rt-detr-v2_barcode-detection

This model is a fine-tuned version of PekingU/rtdetr_v2_r18vd on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 5.0340
• Map: 0.6583
• Map 50: 0.8036
• Map 75: 0.7156
• Map Small: 0.2754
• Map Medium: 0.7152
• Map Large: 0.747
• Mar 1: 0.3589
• Mar 10: 0.833
• Mar 100: 0.8673
• Mar Small: 0.6183
• Mar Medium: 0.8656
• Mar Large: 0.8872
• Map Barcode: 0.6583
• Mar 100 Barcode: 0.8673

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 300
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Map	Map 50	Map 75	Map Barcode	Map Large	Map Medium	Map Small	Mar 1	Mar 10	Mar 100	Mar 100 Barcode	Mar Large	Mar Medium	Mar Small
7.43	1.0	1636	5.0547	0.623	0.7756	0.6841	0.623	0.6594	0.6674	0.2648	0.3542	0.8117	0.8521	0.8521	0.8749	0.8552	0.5325
6.9538	2.0	3272	4.9508	0.6297	0.7829	0.6878	0.6297	0.7109	0.6449	0.261	0.3612	0.8238	0.8604	0.8604	0.8794	0.8624	0.5972
6.6733	3.0	4908	5.0056	0.6553	0.8073	0.7122	0.6553	0.7359	0.6782	0.2319	0.3481	0.8201	0.8542	0.8542	0.8712	0.8573	0.6107
6.5783	4.0	6544	5.0493	0.6524	0.808	0.711	0.6524	0.7472	0.6782	0.2391	0.3581	0.8268	0.8629	0.8629	0.8819	0.8639	0.6066
6.4986	5.0	8180	5.0208	0.6944	0.8578	0.7542	0.6944	0.7651	0.7135	0.2808	0.3706	0.8307	0.8611	0.8611	0.879	0.863	0.6131
6.4056	6.0	9816	5.0116	0.6843	0.8399	0.7445	0.6843	0.7607	0.7126	0.3026	0.371	0.8374	0.8664	0.8664	0.8864	0.8646	0.6176
6.4074	7.0	11452	5.0229	0.653	0.7989	0.7106	0.2768	0.7112	0.731	0.3513	0.8326	0.8694	0.619	0.8675	0.8896	0.653	0.8694
6.2576	8.0	13088	5.0340	0.6583	0.8036	0.7156	0.2754	0.7152	0.747	0.3589	0.833	0.8673	0.6183	0.8656	0.8872	0.6583	0.8673

Framework versions

• Transformers 4.57.3
• Pytorch 2.9.0+cu126
• Datasets 4.4.2
• Tokenizers 0.22.1