rtdetr-v2-r50-finetune-20

This model is a fine-tuned version of PekingU/rtdetr_v2_r101vd on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 6.3887
• Map: 0.5454
• Map 50: 0.8889
• Map 75: 0.6098
• Map Small: 0.5063
• Map Medium: 0.6723
• Map Large: -1.0
• Mar 1: 0.3324
• Mar 10: 0.6612
• Mar 100: 0.6971
• Mar Small: 0.6671
• Mar Medium: 0.7736
• Mar Large: -1.0
• Map Artemia: 0.5454
• Mar 100 Artemia: 0.6971

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: 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: 50

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 Artemia	Mar 100 Artemia
No log	1.0	250	12.9182	0.4007	0.7824	0.3621	0.3112	0.5503	-1.0	0.3327	0.5511	0.6078	0.5366	0.7066	-1.0	0.4007	0.6078
160.8266	2.0	500	8.4407	0.466	0.8694	0.4305	0.3955	0.5728	-1.0	0.3654	0.5829	0.6162	0.5672	0.6869	-1.0	0.466	0.6162
160.8266	3.0	750	8.0965	0.4808	0.8799	0.4675	0.4077	0.604	-1.0	0.3816	0.6078	0.6396	0.5876	0.7117	-1.0	0.4808	0.6396
13.6956	4.0	1000	7.9711	0.4667	0.8521	0.4335	0.3982	0.5755	-1.0	0.3751	0.5857	0.6053	0.5441	0.6912	-1.0	0.4667	0.6053
13.6956	5.0	1250	8.5254	0.466	0.8675	0.4555	0.3846	0.5894	-1.0	0.3692	0.596	0.6146	0.5575	0.6934	-1.0	0.466	0.6146
11.8014	6.0	1500	9.0914	0.462	0.8575	0.4305	0.3842	0.6026	-1.0	0.3707	0.6	0.6159	0.5511	0.7051	-1.0	0.462	0.6159
11.8014	7.0	1750	8.0776	0.4663	0.8672	0.4565	0.389	0.5875	-1.0	0.3779	0.5941	0.6059	0.5376	0.6993	-1.0	0.4663	0.6059
10.8013	8.0	2000	8.5113	0.4432	0.8371	0.4265	0.3649	0.5863	-1.0	0.3654	0.5757	0.5841	0.5113	0.6847	-1.0	0.4432	0.5841
10.8013	9.0	2250	8.6219	0.4568	0.8499	0.4356	0.3821	0.5688	-1.0	0.371	0.5623	0.5632	0.4973	0.6547	-1.0	0.4568	0.5632
9.7358	10.0	2500	8.8358	0.4204	0.8112	0.3862	0.3331	0.5802	-1.0	0.3498	0.5567	0.5607	0.4785	0.6745	-1.0	0.4204	0.5607
9.7358	11.0	2750	8.8985	0.3962	0.7731	0.377	0.3044	0.5679	-1.0	0.3477	0.5421	0.5427	0.4618	0.654	-1.0	0.3962	0.5427
9.2244	12.0	3000	8.9430	0.4299	0.82	0.4395	0.3474	0.5725	-1.0	0.3642	0.543	0.5445	0.4629	0.6569	-1.0	0.4299	0.5445
9.2244	13.0	3250	9.9454	0.3928	0.7416	0.3769	0.2936	0.5794	-1.0	0.338	0.552	0.5533	0.4704	0.6672	-1.0	0.3928	0.5533
8.5888	14.0	3500	9.0124	0.4089	0.7652	0.4107	0.3253	0.5727	-1.0	0.3486	0.5408	0.5408	0.4661	0.6431	-1.0	0.4089	0.5408
8.5888	15.0	3750	8.9630	0.4389	0.8295	0.4405	0.3609	0.5757	-1.0	0.3536	0.5424	0.5424	0.4656	0.6482	-1.0	0.4389	0.5424
8.1171	16.0	4000	9.1986	0.4163	0.796	0.3958	0.3279	0.5813	-1.0	0.3539	0.5411	0.5411	0.4538	0.6613	-1.0	0.4163	0.5411
8.1171	17.0	4250	9.3362	0.428	0.8039	0.4197	0.3481	0.579	-1.0	0.3567	0.543	0.543	0.4667	0.6482	-1.0	0.428	0.543
7.6843	18.0	4500	9.1995	0.4129	0.7953	0.3905	0.3238	0.5697	-1.0	0.3573	0.5396	0.5411	0.4586	0.654	-1.0	0.4129	0.5411
7.6843	19.0	4750	9.4599	0.4268	0.8139	0.4005	0.3435	0.5756	-1.0	0.3486	0.5417	0.5439	0.4575	0.6628	-1.0	0.4268	0.5439
7.3836	20.0	5000	9.2880	0.4298	0.7901	0.4378	0.35	0.5777	-1.0	0.3517	0.5502	0.5508	0.472	0.6584	-1.0	0.4298	0.5508
7.3836	21.0	5250	9.7145	0.4225	0.8022	0.4025	0.3456	0.5663	-1.0	0.3539	0.5427	0.5427	0.464	0.6504	-1.0	0.4225	0.5427
7.0963	22.0	5500	9.7255	0.4111	0.7645	0.3942	0.3189	0.5938	-1.0	0.3498	0.5436	0.5436	0.4559	0.6635	-1.0	0.4111	0.5436
7.0963	23.0	5750	9.9880	0.4074	0.7631	0.403	0.3125	0.5838	-1.0	0.3533	0.5467	0.5467	0.4645	0.6599	-1.0	0.4074	0.5467
6.7639	24.0	6000	9.7641	0.4099	0.7791	0.3933	0.3195	0.5777	-1.0	0.343	0.5414	0.5414	0.4586	0.6555	-1.0	0.4099	0.5414
6.7639	25.0	6250	9.9270	0.4105	0.7848	0.3947	0.319	0.5785	-1.0	0.3461	0.5399	0.5399	0.4522	0.6606	-1.0	0.4105	0.5399
6.569	26.0	6500	10.0126	0.4208	0.7922	0.4009	0.3359	0.5744	-1.0	0.3526	0.5402	0.5402	0.4565	0.654	-1.0	0.4208	0.5402
6.569	27.0	6750	10.1916	0.4084	0.7702	0.4031	0.3165	0.5826	-1.0	0.352	0.5439	0.5439	0.4554	0.665	-1.0	0.4084	0.5439
6.3675	28.0	7000	10.4052	0.4105	0.7708	0.391	0.3197	0.5765	-1.0	0.3502	0.5452	0.5452	0.464	0.6569	-1.0	0.4105	0.5452
6.3675	29.0	7250	10.3041	0.4022	0.7596	0.3662	0.3142	0.5776	-1.0	0.3483	0.5498	0.5498	0.4704	0.6584	-1.0	0.4022	0.5498
6.1285	30.0	7500	10.0512	0.4045	0.7449	0.3813	0.3168	0.5864	-1.0	0.3389	0.548	0.548	0.4656	0.6606	-1.0	0.4045	0.548
6.1285	31.0	7750	10.5979	0.3898	0.7268	0.3663	0.2967	0.5813	-1.0	0.3399	0.5449	0.5449	0.4602	0.6613	-1.0	0.3898	0.5449
5.8026	32.0	8000	10.7262	0.3972	0.7381	0.3798	0.3025	0.5783	-1.0	0.3393	0.5461	0.5461	0.4656	0.6562	-1.0	0.3972	0.5461
5.8026	33.0	8250	10.9765	0.3939	0.7418	0.3756	0.2984	0.5852	-1.0	0.343	0.5486	0.5486	0.464	0.6642	-1.0	0.3939	0.5486
5.6222	34.0	8500	10.8479	0.3967	0.7435	0.3787	0.3012	0.5768	-1.0	0.3445	0.5433	0.5433	0.4575	0.6606	-1.0	0.3967	0.5433
5.6222	35.0	8750	11.5637	0.386	0.7268	0.3595	0.291	0.5897	-1.0	0.3461	0.5486	0.5486	0.4629	0.6657	-1.0	0.386	0.5486
5.3595	36.0	9000	10.9348	0.4127	0.7703	0.3855	0.3226	0.587	-1.0	0.3464	0.5439	0.5439	0.4591	0.6599	-1.0	0.4127	0.5439
5.3595	37.0	9250	11.1356	0.4125	0.7663	0.3807	0.3166	0.5837	-1.0	0.3483	0.5452	0.5452	0.4581	0.6642	-1.0	0.4125	0.5452
5.2688	38.0	9500	11.2445	0.3984	0.7351	0.3744	0.3004	0.598	-1.0	0.3449	0.5502	0.5502	0.4597	0.6745	-1.0	0.3984	0.5502
5.2688	39.0	9750	11.5671	0.3895	0.7224	0.37	0.291	0.5897	-1.0	0.3389	0.5467	0.5467	0.4586	0.6672	-1.0	0.3895	0.5467
5.0146	40.0	10000	11.2258	0.4021	0.7475	0.4054	0.3053	0.5921	-1.0	0.3502	0.5445	0.5445	0.4522	0.6708	-1.0	0.4021	0.5445
5.0146	41.0	10250	11.7360	0.3969	0.7345	0.3857	0.2989	0.5941	-1.0	0.3393	0.5492	0.5492	0.4565	0.6759	-1.0	0.3969	0.5492
4.8117	42.0	10500	11.6230	0.4006	0.7372	0.4001	0.3063	0.5919	-1.0	0.3461	0.5461	0.5461	0.4565	0.6686	-1.0	0.4006	0.5461
4.8117	43.0	10750	11.6356	0.4006	0.736	0.3814	0.3026	0.5861	-1.0	0.3445	0.5452	0.5452	0.4565	0.6672	-1.0	0.4006	0.5452
4.6373	44.0	11000	11.6857	0.4068	0.7474	0.3961	0.3097	0.5878	-1.0	0.3517	0.5445	0.5445	0.4565	0.665	-1.0	0.4068	0.5445
4.6373	45.0	11250	11.9533	0.394	0.7215	0.389	0.2944	0.5925	-1.0	0.3427	0.5474	0.5474	0.4581	0.6701	-1.0	0.394	0.5474
4.4848	46.0	11500	12.0236	0.4047	0.7421	0.3978	0.3101	0.5887	-1.0	0.352	0.547	0.547	0.4597	0.6664	-1.0	0.4047	0.547
4.4848	47.0	11750	12.2714	0.3983	0.7282	0.3854	0.3005	0.59	-1.0	0.3461	0.5445	0.5445	0.4559	0.6657	-1.0	0.3983	0.5445
4.2628	48.0	12000	12.4141	0.397	0.7347	0.38	0.2981	0.5888	-1.0	0.3439	0.5442	0.5442	0.4543	0.6672	-1.0	0.397	0.5442
4.2628	49.0	12250	12.4350	0.3957	0.7305	0.3786	0.2945	0.5878	-1.0	0.3445	0.5458	0.5458	0.4581	0.6657	-1.0	0.3957	0.5458
4.1456	50.0	12500	12.3088	0.3975	0.7338	0.38	0.2965	0.588	-1.0	0.3474	0.5461	0.5461	0.4586	0.6657	-1.0	0.3975	0.5461

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.2