1.0
• Map Large: -1.0
• Mar 1: 0.4097
• Mar 10: 0.8555
• Mar 100: 0.8555
• Mar Small: 0.8555
• Mar Medium: -1.0
• Mar Large: -1.0
• Map Per Class: -1.0
• Mar 100 Per Class: -1.0

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Loss Ce	Loss Bbox	Loss Giou	Cardinality Error	Map	Map 50	Map 75	Map Small	Map Medium	Map Large	Mar 1	Mar 10	Mar 100	Mar Small	Mar Medium	Mar Large	Map Per Class	Mar 100 Per Class
0.4695	1.0	708	0.4327	0.0120	0.0256	0.1404	2.1237	0.7356	0.9796	0.9229	0.7356	-1.0	-1.0	0.3810	0.7910	0.7910	0.7910	-1.0	-1.0	-1.0	-1.0
0.2915	2.0	1416	0.3432	0.0056	0.0217	0.1118	2.1237	0.7640	0.9892	0.9391	0.7640	-1.0	-1.0	0.3900	0.8128	0.8128	0.8128	-1.0	-1.0	-1.0	-1.0
0.2713	3.0	2124	0.3150	0.0063	0.0194	0.1026	2.1237	0.7819	0.9894	0.9494	0.7819	-1.0	-1.0	0.3977	0.8274	0.8274	0.8274	-1.0	-1.0	-1.0	-1.0
0.2583	4.0	2832	0.2754	0.0026	0.0174	0.0915	2.1237	0.7931	0.9898	0.9515	0.7931	-1.0	-1.0	0.4026	0.8387	0.8387	0.8387	-1.0	-1.0	-1.0	-1.0
0.2264	5.0	3540	0.2768	0.0019	0.0178	0.0921	2.1237	0.8011	0.9899	0.9623	0.8011	-1.0	-1.0	0.4057	0.8452	0.8452	0.8452	-1.0	-1.0	-1.0	-1.0
0.2841	6.0	4248	0.3362	0.0049	0.0207	0.1115	2.1237	0.7973	0.9900	0.9614	0.7973	-1.0	-1.0	0.4043	0.8434	0.8434	0.8434	-1.0	-1.0	-1.0	-1.0
0.2929	7.0	4956	0.3310	0.0078	0.0203	0.1071	2.1237	0.7986	0.9899	0.9616	0.7986	-1.0	-1.0	0.4053	0.8445	0.8445	0.8445	-1.0	-1.0	-1.0	-1.0
0.2405	8.0	5664	0.2681	0.0017	0.0168	0.0904	2.1237	0.8018	0.9900	0.9619	0.8018	-1.0	-1.0	0.4067	0.8481	0.8481	0.8481	-1.0	-1.0	-1.0	-1.0
0.1851	9.0	6372	0.2680	0.0019	0.0168	0.0901	2.1237	0.8050	0.9900	0.9622	0.8050	-1.0	-1.0	0.4081	0.8511	0.8511	0.8511	-1.0	-1.0	-1.0	-1.0
0.1842	10.0	7080	0.2553	0.0013	0.0163	0.0856	2.1237	0.8074	0.9900	0.9627	0.8074	-1.0	-1.0	0.4095	0.8544	0.8544	0.8544	-1.0	-1.0	-1.0	-1.0
0.3201	11.0	7788	0.3556	0.0034	0.0226	0.1179	2.1237	0.8040	0.9900	0.9617	0.8040	-1.0	-1.0	0.4080	0.8511	0.8511	0.8511	-1.0	-1.0	-1.0	-1.0
0.266	12.0	8496	0.3296	0.0021	0.0191	0.1151	2.1237	0.7996	0.9900	0.9600	0.7996	-1.0	-1.0	0.4069	0.8489	0.8489	0.8489	-1.0	-1.0	-1.0	-1.0
0.2086	13.0	9204	0.2753	0.0016	0.0178	0.0916	2.1237	0.8007	0.9900	0.9603	0.8007	-1.0	-1.0	0.4076	0.8506	0.8506	0.8506	-1.0	-1.0	-1.0	-1.0
0.1853	14.0	9912	0.2452	0.0009	0.0156	0.0827	2.1237	0.8037	0.9900	0.9606	0.8037	-1.0	-1.0	0.4088	0.8533	0.8533	0.8533	-1.0	-1.0	-1.0	-1.0
0.1588	15.0	10620	0.2502	0.0010	0.0160	0.0842	2.1237	0.8063	0.9901	0.9609	0.8063	-1.0	-1.0	0.4097	0.8555	0.8555	0.8555	-1.0	-1.0	-1.0	-1.0

Framework versions

• Transformers 4.29.2
• Pytorch 2.0.0
• Datasets 2.1.0
• Tokenizers 0.13.3