rtdetr-v2-r50-finetune-14

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: 6.3613
Map: 0.5443
Map 50: 0.8447
Map 75: 0.6384
Map Small: 0.5306
Map Medium: 0.596
Map Large: -1.0
Mar 1: 0.3398
Mar 10: 0.6722
Mar 100: 0.7142
Mar Small: 0.6887
Mar Medium: 0.7793
Mar Large: -1.0
Map Artemia: 0.5443
Mar 100 Artemia: 0.7142

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	13.5415	0.206	0.387	0.2017	0.1328	0.366	-1.0	0.2801	0.5888	0.6704	0.5909	0.7803	-1.0	0.206	0.6704
177.3005	2.0	500	8.5772	0.4058	0.7589	0.3949	0.3226	0.5714	-1.0	0.3511	0.5832	0.6399	0.5925	0.7058	-1.0	0.4058	0.6399
177.3005	3.0	750	8.2305	0.4767	0.8624	0.5058	0.3872	0.6051	-1.0	0.3785	0.5866	0.6393	0.5892	0.7088	-1.0	0.4767	0.6393
13.9362	4.0	1000	8.2316	0.4547	0.8408	0.4242	0.3712	0.5847	-1.0	0.3617	0.5748	0.5891	0.5134	0.6942	-1.0	0.4547	0.5891
13.9362	5.0	1250	8.5846	0.4401	0.8194	0.4179	0.3505	0.5656	-1.0	0.3648	0.5773	0.6022	0.5565	0.665	-1.0	0.4401	0.6022
12.2718	6.0	1500	8.0097	0.4433	0.8223	0.4264	0.3636	0.575	-1.0	0.3592	0.5857	0.6106	0.5634	0.6752	-1.0	0.4433	0.6106
12.2718	7.0	1750	8.1912	0.4436	0.8446	0.3927	0.3765	0.5419	-1.0	0.3589	0.5436	0.553	0.4957	0.6321	-1.0	0.4436	0.553
11.3697	8.0	2000	8.4556	0.4616	0.8624	0.4308	0.3824	0.5826	-1.0	0.3698	0.571	0.576	0.5086	0.6686	-1.0	0.4616	0.576
11.3697	9.0	2250	8.2432	0.4541	0.8653	0.3952	0.3783	0.57	-1.0	0.3685	0.5545	0.5558	0.4855	0.6526	-1.0	0.4541	0.5558
10.4006	10.0	2500	8.3066	0.4558	0.8367	0.449	0.3807	0.5756	-1.0	0.3539	0.5645	0.5651	0.4995	0.6555	-1.0	0.4558	0.5651
10.4006	11.0	2750	8.2486	0.4491	0.8487	0.4281	0.3678	0.5707	-1.0	0.3688	0.5604	0.562	0.486	0.6664	-1.0	0.4491	0.562
9.8687	12.0	3000	8.2646	0.4546	0.8335	0.436	0.3703	0.582	-1.0	0.367	0.5583	0.5583	0.4828	0.662	-1.0	0.4546	0.5583
9.8687	13.0	3250	8.5520	0.4524	0.8333	0.4327	0.3695	0.5822	-1.0	0.3654	0.5539	0.5539	0.4753	0.6613	-1.0	0.4524	0.5539
9.3097	14.0	3500	8.9756	0.4235	0.808	0.3919	0.3374	0.5687	-1.0	0.3495	0.5583	0.5586	0.4849	0.6599	-1.0	0.4235	0.5586
9.3097	15.0	3750	8.9380	0.4177	0.7756	0.3696	0.3316	0.5572	-1.0	0.3436	0.547	0.547	0.4694	0.654	-1.0	0.4177	0.547
8.7227	16.0	4000	8.6756	0.4357	0.8249	0.3962	0.3537	0.5657	-1.0	0.3464	0.5439	0.5439	0.4581	0.662	-1.0	0.4357	0.5439
8.7227	17.0	4250	8.9631	0.4254	0.8103	0.3896	0.335	0.5808	-1.0	0.3682	0.5567	0.5567	0.4817	0.6584	-1.0	0.4254	0.5567
8.1899	18.0	4500	8.6008	0.4369	0.8282	0.4067	0.3537	0.5762	-1.0	0.3564	0.5561	0.5561	0.478	0.6635	-1.0	0.4369	0.5561
8.1899	19.0	4750	8.5243	0.4481	0.8248	0.4855	0.3584	0.5929	-1.0	0.3583	0.5567	0.5567	0.471	0.6745	-1.0	0.4481	0.5567
7.9415	20.0	5000	8.9514	0.4166	0.774	0.3988	0.3277	0.5808	-1.0	0.3427	0.5505	0.5505	0.4774	0.6504	-1.0	0.4166	0.5505
7.9415	21.0	5250	9.1006	0.4106	0.7747	0.3911	0.3198	0.5723	-1.0	0.3386	0.5495	0.5495	0.4769	0.6504	-1.0	0.4106	0.5495
7.5377	22.0	5500	9.2048	0.4165	0.7654	0.3995	0.3206	0.5873	-1.0	0.3464	0.5517	0.5517	0.4694	0.6657	-1.0	0.4165	0.5517
7.5377	23.0	5750	9.3098	0.4099	0.7714	0.3795	0.3244	0.568	-1.0	0.3389	0.552	0.552	0.4849	0.6453	-1.0	0.4099	0.552
7.1178	24.0	6000	9.3530	0.4179	0.7775	0.4008	0.3286	0.5813	-1.0	0.3442	0.5526	0.5526	0.4758	0.6584	-1.0	0.4179	0.5526
7.1178	25.0	6250	9.1295	0.4073	0.7647	0.3846	0.3078	0.5757	-1.0	0.3346	0.5483	0.5483	0.4672	0.6599	-1.0	0.4073	0.5483
6.9818	26.0	6500	9.3428	0.398	0.7421	0.3669	0.3034	0.5751	-1.0	0.3358	0.5486	0.5486	0.4683	0.6591	-1.0	0.398	0.5486
6.9818	27.0	6750	9.2751	0.395	0.743	0.3741	0.2964	0.5806	-1.0	0.3255	0.5452	0.5452	0.457	0.6672	-1.0	0.395	0.5452
6.6702	28.0	7000	9.2444	0.406	0.7597	0.3723	0.3141	0.5766	-1.0	0.3414	0.5511	0.5511	0.4769	0.6533	-1.0	0.406	0.5511
6.6702	29.0	7250	10.0855	0.3552	0.6654	0.3324	0.25	0.5763	-1.0	0.2981	0.5483	0.5483	0.472	0.6533	-1.0	0.3552	0.5483
6.3734	30.0	7500	9.9425	0.3723	0.6927	0.3493	0.2703	0.572	-1.0	0.3106	0.5455	0.5455	0.4688	0.6511	-1.0	0.3723	0.5455
6.3734	31.0	7750	9.9158	0.3807	0.7156	0.3539	0.2723	0.5826	-1.0	0.319	0.5474	0.5474	0.4629	0.6642	-1.0	0.3807	0.5474
6.1584	32.0	8000	9.8123	0.3782	0.6993	0.3572	0.2704	0.5876	-1.0	0.3212	0.5502	0.5502	0.4613	0.6723	-1.0	0.3782	0.5502
6.1584	33.0	8250	10.0163	0.3898	0.7245	0.3604	0.2858	0.5762	-1.0	0.3243	0.5436	0.5436	0.4629	0.654	-1.0	0.3898	0.5436
5.9568	34.0	8500	10.0463	0.3813	0.7001	0.3683	0.2749	0.5873	-1.0	0.3103	0.5492	0.5492	0.4645	0.6657	-1.0	0.3813	0.5492
5.9568	35.0	8750	10.2176	0.3723	0.6889	0.3488	0.2644	0.5864	-1.0	0.3202	0.5502	0.5502	0.4683	0.6628	-1.0	0.3723	0.5502
5.6813	36.0	9000	10.2861	0.3761	0.6971	0.3573	0.2691	0.5803	-1.0	0.3115	0.5492	0.5492	0.4699	0.6584	-1.0	0.3761	0.5492
5.6813	37.0	9250	10.5426	0.3707	0.6825	0.3462	0.2639	0.5839	-1.0	0.304	0.5495	0.5495	0.4661	0.6642	-1.0	0.3707	0.5495
5.5154	38.0	9500	10.5989	0.3645	0.6808	0.342	0.2593	0.5721	-1.0	0.3006	0.5445	0.5445	0.4645	0.6547	-1.0	0.3645	0.5445
5.5154	39.0	9750	10.8408	0.3801	0.7006	0.3563	0.2753	0.5822	-1.0	0.3206	0.5474	0.5474	0.4624	0.6642	-1.0	0.3801	0.5474
5.2204	40.0	10000	10.6278	0.3874	0.7115	0.3689	0.2839	0.5764	-1.0	0.3218	0.5445	0.5445	0.4613	0.6591	-1.0	0.3874	0.5445
5.2204	41.0	10250	10.7575	0.3821	0.7004	0.368	0.2758	0.5852	-1.0	0.3227	0.5486	0.5486	0.471	0.6555	-1.0	0.3821	0.5486
5.0211	42.0	10500	10.8551	0.3825	0.7054	0.3587	0.2757	0.5811	-1.0	0.3227	0.5483	0.5483	0.4677	0.6591	-1.0	0.3825	0.5483
5.0211	43.0	10750	10.9142	0.3816	0.7089	0.3502	0.2744	0.5787	-1.0	0.3268	0.5458	0.5458	0.4613	0.662	-1.0	0.3816	0.5458
4.8453	44.0	11000	11.0659	0.3771	0.6947	0.3545	0.2686	0.5832	-1.0	0.3137	0.5461	0.5461	0.4629	0.6606	-1.0	0.3771	0.5461
4.8453	45.0	11250	11.1620	0.3785	0.6975	0.3654	0.2724	0.58	-1.0	0.3193	0.5486	0.5486	0.4677	0.6599	-1.0	0.3785	0.5486
4.6892	46.0	11500	11.3102	0.3787	0.6972	0.3638	0.2721	0.5804	-1.0	0.3171	0.5477	0.5477	0.4683	0.6569	-1.0	0.3787	0.5477
4.6892	47.0	11750	11.4467	0.3798	0.7042	0.3592	0.2765	0.5798	-1.0	0.3181	0.5455	0.5455	0.4651	0.6562	-1.0	0.3798	0.5455
4.4631	48.0	12000	11.4220	0.3796	0.7016	0.3659	0.2761	0.5799	-1.0	0.3218	0.5474	0.5474	0.4683	0.6562	-1.0	0.3796	0.5474
4.4631	49.0	12250	11.4238	0.3837	0.7067	0.3745	0.2762	0.5841	-1.0	0.3196	0.5474	0.5474	0.4651	0.6606	-1.0	0.3837	0.5474
4.3455	50.0	12500	11.4721	0.3813	0.7026	0.3684	0.2729	0.5824	-1.0	0.3168	0.5464	0.5464	0.4651	0.6584	-1.0	0.3813	0.5464

Framework versions

Transformers 4.57.1
Pytorch 2.8.0+cu128
Datasets 4.2.0
Tokenizers 0.22.2

dariacuna/rtdetr-v2-r50-finetune-14

rtdetr-v2-r50-finetune-14

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Seed this model