rtdetr_finetuned_wgisd

This model is a fine-tuned version of PekingU/rtdetr_r50vd on the None dataset. It achieves the following results on the evaluation set:

• Loss: 9.4202
• Map: 0.4462
• Map 50: 0.7009
• Map 75: 0.4734
• Map Small: -1.0
• Map Medium: 0.234
• Map Large: 0.4809
• Mar 1: 0.0501
• Mar 10: 0.456
• Mar 100: 0.7177
• Mar Small: -1.0
• Mar Medium: 0.4982
• Mar Large: 0.7316
• Map Cdy: 0.4875
• Mar 100 Cdy: 0.7605
• Map Cfr: 0.3124
• Mar 100 Cfr: 0.7061
• Map Csv: 0.5328
• Mar 100 Csv: 0.7537
• Map Svb: 0.447
• Mar 100 Svb: 0.6895
• Map Syh: 0.4512
• Mar 100 Syh: 0.6788

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: 4
• 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: cosine
• num_epochs: 20
• 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 Cdy	Mar 100 Cdy	Map Cfr	Mar 100 Cfr	Map Csv	Mar 100 Csv	Map Svb	Mar 100 Svb	Map Syh	Mar 100 Syh
No log	1.0	60	9.6758	0.4335	0.6535	0.4668	-1.0	0.2255	0.4569	0.0568	0.4307	0.675	-1.0	0.4173	0.6904	0.5298	0.75	0.4614	0.6897	0.5017	0.724	0.1802	0.5432	0.4943	0.6682
No log	2.0	120	11.2094	0.4428	0.6959	0.4866	-1.0	0.231	0.4625	0.0576	0.4462	0.6902	-1.0	0.4869	0.702	0.4535	0.7368	0.3607	0.6588	0.4601	0.7149	0.4514	0.6737	0.4881	0.6671
No log	3.0	180	10.0152	0.4382	0.6656	0.4814	-1.0	0.21	0.4586	0.0584	0.4502	0.7178	-1.0	0.4766	0.7334	0.479	0.7904	0.384	0.697	0.3303	0.7215	0.4465	0.6779	0.5514	0.7024
No log	4.0	240	10.7644	0.392	0.6034	0.4291	-1.0	0.2149	0.4152	0.0553	0.4312	0.693	-1.0	0.4786	0.7056	0.5448	0.7895	0.4015	0.6885	0.3897	0.7372	0.1561	0.5558	0.4677	0.6941
No log	5.0	300	9.2331	0.4517	0.6983	0.4826	-1.0	0.2471	0.4771	0.0582	0.4514	0.7044	-1.0	0.4961	0.7165	0.5238	0.7627	0.3693	0.7115	0.5181	0.7256	0.3906	0.6563	0.4568	0.6659
No log	6.0	360	9.2305	0.4381	0.6823	0.475	-1.0	0.241	0.4668	0.0541	0.4412	0.7076	-1.0	0.5008	0.7196	0.4516	0.7719	0.2702	0.6873	0.5329	0.7264	0.4895	0.6679	0.4465	0.6847
No log	7.0	420	9.5030	0.4688	0.7173	0.5119	-1.0	0.2198	0.495	0.0563	0.4558	0.7114	-1.0	0.501	0.7247	0.4693	0.7654	0.4106	0.7164	0.5062	0.7306	0.444	0.6705	0.514	0.6741
No log	8.0	480	9.2783	0.4527	0.7092	0.4843	-1.0	0.2341	0.4809	0.0571	0.4557	0.7083	-1.0	0.508	0.7197	0.4995	0.7513	0.2783	0.7073	0.514	0.7446	0.4676	0.6711	0.5039	0.6671
12.1364	9.0	540	9.2726	0.4532	0.7098	0.4928	-1.0	0.2228	0.4776	0.0552	0.4615	0.7002	-1.0	0.5189	0.7088	0.474	0.7601	0.3484	0.7194	0.5036	0.7074	0.4495	0.6421	0.4904	0.6718
12.1364	10.0	600	9.4426	0.4414	0.6915	0.4731	-1.0	0.2399	0.4709	0.0549	0.4536	0.7086	-1.0	0.4927	0.722	0.4723	0.7689	0.2985	0.6976	0.4893	0.7372	0.4592	0.6653	0.4875	0.6741
12.1364	11.0	660	9.1977	0.4518	0.7056	0.4996	-1.0	0.2518	0.481	0.0559	0.4525	0.7221	-1.0	0.5173	0.7329	0.4765	0.768	0.2624	0.7127	0.5464	0.7421	0.4787	0.6911	0.4951	0.6965
12.1364	12.0	720	9.0903	0.4565	0.7163	0.4956	-1.0	0.2344	0.4867	0.0583	0.4541	0.7133	-1.0	0.5205	0.7233	0.5302	0.7719	0.2953	0.7152	0.5368	0.7314	0.4543	0.6784	0.4659	0.6694
12.1364	13.0	780	9.4333	0.4314	0.6785	0.4583	-1.0	0.2047	0.4704	0.0545	0.4601	0.7121	-1.0	0.5031	0.7246	0.4716	0.7575	0.2749	0.7218	0.5278	0.7264	0.4165	0.6805	0.4663	0.6741
12.1364	14.0	840	9.5343	0.4349	0.6881	0.4659	-1.0	0.2127	0.4744	0.0545	0.4526	0.7109	-1.0	0.5001	0.7233	0.4847	0.7478	0.2956	0.7067	0.5216	0.7405	0.409	0.6774	0.4634	0.6824
12.1364	15.0	900	9.4543	0.4471	0.7028	0.4869	-1.0	0.2276	0.4811	0.0525	0.4545	0.7144	-1.0	0.4954	0.7281	0.4866	0.7667	0.3023	0.7067	0.5383	0.7455	0.4394	0.6674	0.4687	0.6859
12.1364	16.0	960	9.4275	0.4411	0.69	0.472	-1.0	0.2095	0.4784	0.0519	0.4504	0.7231	-1.0	0.4965	0.7367	0.494	0.7645	0.286	0.7218	0.5394	0.7471	0.4258	0.6832	0.4606	0.6988
9.6941	17.0	1020	9.3709	0.4427	0.6984	0.4786	-1.0	0.2025	0.4817	0.0506	0.4541	0.7209	-1.0	0.4912	0.7347	0.4803	0.7588	0.2906	0.7242	0.5383	0.7421	0.4372	0.6853	0.4673	0.6941
9.6941	18.0	1080	9.3703	0.4494	0.7094	0.4842	-1.0	0.2257	0.483	0.0515	0.459	0.7167	-1.0	0.4959	0.7311	0.4898	0.7614	0.3041	0.7067	0.5427	0.7463	0.4501	0.6905	0.4602	0.6788
9.6941	19.0	1140	9.3795	0.4482	0.7061	0.4776	-1.0	0.2204	0.485	0.0526	0.4621	0.7193	-1.0	0.5019	0.7331	0.4906	0.7627	0.3078	0.7121	0.5366	0.7496	0.4429	0.6874	0.4632	0.6847
9.6941	20.0	1200	9.4202	0.4462	0.7009	0.4734	-1.0	0.234	0.4809	0.0501	0.456	0.7177	-1.0	0.4982	0.7316	0.4875	0.7605	0.3124	0.7061	0.5328	0.7537	0.447	0.6895	0.4512	0.6788

Framework versions

• Transformers 4.57.6
• Pytorch 2.10.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2