rtdetr-v2-r101-finetune-23

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: 6.8074
Map: 0.5389
Map 50: 0.8718
Map 75: 0.6194
Map Small: 0.5007
Map Medium: 0.6359
Map Large: -1.0
Mar 1: 0.3337
Mar 10: 0.6301
Mar 100: 0.6728
Mar Small: 0.6514
Mar Medium: 0.7276
Mar Large: -1.0
Map Artemia: 0.5389
Mar 100 Artemia: 0.6728

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	10.0998	0.2889	0.5533	0.2557	0.2187	0.5121	-1.0	0.3087	0.5237	0.5717	0.5237	0.6394	-1.0	0.2889	0.5717
95.2096	2.0	500	8.2889	0.476	0.8502	0.4796	0.4064	0.5795	-1.0	0.3682	0.5682	0.6237	0.5849	0.6788	-1.0	0.476	0.6237
95.2096	3.0	750	8.0750	0.4878	0.8721	0.5234	0.4123	0.5949	-1.0	0.3701	0.6012	0.6498	0.5909	0.7314	-1.0	0.4878	0.6498
13.5653	4.0	1000	8.2427	0.4908	0.87	0.4907	0.4234	0.5919	-1.0	0.3791	0.5748	0.6255	0.5763	0.6934	-1.0	0.4908	0.6255
13.5653	5.0	1250	8.3591	0.4732	0.8601	0.4663	0.3991	0.5818	-1.0	0.3729	0.5604	0.605	0.5489	0.6825	-1.0	0.4732	0.605
11.7937	6.0	1500	8.6404	0.4755	0.8701	0.4533	0.4094	0.574	-1.0	0.371	0.571	0.6103	0.5737	0.6606	-1.0	0.4755	0.6103
11.7937	7.0	1750	8.1820	0.4611	0.8655	0.437	0.3909	0.5685	-1.0	0.3651	0.5626	0.6065	0.5435	0.6934	-1.0	0.4611	0.6065
10.8248	8.0	2000	8.4100	0.4778	0.8697	0.4814	0.405	0.587	-1.0	0.3695	0.5847	0.5994	0.5392	0.6818	-1.0	0.4778	0.5994
10.8248	9.0	2250	8.3609	0.4814	0.8752	0.4643	0.4113	0.5814	-1.0	0.3769	0.5794	0.5988	0.5468	0.6693	-1.0	0.4814	0.5988
9.8336	10.0	2500	8.6048	0.464	0.8586	0.4632	0.3857	0.5763	-1.0	0.3657	0.5558	0.5607	0.4909	0.6584	-1.0	0.464	0.5607
9.8336	11.0	2750	8.7471	0.464	0.8511	0.4839	0.3911	0.5752	-1.0	0.3682	0.562	0.5732	0.514	0.6555	-1.0	0.464	0.5732
9.2597	12.0	3000	8.5435	0.4666	0.8668	0.4645	0.3917	0.5776	-1.0	0.3717	0.5548	0.5555	0.4742	0.6686	-1.0	0.4666	0.5555
9.2597	13.0	3250	9.0042	0.46	0.858	0.4453	0.3833	0.5731	-1.0	0.3592	0.5508	0.5564	0.4817	0.6599	-1.0	0.46	0.5564
8.7205	14.0	3500	8.7921	0.4561	0.8412	0.4528	0.3868	0.5665	-1.0	0.3632	0.5514	0.5517	0.4849	0.6438	-1.0	0.4561	0.5517
8.7205	15.0	3750	8.8801	0.4584	0.8435	0.4678	0.3873	0.565	-1.0	0.3648	0.5486	0.5492	0.4769	0.6496	-1.0	0.4584	0.5492
8.2055	16.0	4000	8.7329	0.4558	0.8541	0.4141	0.3815	0.5711	-1.0	0.3614	0.5414	0.5414	0.4591	0.6547	-1.0	0.4558	0.5414
8.2055	17.0	4250	9.2023	0.4658	0.8518	0.4487	0.3859	0.5815	-1.0	0.3667	0.5551	0.5564	0.4774	0.665	-1.0	0.4658	0.5564
7.7349	18.0	4500	9.4601	0.4602	0.8394	0.4636	0.3757	0.5817	-1.0	0.3579	0.5424	0.5424	0.4559	0.6613	-1.0	0.4602	0.5424
7.7349	19.0	4750	8.7288	0.4699	0.8449	0.4631	0.3922	0.5888	-1.0	0.3726	0.5536	0.5536	0.4726	0.665	-1.0	0.4699	0.5536
7.42	20.0	5000	9.1784	0.4469	0.8194	0.4327	0.3695	0.5652	-1.0	0.3682	0.5333	0.5333	0.4511	0.6467	-1.0	0.4469	0.5333
7.42	21.0	5250	9.5614	0.4633	0.8463	0.4678	0.3881	0.5717	-1.0	0.3695	0.5495	0.5495	0.4753	0.6526	-1.0	0.4633	0.5495
7.0374	22.0	5500	10.2999	0.4535	0.8365	0.4156	0.3664	0.5762	-1.0	0.3604	0.5352	0.5352	0.4425	0.6628	-1.0	0.4535	0.5352
7.0374	23.0	5750	9.9020	0.4709	0.8487	0.4987	0.3886	0.586	-1.0	0.367	0.5517	0.5517	0.472	0.6613	-1.0	0.4709	0.5517
6.6871	24.0	6000	9.2232	0.4657	0.8355	0.4676	0.3918	0.5778	-1.0	0.3645	0.5455	0.5455	0.464	0.6577	-1.0	0.4657	0.5455
6.6871	25.0	6250	10.1620	0.4682	0.8501	0.453	0.3872	0.5743	-1.0	0.3623	0.5455	0.5455	0.464	0.6577	-1.0	0.4682	0.5455
6.5237	26.0	6500	9.1917	0.4759	0.8466	0.4649	0.3991	0.5828	-1.0	0.3695	0.5502	0.5502	0.4699	0.6606	-1.0	0.4759	0.5502
6.5237	27.0	6750	9.7212	0.4613	0.8357	0.4816	0.3841	0.5818	-1.0	0.3617	0.5439	0.5439	0.4661	0.6511	-1.0	0.4613	0.5439
6.2744	28.0	7000	9.4923	0.4786	0.8478	0.4812	0.4014	0.5893	-1.0	0.3729	0.5526	0.5526	0.4731	0.6628	-1.0	0.4786	0.5526
6.2744	29.0	7250	10.0511	0.4672	0.8314	0.4528	0.3847	0.5859	-1.0	0.3645	0.5542	0.5548	0.4737	0.6672	-1.0	0.4672	0.5548
6.1229	30.0	7500	9.9844	0.4731	0.8475	0.4857	0.3945	0.5841	-1.0	0.3657	0.5455	0.5455	0.4651	0.6562	-1.0	0.4731	0.5455
6.1229	31.0	7750	10.4688	0.4722	0.8483	0.4524	0.3892	0.5828	-1.0	0.3651	0.5489	0.5489	0.4667	0.662	-1.0	0.4722	0.5489
5.8374	32.0	8000	9.9301	0.4702	0.8375	0.4911	0.395	0.5815	-1.0	0.367	0.5421	0.5421	0.4624	0.6518	-1.0	0.4702	0.5421
5.8374	33.0	8250	10.9157	0.4695	0.8342	0.4705	0.3898	0.5833	-1.0	0.3623	0.5514	0.5514	0.4715	0.662	-1.0	0.4695	0.5514
5.6009	34.0	8500	10.2799	0.4667	0.8273	0.4598	0.3817	0.5875	-1.0	0.3614	0.5427	0.5427	0.4613	0.6547	-1.0	0.4667	0.5427
5.6009	35.0	8750	10.6563	0.4694	0.8353	0.4837	0.3865	0.5891	-1.0	0.3664	0.5517	0.5517	0.471	0.6628	-1.0	0.4694	0.5517
5.3954	36.0	9000	10.8796	0.4749	0.8415	0.4884	0.3956	0.5876	-1.0	0.3685	0.5511	0.5511	0.4694	0.6635	-1.0	0.4749	0.5511
5.3954	37.0	9250	10.9810	0.4679	0.8314	0.4783	0.3839	0.5935	-1.0	0.3673	0.552	0.552	0.4694	0.6657	-1.0	0.4679	0.552
5.2373	38.0	9500	10.8722	0.4681	0.8259	0.4596	0.3833	0.5969	-1.0	0.3607	0.5498	0.5498	0.4661	0.665	-1.0	0.4681	0.5498
5.2373	39.0	9750	10.8305	0.4712	0.833	0.4798	0.3886	0.5898	-1.0	0.3682	0.5467	0.5467	0.4608	0.665	-1.0	0.4712	0.5467
5.0038	40.0	10000	10.7798	0.465	0.8293	0.4649	0.3821	0.5901	-1.0	0.3592	0.5439	0.5439	0.457	0.6635	-1.0	0.465	0.5439
5.0038	41.0	10250	10.8215	0.4727	0.8337	0.4808	0.3914	0.5901	-1.0	0.3623	0.547	0.547	0.4602	0.6664	-1.0	0.4727	0.547
4.841	42.0	10500	11.3329	0.4684	0.8275	0.4807	0.3816	0.5989	-1.0	0.3567	0.547	0.547	0.4581	0.6693	-1.0	0.4684	0.547
4.841	43.0	10750	11.2774	0.4673	0.8266	0.4856	0.3858	0.5884	-1.0	0.3611	0.548	0.548	0.4667	0.6599	-1.0	0.4673	0.548
4.6183	44.0	11000	11.4358	0.4712	0.8284	0.5025	0.3865	0.5939	-1.0	0.3611	0.5474	0.5474	0.4624	0.6642	-1.0	0.4712	0.5474
4.6183	45.0	11250	11.5586	0.4656	0.8243	0.4834	0.3829	0.5877	-1.0	0.3576	0.5452	0.5452	0.4613	0.6606	-1.0	0.4656	0.5452
4.4687	46.0	11500	11.5333	0.4682	0.8182	0.4863	0.3839	0.5946	-1.0	0.3583	0.5455	0.5455	0.4597	0.6635	-1.0	0.4682	0.5455
4.4687	47.0	11750	11.9874	0.4722	0.8297	0.4905	0.389	0.5976	-1.0	0.3642	0.5483	0.5483	0.4634	0.665	-1.0	0.4722	0.5483
4.2431	48.0	12000	12.0282	0.4742	0.8307	0.4959	0.3922	0.5922	-1.0	0.3676	0.548	0.548	0.4634	0.6642	-1.0	0.4742	0.548
4.2431	49.0	12250	12.0521	0.4692	0.8262	0.4812	0.3824	0.5971	-1.0	0.3636	0.5461	0.5461	0.4597	0.665	-1.0	0.4692	0.5461
4.1432	50.0	12500	12.0597	0.4688	0.8185	0.4884	0.3825	0.5966	-1.0	0.3604	0.5445	0.5445	0.457	0.665	-1.0	0.4688	0.5445

Framework versions

Transformers 4.57.1
Pytorch 2.8.0+cu128
Datasets 4.2.0
Tokenizers 0.22.2

dariacuna/rtdetr-v2-r101-finetune-23

rtdetr-v2-r101-finetune-23

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

为此模型做种