rt_detrv2_finetuned_trashify_box_detector_v1

This model is a fine-tuned version of PekingU/rtdetr_v2_r50vd on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 9.6275
• Map: 0.4672
• Map 50: 0.6559
• Map 75: 0.546
• Map Bin: 0.7377
• Map Hand: 0.4683
• Map Large: 0.487
• Map Medium: 0.2697
• Map Not Bin: 0.0629
• Map Not Hand: -1.0
• Map Not Trash: 0.2531
• Map Small: 0.0186
• Map Trash: 0.6416
• Map Trash Arm: 0.6394
• Mar 1: 0.5265
• Mar 10: 0.6818
• Mar 100: 0.7353
• Mar 100 Bin: 0.878
• Mar 100 Hand: 0.7902
• Mar 100 Not Bin: 0.5214
• Mar 100 Not Hand: -1.0
• Mar 100 Not Trash: 0.6625
• Mar 100 Trash: 0.7929
• Mar 100 Trash Arm: 0.7667
• Mar Large: 0.7603
• Mar Medium: 0.4727
• Mar Small: 0.4

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: 0.0001
• train_batch_size: 8
• 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_ratio: 0.05
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Map	Map 50	Map 75	Map Bin	Map Hand	Map Large	Map Medium	Map Not Bin	Map Not Hand	Map Not Trash	Map Small	Map Trash	Map Trash Arm	Mar 1	Mar 10	Mar 100	Mar 100 Bin	Mar 100 Hand	Mar 100 Not Bin	Mar 100 Not Hand	Mar 100 Not Trash	Mar 100 Trash	Mar 100 Trash Arm	Mar Large	Mar Medium	Mar Small
61.0002	1.0	99	15.2461	0.2878	0.4135	0.3138	0.6102	0.5428	0.2955	0.0907	0.013	-1.0	0.0391	0.0188	0.5218	0.0	0.3104	0.4788	0.5515	0.8511	0.8176	0.3286	-1.0	0.5611	0.7504	0.0	0.5801	0.3756	0.15
20.2622	2.0	198	11.0607	0.3966	0.5492	0.4472	0.682	0.4783	0.416	0.2504	0.0315	-1.0	0.1484	0.065	0.5927	0.4468	0.4785	0.6461	0.7488	0.8837	0.802	0.6214	-1.0	0.5806	0.7717	0.8333	0.7825	0.433	0.3
16.8192	3.0	297	10.6279	0.4234	0.5866	0.4979	0.6857	0.4725	0.4404	0.1608	0.0297	-1.0	0.1889	0.1143	0.6056	0.5582	0.5068	0.6498	0.7181	0.8674	0.8	0.5	-1.0	0.5903	0.7841	0.7667	0.753	0.3835	0.3
15.062	4.0	396	9.9305	0.4518	0.6257	0.5146	0.7459	0.5363	0.466	0.2175	0.054	-1.0	0.2358	0.185	0.585	0.5539	0.5164	0.6851	0.7499	0.8943	0.8147	0.55	-1.0	0.6222	0.785	0.8333	0.7807	0.4193	0.4
13.7257	5.0	495	9.8895	0.436	0.6177	0.512	0.7475	0.4603	0.4533	0.2587	0.0544	-1.0	0.2198	0.06	0.598	0.5362	0.5093	0.6776	0.7229	0.8957	0.8088	0.5	-1.0	0.6208	0.7788	0.7333	0.7573	0.4091	0.3
12.8647	6.0	594	9.9595	0.4303	0.6248	0.5199	0.723	0.4792	0.4452	0.2223	0.0486	-1.0	0.2155	0.1286	0.6215	0.4942	0.4933	0.6604	0.7055	0.873	0.802	0.5071	-1.0	0.6167	0.7673	0.6667	0.7338	0.4	0.3
12.0576	7.0	693	9.4153	0.4513	0.6264	0.5265	0.7643	0.4492	0.4693	0.211	0.0415	-1.0	0.2474	0.0675	0.6363	0.5692	0.5277	0.6996	0.7426	0.8922	0.7824	0.5643	-1.0	0.6556	0.7947	0.7667	0.7712	0.4614	0.35
11.2359	8.0	792	9.5170	0.4771	0.655	0.5539	0.7583	0.4758	0.494	0.2926	0.0661	-1.0	0.2451	0.0303	0.6501	0.6673	0.5331	0.6997	0.7398	0.8943	0.799	0.4857	-1.0	0.6625	0.7973	0.8	0.7641	0.4932	0.3
10.6301	9.0	891	9.7439	0.4534	0.6356	0.5248	0.742	0.4354	0.4738	0.2788	0.0892	-1.0	0.246	0.0227	0.6385	0.5692	0.5247	0.6838	0.739	0.8865	0.7843	0.5357	-1.0	0.6569	0.8035	0.7667	0.7622	0.4864	0.4
10.2602	10.0	990	9.6275	0.4672	0.6559	0.546	0.7377	0.4683	0.487	0.2697	0.0629	-1.0	0.2531	0.0186	0.6416	0.6394	0.5265	0.6818	0.7353	0.878	0.7902	0.5214	-1.0	0.6625	0.7929	0.7667	0.7603	0.4727	0.4

Framework versions

• Transformers 4.55.2
• Pytorch 2.9.1
• Datasets 4.5.0
• Tokenizers 0.21.4