rtdetr-v2-r50-finetune-15

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: 6.0657
Map: 0.5656
Map 50: 0.8817
Map 75: 0.6676
Map Small: 0.5345
Map Medium: 0.6555
Map Large: -1.0
Mar 1: 0.3388
Mar 10: 0.6437
Mar 100: 0.6803
Mar Small: 0.6613
Mar Medium: 0.7287
Mar Large: -1.0
Map Artemia: 0.5656
Mar 100 Artemia: 0.6803

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	10.8091	0.461	0.8168	0.4878	0.3588	0.5839	-1.0	0.3598	0.5782	0.6249	0.5592	0.7144	-1.0	0.461	0.6249
113.4413	2.0	500	7.7510	0.4879	0.8857	0.5081	0.4204	0.5813	-1.0	0.3757	0.5869	0.624	0.5821	0.682	-1.0	0.4879	0.624
113.4413	3.0	750	7.7158	0.4959	0.8756	0.4886	0.4298	0.5837	-1.0	0.3769	0.5819	0.6162	0.5755	0.6712	-1.0	0.4959	0.6162
13.5624	4.0	1000	8.6789	0.478	0.8615	0.4507	0.4037	0.5851	-1.0	0.3788	0.5807	0.6037	0.5495	0.6777	-1.0	0.478	0.6037
13.5624	5.0	1250	8.1871	0.4645	0.8673	0.3991	0.377	0.5846	-1.0	0.371	0.567	0.5903	0.5348	0.6655	-1.0	0.4645	0.5903
11.903	6.0	1500	7.6425	0.4746	0.8603	0.4375	0.3893	0.5983	-1.0	0.3757	0.5704	0.6006	0.5391	0.6835	-1.0	0.4746	0.6006
11.903	7.0	1750	7.4409	0.4683	0.8799	0.4172	0.3839	0.5855	-1.0	0.3657	0.5595	0.5726	0.5082	0.6604	-1.0	0.4683	0.5726
11.0061	8.0	2000	8.0896	0.453	0.8647	0.4299	0.3769	0.5734	-1.0	0.371	0.5511	0.5757	0.5179	0.654	-1.0	0.453	0.5757
11.0061	9.0	2250	8.6036	0.4815	0.8647	0.4468	0.3982	0.596	-1.0	0.3826	0.5735	0.5925	0.5293	0.6784	-1.0	0.4815	0.5925
10.0433	10.0	2500	7.9387	0.4737	0.862	0.445	0.3918	0.5906	-1.0	0.3832	0.5679	0.5829	0.5147	0.6755	-1.0	0.4737	0.5829
10.0433	11.0	2750	8.7890	0.4569	0.8423	0.4349	0.3676	0.5853	-1.0	0.3788	0.5648	0.5701	0.5022	0.6626	-1.0	0.4569	0.5701
9.3861	12.0	3000	8.4122	0.4512	0.8446	0.4083	0.3684	0.5725	-1.0	0.3695	0.5439	0.5477	0.4652	0.6583	-1.0	0.4512	0.5477
9.3861	13.0	3250	8.4957	0.474	0.8674	0.4603	0.3876	0.5938	-1.0	0.3698	0.5617	0.5664	0.4826	0.6799	-1.0	0.474	0.5664
8.7579	14.0	3500	8.9267	0.4432	0.8394	0.4132	0.3576	0.5764	-1.0	0.3707	0.5545	0.5589	0.4864	0.6568	-1.0	0.4432	0.5589
8.7579	15.0	3750	8.8075	0.4621	0.865	0.43	0.3816	0.5848	-1.0	0.3788	0.566	0.5695	0.4995	0.664	-1.0	0.4621	0.5695
8.2244	16.0	4000	9.2576	0.4502	0.8285	0.4147	0.3581	0.581	-1.0	0.3717	0.5486	0.5498	0.4641	0.6647	-1.0	0.4502	0.5498
8.2244	17.0	4250	9.6513	0.4604	0.8515	0.4364	0.3741	0.5774	-1.0	0.3698	0.5514	0.5514	0.4734	0.6554	-1.0	0.4604	0.5514
7.636	18.0	4500	11.1046	0.4618	0.8406	0.4382	0.3711	0.5884	-1.0	0.3738	0.5664	0.5698	0.4918	0.6741	-1.0	0.4618	0.5698
7.636	19.0	4750	8.5267	0.4641	0.8458	0.4441	0.3849	0.5833	-1.0	0.3738	0.5555	0.5561	0.475	0.6655	-1.0	0.4641	0.5561
7.2484	20.0	5000	9.1166	0.4779	0.8581	0.4521	0.404	0.5864	-1.0	0.3794	0.5776	0.5826	0.5217	0.664	-1.0	0.4779	0.5826
7.2484	21.0	5250	9.3585	0.4664	0.8423	0.4443	0.384	0.583	-1.0	0.3717	0.5601	0.5626	0.4864	0.6662	-1.0	0.4664	0.5626
6.8248	22.0	5500	10.1572	0.4529	0.834	0.4218	0.3623	0.5764	-1.0	0.3645	0.5461	0.5461	0.4663	0.654	-1.0	0.4529	0.5461
6.8248	23.0	5750	8.9893	0.4714	0.8459	0.4437	0.392	0.5816	-1.0	0.367	0.5533	0.5533	0.4783	0.6547	-1.0	0.4714	0.5533
6.5122	24.0	6000	11.0503	0.46	0.8385	0.4326	0.3668	0.5869	-1.0	0.3682	0.5592	0.5611	0.4848	0.664	-1.0	0.46	0.5611
6.5122	25.0	6250	10.4206	0.4701	0.851	0.4487	0.3899	0.5819	-1.0	0.3726	0.5564	0.5576	0.4815	0.6604	-1.0	0.4701	0.5576
6.2986	26.0	6500	9.8390	0.4673	0.853	0.4556	0.3834	0.5843	-1.0	0.3685	0.5555	0.5558	0.4777	0.6619	-1.0	0.4673	0.5558
6.2986	27.0	6750	10.7789	0.4595	0.8459	0.4335	0.3765	0.5702	-1.0	0.3604	0.5498	0.5508	0.4766	0.6518	-1.0	0.4595	0.5508
5.9294	28.0	7000	10.4846	0.4748	0.8523	0.4727	0.4019	0.5762	-1.0	0.3704	0.5589	0.5589	0.4929	0.6475	-1.0	0.4748	0.5589
5.9294	29.0	7250	10.2912	0.4737	0.8553	0.4511	0.3968	0.5805	-1.0	0.3745	0.5607	0.5617	0.4848	0.6655	-1.0	0.4737	0.5617
5.7099	30.0	7500	10.6608	0.4776	0.8573	0.4481	0.398	0.5876	-1.0	0.3704	0.5607	0.5607	0.4859	0.6619	-1.0	0.4776	0.5607
5.7099	31.0	7750	11.0849	0.475	0.8518	0.4529	0.3935	0.5823	-1.0	0.3779	0.5607	0.5607	0.4837	0.6647	-1.0	0.475	0.5607
5.3874	32.0	8000	10.6315	0.476	0.8432	0.4576	0.3934	0.5855	-1.0	0.3723	0.5558	0.5558	0.4772	0.6619	-1.0	0.476	0.5558
5.3874	33.0	8250	11.4011	0.484	0.8574	0.4708	0.4077	0.5854	-1.0	0.3769	0.5642	0.5642	0.4918	0.6612	-1.0	0.484	0.5642
5.1844	34.0	8500	10.8288	0.4794	0.8529	0.4652	0.3996	0.5903	-1.0	0.3729	0.5595	0.5595	0.4766	0.6712	-1.0	0.4794	0.5595
5.1844	35.0	8750	11.1135	0.4836	0.8562	0.4823	0.4041	0.5928	-1.0	0.3766	0.562	0.562	0.4804	0.6719	-1.0	0.4836	0.562
4.9566	36.0	9000	11.1931	0.4703	0.8507	0.4481	0.3936	0.5822	-1.0	0.3704	0.5592	0.5592	0.4832	0.6619	-1.0	0.4703	0.5592
4.9566	37.0	9250	12.2400	0.4747	0.8601	0.447	0.3914	0.5809	-1.0	0.3688	0.5586	0.5586	0.4821	0.6619	-1.0	0.4747	0.5586
4.7564	38.0	9500	11.5422	0.4795	0.8562	0.4608	0.4018	0.5847	-1.0	0.3707	0.5598	0.5598	0.4815	0.6655	-1.0	0.4795	0.5598
4.7564	39.0	9750	11.8137	0.4816	0.8542	0.4693	0.3987	0.5915	-1.0	0.372	0.5601	0.5601	0.4799	0.6683	-1.0	0.4816	0.5601
4.4493	40.0	10000	11.9460	0.4826	0.8549	0.4639	0.4052	0.587	-1.0	0.3751	0.5617	0.5617	0.4826	0.6683	-1.0	0.4826	0.5617
4.4493	41.0	10250	11.3761	0.4777	0.8527	0.4616	0.3966	0.5856	-1.0	0.3717	0.5604	0.5604	0.4821	0.6662	-1.0	0.4777	0.5604
4.2655	42.0	10500	11.8697	0.4775	0.8537	0.4563	0.3953	0.586	-1.0	0.371	0.5579	0.5579	0.4766	0.6676	-1.0	0.4775	0.5579
4.2655	43.0	10750	11.7757	0.4793	0.8521	0.4711	0.3941	0.5881	-1.0	0.3723	0.5611	0.5611	0.481	0.6691	-1.0	0.4793	0.5611
4.0564	44.0	11000	12.3206	0.4821	0.855	0.467	0.4027	0.5877	-1.0	0.3748	0.562	0.562	0.4837	0.6676	-1.0	0.4821	0.562
4.0564	45.0	11250	12.3939	0.4794	0.8535	0.4659	0.3948	0.5919	-1.0	0.3735	0.5604	0.5604	0.4804	0.6683	-1.0	0.4794	0.5604
3.8844	46.0	11500	13.0398	0.4768	0.8519	0.4671	0.3965	0.5865	-1.0	0.3732	0.5595	0.5595	0.4793	0.6676	-1.0	0.4768	0.5595
3.8844	47.0	11750	12.5914	0.4779	0.8459	0.4644	0.3965	0.586	-1.0	0.3738	0.5589	0.5589	0.4804	0.6647	-1.0	0.4779	0.5589
3.6609	48.0	12000	12.6489	0.4819	0.8464	0.4776	0.397	0.5951	-1.0	0.3748	0.5614	0.5614	0.4788	0.6727	-1.0	0.4819	0.5614
3.6609	49.0	12250	12.6018	0.4813	0.8542	0.4652	0.4004	0.5914	-1.0	0.3757	0.5629	0.5629	0.4832	0.6705	-1.0	0.4813	0.5629
3.5309	50.0	12500	12.5405	0.4806	0.8541	0.4648	0.3995	0.5894	-1.0	0.3757	0.5626	0.5626	0.4853	0.6669	-1.0	0.4806	0.5626

Framework versions

Transformers 4.57.1
Pytorch 2.8.0+cu128
Datasets 4.2.0
Tokenizers 0.22.2

dariacuna/rtdetr-v2-r50-finetune-15

rtdetr-v2-r50-finetune-15

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

为此模型做种