rtdetr-v2-r34-finetune-22

This model is a fine-tuned version of PekingU/rtdetr_v2_r34vd on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 5.0884
Map: 0.5388
Map 50: 0.8433
Map 75: 0.6187
Map Small: 0.491
Map Medium: 0.6567
Map Large: -1.0
Mar 1: 0.3294
Mar 10: 0.6537
Mar 100: 0.6896
Mar Small: 0.6703
Mar Medium: 0.7391
Mar Large: -1.0
Map Artemia: 0.5388
Mar 100 Artemia: 0.6896

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	13.3221	0.3224	0.6104	0.3093	0.2291	0.5225	-1.0	0.3028	0.5667	0.6187	0.5333	0.735	-1.0	0.3224	0.6187
200.0611	2.0	500	6.5405	0.4053	0.782	0.3699	0.3144	0.557	-1.0	0.3442	0.5617	0.6368	0.5898	0.7022	-1.0	0.4053	0.6368
200.0611	3.0	750	6.5873	0.4254	0.7679	0.3786	0.3153	0.6004	-1.0	0.3632	0.5922	0.6595	0.6075	0.7307	-1.0	0.4254	0.6595
10.4117	4.0	1000	6.2695	0.436	0.7958	0.4268	0.3326	0.5969	-1.0	0.3601	0.5947	0.6302	0.5871	0.6905	-1.0	0.436	0.6302
10.4117	5.0	1250	6.4432	0.3917	0.7557	0.3527	0.2956	0.5821	-1.0	0.3489	0.576	0.6299	0.5973	0.6759	-1.0	0.3917	0.6299
9.2935	6.0	1500	6.5503	0.3634	0.6868	0.3494	0.2658	0.5824	-1.0	0.3358	0.586	0.6075	0.5527	0.6839	-1.0	0.3634	0.6075
9.2935	7.0	1750	6.3442	0.4099	0.7757	0.3383	0.3222	0.5634	-1.0	0.3567	0.572	0.5841	0.5317	0.6569	-1.0	0.4099	0.5841
8.7572	8.0	2000	6.6003	0.3646	0.7119	0.3303	0.2647	0.5776	-1.0	0.3234	0.5754	0.5872	0.5253	0.6745	-1.0	0.3646	0.5872
8.7572	9.0	2250	6.6029	0.3579	0.6999	0.3023	0.2627	0.5718	-1.0	0.3252	0.5713	0.5816	0.5194	0.6686	-1.0	0.3579	0.5816
8.1434	10.0	2500	6.5649	0.4028	0.7703	0.3497	0.3152	0.5788	-1.0	0.3452	0.5869	0.5928	0.5349	0.6737	-1.0	0.4028	0.5928
8.1434	11.0	2750	6.9758	0.368	0.7204	0.3357	0.2646	0.575	-1.0	0.3327	0.5673	0.5723	0.4973	0.6766	-1.0	0.368	0.5723
7.804	12.0	3000	7.1631	0.3319	0.6345	0.292	0.2265	0.5823	-1.0	0.3065	0.5757	0.585	0.5124	0.6861	-1.0	0.3319	0.585
7.804	13.0	3250	6.9533	0.3594	0.6917	0.3205	0.2658	0.5724	-1.0	0.3402	0.5707	0.5763	0.5081	0.6715	-1.0	0.3594	0.5763
7.381	14.0	3500	7.2171	0.3323	0.665	0.2857	0.2351	0.5553	-1.0	0.3106	0.5567	0.5598	0.4989	0.6453	-1.0	0.3323	0.5598
7.381	15.0	3750	7.1963	0.3347	0.6599	0.2944	0.244	0.5658	-1.0	0.3125	0.5589	0.5611	0.4892	0.6613	-1.0	0.3347	0.5611
7.0061	16.0	4000	6.9970	0.3826	0.7418	0.3294	0.2901	0.5673	-1.0	0.3424	0.5607	0.5654	0.4952	0.6642	-1.0	0.3826	0.5654
7.0061	17.0	4250	7.2147	0.3679	0.7125	0.3379	0.2716	0.5803	-1.0	0.3414	0.5607	0.5617	0.493	0.6569	-1.0	0.3679	0.5617
6.6717	18.0	4500	7.1103	0.3717	0.7277	0.3285	0.2855	0.5435	-1.0	0.3355	0.5486	0.5508	0.4935	0.6307	-1.0	0.3717	0.5508
6.6717	19.0	4750	7.2940	0.3573	0.6839	0.3293	0.2633	0.5665	-1.0	0.3178	0.5673	0.5698	0.5113	0.6511	-1.0	0.3573	0.5698
6.3706	20.0	5000	7.6389	0.3188	0.6285	0.2765	0.2308	0.5541	-1.0	0.3097	0.5611	0.5636	0.5054	0.6445	-1.0	0.3188	0.5636
6.3706	21.0	5250	7.5861	0.3519	0.6815	0.3186	0.252	0.5673	-1.0	0.3224	0.5474	0.5492	0.4699	0.6591	-1.0	0.3519	0.5492
6.143	22.0	5500	7.4499	0.336	0.67	0.2985	0.248	0.5452	-1.0	0.3069	0.548	0.5502	0.4812	0.6467	-1.0	0.336	0.5502
6.143	23.0	5750	7.4256	0.3324	0.6362	0.2893	0.2409	0.5686	-1.0	0.3137	0.5592	0.5617	0.4866	0.6664	-1.0	0.3324	0.5617
5.9266	24.0	6000	7.4874	0.3532	0.6797	0.3189	0.2614	0.5603	-1.0	0.3321	0.5576	0.5601	0.4946	0.6518	-1.0	0.3532	0.5601
5.9266	25.0	6250	7.5070	0.3506	0.6737	0.3225	0.2518	0.566	-1.0	0.3268	0.5508	0.5536	0.4763	0.6613	-1.0	0.3506	0.5536
5.6974	26.0	6500	7.6325	0.327	0.6254	0.2853	0.2352	0.5664	-1.0	0.3262	0.5589	0.5604	0.493	0.6547	-1.0	0.327	0.5604
5.6974	27.0	6750	7.8022	0.3408	0.6595	0.3035	0.246	0.5586	-1.0	0.3308	0.5452	0.547	0.4742	0.6489	-1.0	0.3408	0.547
5.5361	28.0	7000	7.6712	0.3416	0.6567	0.3064	0.2488	0.5674	-1.0	0.329	0.5505	0.5511	0.4774	0.654	-1.0	0.3416	0.5511
5.5361	29.0	7250	7.7049	0.3422	0.6736	0.2931	0.2512	0.5493	-1.0	0.3259	0.5349	0.5355	0.4651	0.6343	-1.0	0.3422	0.5355
5.3672	30.0	7500	7.8047	0.3412	0.6611	0.2915	0.2432	0.5611	-1.0	0.3212	0.5474	0.548	0.4737	0.6518	-1.0	0.3412	0.548
5.3672	31.0	7750	7.8350	0.3466	0.6802	0.2966	0.2522	0.5459	-1.0	0.328	0.5389	0.5402	0.4715	0.6358	-1.0	0.3466	0.5402
5.195	32.0	8000	7.8076	0.3332	0.6409	0.2883	0.2331	0.569	-1.0	0.3181	0.5539	0.5539	0.4828	0.6533	-1.0	0.3332	0.5539
5.195	33.0	8250	7.8677	0.3343	0.6377	0.2959	0.2311	0.5659	-1.0	0.3174	0.5502	0.5511	0.4758	0.6562	-1.0	0.3343	0.5511
5.0533	34.0	8500	8.0200	0.328	0.6337	0.2811	0.2293	0.5615	-1.0	0.3171	0.5464	0.5467	0.471	0.6526	-1.0	0.328	0.5467
5.0533	35.0	8750	8.0928	0.3351	0.6494	0.2865	0.2354	0.5637	-1.0	0.3171	0.5427	0.543	0.4645	0.6526	-1.0	0.3351	0.543
4.9168	36.0	9000	7.9054	0.3643	0.6925	0.3218	0.2652	0.567	-1.0	0.3402	0.5483	0.5492	0.472	0.6569	-1.0	0.3643	0.5492
4.9168	37.0	9250	8.1647	0.3344	0.6368	0.294	0.23	0.5727	-1.0	0.3184	0.5551	0.5561	0.4823	0.6591	-1.0	0.3344	0.5561
4.7956	38.0	9500	8.1079	0.3461	0.6781	0.304	0.2449	0.5643	-1.0	0.3315	0.5393	0.5393	0.4591	0.6511	-1.0	0.3461	0.5393
4.7956	39.0	9750	8.2293	0.34	0.6446	0.2985	0.237	0.5652	-1.0	0.3349	0.5483	0.5483	0.4677	0.6606	-1.0	0.34	0.5483
4.6033	40.0	10000	8.1899	0.3448	0.662	0.2937	0.2476	0.5586	-1.0	0.3315	0.5386	0.5386	0.4661	0.6401	-1.0	0.3448	0.5386
4.6033	41.0	10250	8.1783	0.3498	0.6652	0.3126	0.2519	0.557	-1.0	0.3389	0.5411	0.5411	0.4683	0.6431	-1.0	0.3498	0.5411
4.4986	42.0	10500	8.4910	0.338	0.6519	0.2859	0.242	0.5606	-1.0	0.3277	0.5417	0.5417	0.4613	0.654	-1.0	0.338	0.5417
4.4986	43.0	10750	8.4450	0.3378	0.6569	0.2926	0.241	0.5634	-1.0	0.3302	0.5427	0.5427	0.4656	0.6504	-1.0	0.3378	0.5427
4.3994	44.0	11000	8.3464	0.3506	0.6722	0.2953	0.2589	0.5579	-1.0	0.343	0.5445	0.5445	0.4715	0.6467	-1.0	0.3506	0.5445
4.3994	45.0	11250	8.3953	0.3461	0.6698	0.2994	0.2482	0.5609	-1.0	0.3458	0.5414	0.5414	0.4613	0.6533	-1.0	0.3461	0.5414
4.3239	46.0	11500	8.5365	0.3434	0.6633	0.3022	0.244	0.5634	-1.0	0.3383	0.543	0.543	0.4656	0.6511	-1.0	0.3434	0.543
4.3239	47.0	11750	8.6661	0.3335	0.6513	0.2895	0.239	0.5583	-1.0	0.3315	0.5399	0.5399	0.4634	0.6467	-1.0	0.3335	0.5399
4.1758	48.0	12000	8.6470	0.3399	0.6565	0.2928	0.2422	0.5618	-1.0	0.3358	0.5424	0.5424	0.464	0.6518	-1.0	0.3399	0.5424
4.1758	49.0	12250	8.5784	0.3428	0.6683	0.2978	0.2461	0.5581	-1.0	0.334	0.5414	0.5414	0.4624	0.6518	-1.0	0.3428	0.5414
4.1144	50.0	12500	8.5991	0.3427	0.6662	0.297	0.246	0.5603	-1.0	0.3368	0.5417	0.5417	0.4629	0.6518	-1.0	0.3427	0.5417

Framework versions

Transformers 4.57.1
Pytorch 2.8.0+cu128
Datasets 4.2.0
Tokenizers 0.22.2

dariacuna/rtdetr-v2-r34-finetune-22

rtdetr-v2-r34-finetune-22

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Seed this model