rtdetr-v2-r50-finetune-16

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: 5.9871
Map: 0.5911
Map 50: 0.932
Map 75: 0.6857
Map Small: 0.5683
Map Medium: 0.6486
Map Large: -1.0
Mar 1: 0.3375
Mar 10: 0.6563
Mar 100: 0.6693
Mar Small: 0.6541
Mar Medium: 0.708
Mar Large: -1.0
Map Artemia: 0.5911
Mar 100 Artemia: 0.6693

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	14.1391	0.449	0.8326	0.4149	0.3581	0.5706	-1.0	0.3483	0.5969	0.6586	0.5753	0.773	-1.0	0.449	0.6586
98.285	2.0	500	8.6543	0.4742	0.8485	0.5185	0.3998	0.5723	-1.0	0.3604	0.5548	0.6165	0.5505	0.708	-1.0	0.4742	0.6165
98.285	3.0	750	7.8437	0.4785	0.8849	0.4631	0.3931	0.5937	-1.0	0.3598	0.5642	0.6386	0.5769	0.7234	-1.0	0.4785	0.6386
14.5506	4.0	1000	7.5231	0.4916	0.8678	0.4946	0.4147	0.6006	-1.0	0.3732	0.5801	0.6324	0.5758	0.7117	-1.0	0.4916	0.6324
14.5506	5.0	1250	7.5783	0.4941	0.8647	0.4984	0.4281	0.5918	-1.0	0.3804	0.585	0.6393	0.5833	0.7153	-1.0	0.4941	0.6393
13.5987	6.0	1500	7.4946	0.4919	0.8801	0.504	0.4077	0.6068	-1.0	0.3779	0.5879	0.6221	0.5667	0.6985	-1.0	0.4919	0.6221
13.5987	7.0	1750	7.7443	0.4836	0.8721	0.4557	0.3982	0.5926	-1.0	0.3707	0.5819	0.6069	0.5559	0.6774	-1.0	0.4836	0.6069
13.2256	8.0	2000	7.6066	0.4777	0.8738	0.4548	0.3996	0.5868	-1.0	0.3698	0.5555	0.5822	0.5253	0.6613	-1.0	0.4777	0.5822
13.2256	9.0	2250	7.7113	0.4957	0.8712	0.4962	0.4187	0.5953	-1.0	0.381	0.5838	0.6097	0.571	0.6628	-1.0	0.4957	0.6097
12.8339	10.0	2500	7.2662	0.4878	0.8829	0.4568	0.4082	0.5921	-1.0	0.3835	0.5701	0.5897	0.5387	0.6599	-1.0	0.4878	0.5897
12.8339	11.0	2750	7.5464	0.4816	0.8832	0.4757	0.406	0.5902	-1.0	0.3791	0.566	0.5978	0.5565	0.6562	-1.0	0.4816	0.5978
12.614	12.0	3000	7.6500	0.4724	0.8679	0.4303	0.3809	0.5985	-1.0	0.3729	0.5604	0.5854	0.5145	0.6832	-1.0	0.4724	0.5854
12.614	13.0	3250	7.6265	0.4673	0.8635	0.4743	0.3898	0.5782	-1.0	0.3695	0.5642	0.5729	0.514	0.6547	-1.0	0.4673	0.5729
12.3342	14.0	3500	7.4912	0.4868	0.883	0.4945	0.4107	0.5897	-1.0	0.3826	0.5766	0.5916	0.5392	0.6642	-1.0	0.4868	0.5916
12.3342	15.0	3750	7.4118	0.4758	0.8742	0.4576	0.3927	0.5976	-1.0	0.3779	0.5701	0.5769	0.5054	0.6752	-1.0	0.4758	0.5769
11.9633	16.0	4000	8.3546	0.4576	0.8362	0.417	0.3687	0.593	-1.0	0.3688	0.5502	0.5545	0.4742	0.6657	-1.0	0.4576	0.5545
11.9633	17.0	4250	7.5139	0.4684	0.868	0.4543	0.3967	0.5734	-1.0	0.371	0.562	0.5748	0.5204	0.6496	-1.0	0.4684	0.5748
11.6809	18.0	4500	7.6086	0.4679	0.8576	0.4406	0.3861	0.5924	-1.0	0.376	0.5611	0.5664	0.4957	0.665	-1.0	0.4679	0.5664
11.6809	19.0	4750	7.6561	0.4753	0.858	0.4685	0.3973	0.5824	-1.0	0.3717	0.5617	0.5667	0.5027	0.6555	-1.0	0.4753	0.5667
11.5301	20.0	5000	8.0791	0.4499	0.8608	0.4214	0.3684	0.5655	-1.0	0.3682	0.5517	0.5526	0.479	0.654	-1.0	0.4499	0.5526
11.5301	21.0	5250	7.8354	0.4753	0.8712	0.45	0.3991	0.585	-1.0	0.3832	0.566	0.5766	0.5081	0.6708	-1.0	0.4753	0.5766
11.1755	22.0	5500	7.6411	0.4613	0.8548	0.4585	0.3816	0.577	-1.0	0.3701	0.5489	0.5604	0.4968	0.6482	-1.0	0.4613	0.5604
11.1755	23.0	5750	8.0927	0.4636	0.8768	0.4525	0.3858	0.5748	-1.0	0.3776	0.5567	0.5601	0.4919	0.6533	-1.0	0.4636	0.5601
10.7355	24.0	6000	7.6829	0.4545	0.848	0.4217	0.3759	0.5657	-1.0	0.3664	0.5439	0.5464	0.4694	0.6526	-1.0	0.4545	0.5464
10.7355	25.0	6250	7.8003	0.4544	0.8462	0.4388	0.371	0.5819	-1.0	0.3707	0.5474	0.552	0.4699	0.6664	-1.0	0.4544	0.552
10.6738	26.0	6500	7.5889	0.4606	0.8584	0.4429	0.3781	0.575	-1.0	0.3682	0.5498	0.5502	0.4758	0.6518	-1.0	0.4606	0.5502
10.6738	27.0	6750	8.1307	0.4348	0.8291	0.3905	0.3592	0.5515	-1.0	0.3576	0.5327	0.5333	0.4597	0.635	-1.0	0.4348	0.5333
10.3283	28.0	7000	8.5342	0.4499	0.8469	0.4397	0.3703	0.5652	-1.0	0.366	0.5467	0.5505	0.4876	0.6372	-1.0	0.4499	0.5505
10.3283	29.0	7250	8.0470	0.4566	0.8545	0.4431	0.3743	0.5748	-1.0	0.3676	0.5486	0.5511	0.4731	0.6584	-1.0	0.4566	0.5511
10.1173	30.0	7500	8.1874	0.4382	0.8406	0.4155	0.3589	0.5594	-1.0	0.3636	0.5371	0.5377	0.4618	0.6423	-1.0	0.4382	0.5377
10.1173	31.0	7750	8.1895	0.4461	0.8379	0.4093	0.3647	0.5632	-1.0	0.3645	0.5333	0.5352	0.4505	0.6518	-1.0	0.4461	0.5352
9.7767	32.0	8000	7.8706	0.4493	0.8466	0.4259	0.3784	0.5619	-1.0	0.3679	0.5374	0.5374	0.4624	0.6401	-1.0	0.4493	0.5374
9.7767	33.0	8250	8.0987	0.4521	0.8434	0.4117	0.3761	0.5642	-1.0	0.3673	0.5383	0.5399	0.4597	0.6489	-1.0	0.4521	0.5399
9.5652	34.0	8500	8.0303	0.4537	0.8498	0.4155	0.3761	0.5678	-1.0	0.3626	0.5449	0.5464	0.4634	0.6606	-1.0	0.4537	0.5464
9.5652	35.0	8750	8.2618	0.4442	0.8364	0.4518	0.3544	0.5766	-1.0	0.3657	0.5477	0.5486	0.472	0.6533	-1.0	0.4442	0.5486
9.3014	36.0	9000	8.0328	0.4486	0.8537	0.4471	0.3773	0.5585	-1.0	0.3639	0.5433	0.547	0.4753	0.646	-1.0	0.4486	0.547
9.3014	37.0	9250	8.4699	0.4433	0.8468	0.4032	0.3599	0.5661	-1.0	0.3626	0.5389	0.5389	0.4511	0.6591	-1.0	0.4433	0.5389
9.0996	38.0	9500	8.5206	0.4407	0.8446	0.4348	0.355	0.5683	-1.0	0.3611	0.5368	0.5371	0.4581	0.6453	-1.0	0.4407	0.5371
9.0996	39.0	9750	8.3000	0.4455	0.8411	0.422	0.3577	0.5782	-1.0	0.366	0.5399	0.5399	0.4559	0.6547	-1.0	0.4455	0.5399
8.7292	40.0	10000	8.7367	0.4361	0.8407	0.3889	0.3525	0.5596	-1.0	0.3629	0.5349	0.5358	0.4532	0.6489	-1.0	0.4361	0.5358
8.7292	41.0	10250	8.7259	0.4439	0.8356	0.4237	0.3585	0.5707	-1.0	0.3676	0.5455	0.5461	0.4656	0.6577	-1.0	0.4439	0.5461
8.4869	42.0	10500	8.7273	0.4386	0.8442	0.3962	0.3472	0.5707	-1.0	0.3579	0.533	0.5361	0.4484	0.6569	-1.0	0.4386	0.5361
8.4869	43.0	10750	9.2430	0.4335	0.8345	0.3995	0.345	0.5588	-1.0	0.357	0.5318	0.5318	0.4489	0.6453	-1.0	0.4335	0.5318
8.2548	44.0	11000	8.6236	0.4359	0.8343	0.3999	0.3447	0.5695	-1.0	0.3607	0.5299	0.5299	0.4382	0.6547	-1.0	0.4359	0.5299
8.2548	45.0	11250	8.6935	0.439	0.8429	0.4106	0.3659	0.5548	-1.0	0.3589	0.5361	0.5377	0.4613	0.6423	-1.0	0.439	0.5377
7.9558	46.0	11500	8.9859	0.435	0.83	0.4006	0.3482	0.5604	-1.0	0.367	0.5318	0.5318	0.4484	0.6453	-1.0	0.435	0.5318
7.9558	47.0	11750	9.0123	0.4331	0.8294	0.3885	0.3524	0.5518	-1.0	0.3648	0.5368	0.5368	0.4602	0.6409	-1.0	0.4331	0.5368
7.7147	48.0	12000	9.3183	0.4305	0.832	0.3768	0.3454	0.5593	-1.0	0.3586	0.5349	0.5349	0.4511	0.6489	-1.0	0.4305	0.5349
7.7147	49.0	12250	8.4032	0.446	0.8439	0.4356	0.3623	0.5742	-1.0	0.3614	0.5396	0.5396	0.4532	0.6577	-1.0	0.446	0.5396
7.5154	50.0	12500	9.0939	0.4249	0.8357	0.3687	0.3393	0.554	-1.0	0.3573	0.5268	0.5268	0.4409	0.6453	-1.0	0.4249	0.5268

Framework versions

Transformers 4.57.1
Pytorch 2.8.0+cu128
Datasets 4.2.0
Tokenizers 0.22.2

dariacuna/rtdetr-v2-r50-finetune-16

rtdetr-v2-r50-finetune-16

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

为此模型做种