rtdetr-v2-r101-finetune-27

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: 6.4728
Map: 0.5378
Map 50: 0.8765
Map 75: 0.6023
Map Small: 0.4923
Map Medium: 0.6727
Map Large: -1.0
Mar 1: 0.3239
Mar 10: 0.6696
Mar 100: 0.7091
Mar Small: 0.6766
Mar Medium: 0.792
Mar Large: -1.0
Map Artemia: 0.5378
Mar 100 Artemia: 0.7091

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	9.8250	0.4233	0.8064	0.3963	0.3408	0.5617	-1.0	0.3526	0.5807	0.6305	0.5656	0.7212	-1.0	0.4233	0.6305
131.9463	2.0	500	8.2747	0.4667	0.8571	0.467	0.3882	0.5891	-1.0	0.3586	0.5997	0.638	0.5806	0.719	-1.0	0.4667	0.638
131.9463	3.0	750	8.1930	0.4748	0.8776	0.4682	0.3939	0.599	-1.0	0.3639	0.5879	0.648	0.5866	0.7328	-1.0	0.4748	0.648
13.5316	4.0	1000	8.0508	0.4249	0.761	0.4063	0.4124	0.5774	-1.0	0.3748	0.591	0.6184	0.5731	0.681	-1.0	0.4249	0.6184
13.5316	5.0	1250	8.3392	0.4486	0.8569	0.417	0.3762	0.5648	-1.0	0.3645	0.5614	0.5935	0.5414	0.665	-1.0	0.4486	0.5935
11.6588	6.0	1500	8.1256	0.4641	0.8655	0.4312	0.3871	0.5938	-1.0	0.3707	0.5981	0.6178	0.5629	0.6934	-1.0	0.4641	0.6178
11.6588	7.0	1750	8.2882	0.4679	0.8589	0.4616	0.3933	0.5848	-1.0	0.3707	0.5875	0.595	0.5387	0.6723	-1.0	0.4679	0.595
10.6185	8.0	2000	8.5228	0.4484	0.8493	0.3911	0.371	0.5738	-1.0	0.3648	0.5645	0.5788	0.5167	0.665	-1.0	0.4484	0.5788
10.6185	9.0	2250	8.6664	0.4566	0.8465	0.4452	0.3806	0.5819	-1.0	0.3713	0.5729	0.5766	0.514	0.662	-1.0	0.4566	0.5766
9.756	10.0	2500	8.7909	0.4111	0.7735	0.3792	0.3245	0.584	-1.0	0.3374	0.5763	0.5816	0.5231	0.6613	-1.0	0.4111	0.5816
9.756	11.0	2750	8.8580	0.4283	0.8231	0.4103	0.3434	0.5812	-1.0	0.3433	0.5589	0.5651	0.4962	0.6599	-1.0	0.4283	0.5651
9.189	12.0	3000	8.9836	0.4119	0.8188	0.3734	0.3291	0.5749	-1.0	0.3383	0.5492	0.5502	0.4715	0.6584	-1.0	0.4119	0.5502
9.189	13.0	3250	9.6157	0.3735	0.7107	0.3538	0.2875	0.587	-1.0	0.3221	0.5676	0.5698	0.5027	0.6613	-1.0	0.3735	0.5698
8.5802	14.0	3500	9.4162	0.3693	0.7056	0.3355	0.2761	0.5812	-1.0	0.3283	0.5701	0.5735	0.5065	0.6672	-1.0	0.3693	0.5735
8.5802	15.0	3750	9.0857	0.3971	0.7832	0.3681	0.3019	0.5829	-1.0	0.3352	0.5564	0.5626	0.4876	0.6657	-1.0	0.3971	0.5626
8.2245	16.0	4000	9.4847	0.4007	0.7838	0.3502	0.3083	0.5625	-1.0	0.333	0.5343	0.5346	0.4505	0.6511	-1.0	0.4007	0.5346
8.2245	17.0	4250	9.5033	0.3978	0.745	0.4007	0.3018	0.5882	-1.0	0.3343	0.5654	0.566	0.4995	0.6569	-1.0	0.3978	0.566
7.7021	18.0	4500	9.4326	0.3971	0.7523	0.3566	0.3028	0.586	-1.0	0.3302	0.5561	0.5579	0.4855	0.6584	-1.0	0.3971	0.5579
7.7021	19.0	4750	9.4041	0.401	0.7585	0.373	0.3062	0.5769	-1.0	0.3271	0.5555	0.5564	0.4828	0.6577	-1.0	0.401	0.5564
7.3848	20.0	5000	9.4079	0.3976	0.7487	0.374	0.2972	0.5895	-1.0	0.3302	0.5561	0.5576	0.4839	0.6591	-1.0	0.3976	0.5576
7.3848	21.0	5250	9.8852	0.3975	0.7556	0.3599	0.2969	0.5812	-1.0	0.3249	0.5449	0.5449	0.464	0.6562	-1.0	0.3975	0.5449
7.0775	22.0	5500	9.3829	0.4013	0.7527	0.4	0.3096	0.589	-1.0	0.3346	0.5498	0.5498	0.4753	0.6518	-1.0	0.4013	0.5498
7.0775	23.0	5750	9.8199	0.3774	0.702	0.3634	0.2737	0.5902	-1.0	0.324	0.5604	0.5604	0.4844	0.6657	-1.0	0.3774	0.5604
6.7373	24.0	6000	9.8633	0.3655	0.6871	0.3512	0.259	0.585	-1.0	0.3006	0.5551	0.5551	0.4726	0.6693	-1.0	0.3655	0.5551
6.7373	25.0	6250	10.1601	0.3761	0.7098	0.3495	0.2687	0.5732	-1.0	0.3196	0.5458	0.5461	0.4645	0.6584	-1.0	0.3761	0.5461
6.501	26.0	6500	10.0421	0.3912	0.7274	0.3923	0.2869	0.589	-1.0	0.3252	0.5567	0.5567	0.4785	0.6642	-1.0	0.3912	0.5567
6.501	27.0	6750	9.8351	0.3845	0.7291	0.354	0.2809	0.5804	-1.0	0.3181	0.5449	0.5449	0.4629	0.6577	-1.0	0.3845	0.5449
6.2606	28.0	7000	10.2445	0.3657	0.6824	0.3504	0.2583	0.5819	-1.0	0.3078	0.5533	0.5533	0.4763	0.6591	-1.0	0.3657	0.5533
6.2606	29.0	7250	10.3503	0.3681	0.6877	0.3596	0.2585	0.5851	-1.0	0.3115	0.5467	0.5467	0.4667	0.6569	-1.0	0.3681	0.5467
6.0109	30.0	7500	10.1341	0.3764	0.6983	0.3694	0.2726	0.5844	-1.0	0.3224	0.5542	0.5542	0.4774	0.6606	-1.0	0.3764	0.5542
6.0109	31.0	7750	10.1786	0.3995	0.7504	0.3732	0.2964	0.5896	-1.0	0.3305	0.5489	0.5489	0.4634	0.6664	-1.0	0.3995	0.5489
5.7401	32.0	8000	10.2363	0.3924	0.7253	0.3833	0.2885	0.5916	-1.0	0.3215	0.5523	0.5523	0.4747	0.6599	-1.0	0.3924	0.5523
5.7401	33.0	8250	10.5037	0.3611	0.6713	0.3341	0.2549	0.5893	-1.0	0.3	0.5548	0.5548	0.4774	0.6613	-1.0	0.3611	0.5548
5.6219	34.0	8500	10.5663	0.3825	0.7091	0.3609	0.2796	0.6003	-1.0	0.3159	0.5505	0.5505	0.4688	0.6628	-1.0	0.3825	0.5505
5.6219	35.0	8750	10.4490	0.3718	0.695	0.3583	0.2613	0.5975	-1.0	0.3131	0.5523	0.5523	0.4704	0.665	-1.0	0.3718	0.5523
5.4013	36.0	9000	10.4777	0.3825	0.7095	0.3687	0.2797	0.5943	-1.0	0.3212	0.5542	0.5542	0.4812	0.6547	-1.0	0.3825	0.5542
5.4013	37.0	9250	10.7519	0.3725	0.6898	0.3536	0.2638	0.5945	-1.0	0.3174	0.5551	0.5551	0.4774	0.662	-1.0	0.3725	0.5551
5.2497	38.0	9500	10.8456	0.3675	0.6924	0.3515	0.2576	0.597	-1.0	0.3159	0.5548	0.5548	0.4731	0.6679	-1.0	0.3675	0.5548
5.2497	39.0	9750	10.9667	0.3679	0.6853	0.3567	0.261	0.5931	-1.0	0.3106	0.5489	0.5489	0.4715	0.6555	-1.0	0.3679	0.5489
4.9763	40.0	10000	10.8084	0.3927	0.7288	0.3694	0.292	0.5941	-1.0	0.3262	0.5486	0.5486	0.4672	0.6606	-1.0	0.3927	0.5486
4.9763	41.0	10250	10.8820	0.3906	0.7272	0.3698	0.287	0.5927	-1.0	0.3262	0.5492	0.5495	0.4683	0.6613	-1.0	0.3906	0.5495
4.7855	42.0	10500	10.9702	0.3879	0.7203	0.3705	0.2825	0.596	-1.0	0.3199	0.5502	0.5508	0.4677	0.665	-1.0	0.3879	0.5508
4.7855	43.0	10750	11.0573	0.387	0.7129	0.3674	0.2818	0.5932	-1.0	0.3246	0.5523	0.553	0.4731	0.6628	-1.0	0.387	0.553
4.6116	44.0	11000	11.2676	0.3757	0.6976	0.3503	0.2666	0.5932	-1.0	0.3178	0.5498	0.5502	0.4699	0.6606	-1.0	0.3757	0.5502
4.6116	45.0	11250	11.5526	0.3666	0.6843	0.3528	0.2567	0.5934	-1.0	0.3059	0.5523	0.5523	0.471	0.6642	-1.0	0.3666	0.5523
4.4386	46.0	11500	11.5747	0.375	0.699	0.3508	0.2651	0.592	-1.0	0.315	0.5505	0.5505	0.4715	0.6591	-1.0	0.375	0.5505
4.4386	47.0	11750	11.7096	0.3733	0.6919	0.3599	0.2639	0.5968	-1.0	0.3174	0.5533	0.5533	0.4731	0.6635	-1.0	0.3733	0.5533
4.2401	48.0	12000	11.7366	0.3753	0.6924	0.3593	0.265	0.5935	-1.0	0.3137	0.5505	0.5505	0.4699	0.6613	-1.0	0.3753	0.5505
4.2401	49.0	12250	11.7011	0.3805	0.6991	0.3564	0.2733	0.5929	-1.0	0.3143	0.5514	0.5514	0.472	0.6606	-1.0	0.3805	0.5514
4.1303	50.0	12500	11.7229	0.38	0.6999	0.3561	0.2732	0.5932	-1.0	0.3134	0.5508	0.5508	0.472	0.6591	-1.0	0.38	0.5508

Framework versions

Transformers 4.57.1
Pytorch 2.8.0+cu128
Datasets 4.2.0
Tokenizers 0.22.2

dariacuna/rtdetr-v2-r101-finetune-27

rtdetr-v2-r101-finetune-27

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Seed this model