rtdetr-v2-r18-finetune-25

This model is a fine-tuned version of PekingU/rtdetr_v2_r18vd on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 4.2341
Map: 0.5136
Map 50: 0.8183
Map 75: 0.6091
Map Small: 0.4905
Map Medium: 0.601
Map Large: -1.0
Mar 1: 0.3239
Mar 10: 0.6524
Mar 100: 0.6987
Mar Small: 0.6703
Mar Medium: 0.7713
Mar Large: -1.0
Map Artemia: 0.5136
Mar 100 Artemia: 0.6987

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	8.8680	0.1829	0.3349	0.184	0.1347	0.3312	-1.0	0.2417	0.5271	0.6065	0.5124	0.7365	-1.0	0.1829	0.6065
80.1538	2.0	500	5.4557	0.4332	0.7886	0.4078	0.3658	0.5522	-1.0	0.3611	0.5953	0.653	0.5925	0.7372	-1.0	0.4332	0.653
80.1538	3.0	750	5.1594	0.4676	0.8146	0.5027	0.3813	0.5978	-1.0	0.372	0.609	0.6667	0.6151	0.7372	-1.0	0.4676	0.6667
8.7802	4.0	1000	5.2335	0.4425	0.8299	0.442	0.3621	0.5799	-1.0	0.3576	0.5875	0.6439	0.5973	0.708	-1.0	0.4425	0.6439
8.7802	5.0	1250	5.2661	0.398	0.757	0.3809	0.3119	0.5667	-1.0	0.343	0.5885	0.6495	0.6032	0.7139	-1.0	0.398	0.6495
7.9468	6.0	1500	5.2284	0.4026	0.7746	0.3847	0.321	0.5753	-1.0	0.3389	0.5782	0.6243	0.5742	0.6934	-1.0	0.4026	0.6243
7.9468	7.0	1750	5.0992	0.4222	0.7979	0.4192	0.3345	0.5702	-1.0	0.3474	0.5963	0.6196	0.5613	0.7007	-1.0	0.4222	0.6196
7.5207	8.0	2000	5.3329	0.4145	0.8017	0.3945	0.3278	0.5691	-1.0	0.3302	0.5769	0.5984	0.5478	0.6686	-1.0	0.4145	0.5984
7.5207	9.0	2250	5.3627	0.3899	0.7398	0.3776	0.3005	0.5599	-1.0	0.3224	0.567	0.5894	0.5371	0.6613	-1.0	0.3899	0.5894
7.024	10.0	2500	5.3062	0.3915	0.7531	0.3797	0.3085	0.5704	-1.0	0.3246	0.5723	0.591	0.5355	0.6672	-1.0	0.3915	0.591
7.024	11.0	2750	5.3825	0.4012	0.8	0.3716	0.3164	0.5622	-1.0	0.3318	0.5629	0.5726	0.521	0.6438	-1.0	0.4012	0.5726
6.7447	12.0	3000	5.4242	0.3836	0.7564	0.3568	0.2827	0.5696	-1.0	0.3156	0.5726	0.5829	0.5081	0.6869	-1.0	0.3836	0.5829
6.7447	13.0	3250	5.5056	0.369	0.7402	0.3523	0.2696	0.5689	-1.0	0.3212	0.5692	0.5829	0.5194	0.6715	-1.0	0.369	0.5829
6.3815	14.0	3500	5.7133	0.351	0.6955	0.3346	0.2525	0.5646	-1.0	0.3143	0.5685	0.5763	0.5134	0.6642	-1.0	0.351	0.5763
6.3815	15.0	3750	5.6219	0.3795	0.7303	0.3665	0.2843	0.5737	-1.0	0.3206	0.5801	0.5872	0.5237	0.6752	-1.0	0.3795	0.5872
6.1381	16.0	4000	5.4845	0.3844	0.7441	0.3704	0.286	0.5717	-1.0	0.3283	0.586	0.5963	0.5274	0.692	-1.0	0.3844	0.5963
6.1381	17.0	4250	5.5927	0.4121	0.7885	0.4219	0.3207	0.5753	-1.0	0.3252	0.5729	0.5769	0.514	0.6635	-1.0	0.4121	0.5769
5.8175	18.0	4500	5.6480	0.3965	0.7462	0.3824	0.2975	0.5673	-1.0	0.3312	0.5685	0.5732	0.5048	0.6686	-1.0	0.3965	0.5732
5.8175	19.0	4750	5.7352	0.393	0.7404	0.3852	0.3018	0.5655	-1.0	0.3184	0.5667	0.5723	0.5059	0.6635	-1.0	0.393	0.5723
5.6458	20.0	5000	5.9227	0.3707	0.7182	0.3351	0.2683	0.5695	-1.0	0.3087	0.5698	0.5776	0.5075	0.6752	-1.0	0.3707	0.5776
5.6458	21.0	5250	5.8445	0.3939	0.7448	0.3828	0.2959	0.5734	-1.0	0.329	0.5732	0.5773	0.5027	0.681	-1.0	0.3939	0.5773
5.456	22.0	5500	5.9328	0.3902	0.7432	0.3746	0.2953	0.5679	-1.0	0.3224	0.5667	0.5685	0.4978	0.6672	-1.0	0.3902	0.5685
5.456	23.0	5750	5.8901	0.4011	0.7566	0.4112	0.3111	0.5552	-1.0	0.3349	0.5676	0.571	0.5005	0.6693	-1.0	0.4011	0.571
5.2706	24.0	6000	5.9919	0.3764	0.725	0.3638	0.2755	0.5614	-1.0	0.3109	0.5567	0.5592	0.4849	0.662	-1.0	0.3764	0.5592
5.2706	25.0	6250	6.0257	0.3856	0.7379	0.3628	0.2839	0.5724	-1.0	0.3259	0.5551	0.5579	0.4823	0.662	-1.0	0.3856	0.5579
5.1699	26.0	6500	6.1196	0.3912	0.7332	0.3921	0.2871	0.5739	-1.0	0.3187	0.5692	0.5717	0.4914	0.6832	-1.0	0.3912	0.5717
5.1699	27.0	6750	6.1308	0.3858	0.727	0.3851	0.2828	0.5712	-1.0	0.3118	0.5592	0.5607	0.479	0.6737	-1.0	0.3858	0.5607
5.0323	28.0	7000	6.2074	0.4008	0.7487	0.3983	0.2992	0.5691	-1.0	0.3308	0.557	0.5579	0.4796	0.6664	-1.0	0.4008	0.5579
5.0323	29.0	7250	6.2535	0.3866	0.7293	0.3962	0.2845	0.5627	-1.0	0.3171	0.5607	0.5607	0.4909	0.6577	-1.0	0.3866	0.5607
4.9072	30.0	7500	6.2425	0.388	0.7302	0.3886	0.2874	0.5559	-1.0	0.3218	0.5558	0.5564	0.4812	0.6613	-1.0	0.388	0.5564
4.9072	31.0	7750	6.3176	0.398	0.7472	0.3807	0.298	0.5669	-1.0	0.3249	0.5539	0.5545	0.4774	0.6613	-1.0	0.398	0.5545
4.7342	32.0	8000	6.3737	0.3944	0.7299	0.398	0.2885	0.5711	-1.0	0.3209	0.557	0.5576	0.4801	0.665	-1.0	0.3944	0.5576
4.7342	33.0	8250	6.5880	0.3869	0.7174	0.3947	0.283	0.5752	-1.0	0.3231	0.5598	0.5604	0.4828	0.6686	-1.0	0.3869	0.5604
4.7047	34.0	8500	6.3981	0.3852	0.7267	0.3741	0.2887	0.5636	-1.0	0.3215	0.5502	0.5508	0.4667	0.6672	-1.0	0.3852	0.5508
4.7047	35.0	8750	6.4883	0.3791	0.7138	0.3601	0.2773	0.5605	-1.0	0.3168	0.5539	0.5551	0.4667	0.6781	-1.0	0.3791	0.5551
4.5866	36.0	9000	6.5402	0.4004	0.7471	0.3942	0.3014	0.5767	-1.0	0.3271	0.5561	0.5573	0.478	0.6672	-1.0	0.4004	0.5573
4.5866	37.0	9250	6.6519	0.3868	0.7348	0.3709	0.2851	0.5684	-1.0	0.3249	0.5536	0.5536	0.4753	0.662	-1.0	0.3868	0.5536
4.5196	38.0	9500	6.5290	0.3962	0.7419	0.3931	0.2986	0.5688	-1.0	0.3252	0.5579	0.5589	0.4801	0.6679	-1.0	0.3962	0.5589
4.5196	39.0	9750	6.6981	0.3958	0.7395	0.3905	0.2949	0.566	-1.0	0.3308	0.5589	0.5595	0.4817	0.6672	-1.0	0.3958	0.5595
4.4211	40.0	10000	6.7341	0.3902	0.7301	0.3701	0.2851	0.5752	-1.0	0.3218	0.5623	0.5626	0.4817	0.6745	-1.0	0.3902	0.5626
4.4211	41.0	10250	6.8023	0.3936	0.7304	0.3793	0.2895	0.5736	-1.0	0.3224	0.5558	0.5564	0.4726	0.6723	-1.0	0.3936	0.5564
4.3384	42.0	10500	6.7279	0.3973	0.7365	0.3819	0.2992	0.5633	-1.0	0.3324	0.5567	0.5573	0.4763	0.6693	-1.0	0.3973	0.5573
4.3384	43.0	10750	6.6842	0.4041	0.7502	0.3839	0.3022	0.571	-1.0	0.3302	0.5548	0.5551	0.4742	0.6672	-1.0	0.4041	0.5551
4.2885	44.0	11000	6.8704	0.3992	0.7431	0.3811	0.2946	0.5766	-1.0	0.3299	0.5561	0.5561	0.471	0.6737	-1.0	0.3992	0.5561
4.2885	45.0	11250	6.8004	0.3993	0.7428	0.3891	0.2998	0.5677	-1.0	0.3287	0.5533	0.5533	0.472	0.6657	-1.0	0.3993	0.5533
4.2213	46.0	11500	6.9365	0.3965	0.7395	0.3898	0.2926	0.5718	-1.0	0.3299	0.5517	0.5517	0.4731	0.6606	-1.0	0.3965	0.5517
4.2213	47.0	11750	6.9643	0.3962	0.7392	0.3777	0.2897	0.5768	-1.0	0.3302	0.5558	0.5564	0.4753	0.6686	-1.0	0.3962	0.5564
4.1368	48.0	12000	6.9463	0.3964	0.7389	0.3804	0.2901	0.5746	-1.0	0.3293	0.5539	0.5539	0.472	0.6672	-1.0	0.3964	0.5539
4.1368	49.0	12250	6.9413	0.3971	0.7347	0.3839	0.2971	0.5725	-1.0	0.3299	0.5551	0.5555	0.4769	0.6642	-1.0	0.3971	0.5555
4.1304	50.0	12500	6.9121	0.3985	0.7426	0.3811	0.296	0.5699	-1.0	0.328	0.5542	0.5545	0.4731	0.6672	-1.0	0.3985	0.5545

Framework versions

Transformers 4.57.1
Pytorch 2.8.0+cu128
Datasets 4.2.0
Tokenizers 0.22.2

dariacuna/rtdetr-v2-r18-finetune-25

rtdetr-v2-r18-finetune-25

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions