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: 9.3610
• Map: 0.4519
• Map 50: 0.6127
• Map 75: 0.5328
• Map Small: 0.0
• Map Medium: 0.2634
• Map Large: 0.4637
• Mar 1: 0.5091
• Mar 10: 0.6874
• Mar 100: 0.7327
• Mar Small: 0.0
• Mar Medium: 0.5854
• Mar Large: 0.7464
• Map Bin: 0.7301
• Mar 100 Bin: 0.89
• Map Hand: 0.5973
• Mar 100 Hand: 0.8465
• Map Not Bin: 0.1381
• Mar 100 Not Bin: 0.5636
• Map Not Hand: 0.0086
• Mar 100 Not Hand: 0.55
• Map Not Trash: 0.2291
• Mar 100 Not Trash: 0.6151
• Map Trash: 0.6786
• Mar 100 Trash: 0.8211
• Map Trash Arm: 0.7817
• Mar 100 Trash Arm: 0.8429

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: 16
• eval_batch_size: 16
• 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
• 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 Bin	Mar 100 Bin	Map Hand	Mar 100 Hand	Map Not Bin	Mar 100 Not Bin	Map Not Hand	Mar 100 Not Hand	Map Not Trash	Mar 100 Not Trash	Map Trash	Mar 100 Trash	Map Trash Arm	Mar 100 Trash Arm
72.6397	1.0	50	16.7480	0.2175	0.3089	0.243	0.0333	0.0679	0.2262	0.2629	0.4365	0.4851	0.3	0.3328	0.5034	0.5808	0.8827	0.4481	0.8364	0.0021	0.4182	0.0	0.0	0.0595	0.4857	0.4317	0.7725	0.0	0.0
23.5316	2.0	100	11.1267	0.2822	0.3971	0.3194	0.4	0.2001	0.2966	0.3528	0.5543	0.6681	0.4	0.4565	0.6904	0.6808	0.8363	0.5507	0.8202	0.0041	0.6182	0.0014	0.3333	0.103	0.5786	0.6167	0.7902	0.0184	0.7
17.7918	3.0	150	10.0192	0.3915	0.5264	0.4572	0.0	0.2399	0.4057	0.4833	0.6719	0.7029	0.0	0.4319	0.7294	0.7086	0.873	0.5875	0.8202	0.0742	0.4818	0.005	0.45	0.1319	0.5913	0.6569	0.8186	0.5766	0.8857
15.6436	4.0	200	9.8303	0.4036	0.5471	0.4809	0.0	0.2822	0.4134	0.4615	0.6831	0.7245	0.0	0.4427	0.7506	0.7117	0.8789	0.5855	0.8404	0.0379	0.6182	0.0021	0.4833	0.2024	0.6032	0.6659	0.8049	0.6193	0.8429
14.251	5.0	250	9.4803	0.4514	0.5947	0.531	0.0	0.2404	0.46	0.491	0.677	0.7205	0.0	0.4481	0.7469	0.734	0.8817	0.6008	0.846	0.1653	0.5545	0.0023	0.5	0.1883	0.604	0.6719	0.8147	0.7974	0.8429
12.9869	6.0	300	9.3610	0.4519	0.6127	0.5328	0.0	0.2634	0.4637	0.5091	0.6874	0.7327	0.0	0.5854	0.7464	0.7301	0.89	0.5973	0.8465	0.1381	0.5636	0.0086	0.55	0.2291	0.6151	0.6786	0.8211	0.7817	0.8429
12.0088	7.0	350	9.4350	0.4505	0.6093	0.5337	0.0	0.2782	0.462	0.5139	0.6792	0.7209	0.0	0.472	0.7419	0.7243	0.882	0.6092	0.8444	0.188	0.6182	0.0015	0.35	0.2118	0.6071	0.6751	0.8157	0.7435	0.9286
11.0783	8.0	400	9.4135	0.4684	0.6146	0.5401	0.0	0.3219	0.4795	0.4938	0.6805	0.7165	0.0	0.4786	0.7366	0.7471	0.8858	0.6067	0.8333	0.1835	0.5909	0.0024	0.3833	0.2233	0.6048	0.6661	0.8176	0.8496	0.9
10.4565	9.0	450	9.4666	0.4555	0.6103	0.5368	0.0	0.2937	0.4669	0.5006	0.6709	0.7175	0.0	0.4578	0.7404	0.7394	0.8837	0.5975	0.8374	0.1835	0.5818	0.002	0.4167	0.2173	0.5968	0.6642	0.8059	0.785	0.9
9.8849	10.0	500	9.5216	0.4576	0.6098	0.5349	0.0	0.3041	0.4686	0.5027	0.6846	0.7265	0.0	0.4688	0.7497	0.7256	0.8855	0.603	0.8369	0.1843	0.6	0.0026	0.45	0.2158	0.6063	0.6652	0.8069	0.8066	0.9

Framework versions

• Transformers 4.57.3
• Pytorch 2.9.0+cu126
• Datasets 4.5.0
• Tokenizers 0.22.2