deta-swin-large

This model is a fine-tuned version of jozhang97/deta-swin-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 7.7323
• Map: 0.448
• Map 50: 0.6146
• Map 75: 0.4994
• Map Small: 0.4453
• Map Medium: 0.5526
• Map Large: -1.0
• Mar 1: 0.286
• Mar 10: 0.4045
• Mar 100: 0.5073
• Mar Small: 0.5039
• Mar Medium: 0.571
• Mar Large: -1.0
• Map Ball: 0.2858
• Mar 100 Ball: 0.4068
• Map Goalkeeper: 0.7008
• Mar 100 Goalkeeper: 0.7649
• Map Player: 0.8053
• Mar 100 Player: 0.8576
• Map Referee: 0.0
• Mar 100 Referee: 0.0

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: 0.0001
• train_batch_size: 2
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20

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 Ball	Mar 100 Ball	Map Goalkeeper	Mar 100 Goalkeeper	Map Player	Mar 100 Player	Map Referee	Mar 100 Referee
296.278	1.0	149	62.3890	0.0999	0.1535	0.1216	0.0976	0.1516	-1.0	0.0078	0.0616	0.1808	0.1808	0.2414	-1.0	0.0	0.0	0.0	0.0	0.3997	0.7232	0.0	0.0
74.5246	2.0	298	25.7949	0.1643	0.236	0.1843	0.1638	0.2281	-1.0	0.0162	0.1175	0.3643	0.3753	0.2564	-1.0	0.0	0.0	0.0189	0.6969	0.6384	0.7605	0.0	0.0
23.5211	3.0	447	16.6520	0.1984	0.25	0.2246	0.1971	0.2759	-1.0	0.0111	0.2288	0.4151	0.4131	0.5609	-1.0	0.0	0.0	0.0371	0.8114	0.7566	0.849	0.0	0.0
18.7573	4.0	596	13.6105	0.2271	0.3101	0.2518	0.2264	0.2742	-1.0	0.0608	0.2627	0.4917	0.4901	0.5426	-1.0	0.1064	0.3159	0.0287	0.8139	0.7732	0.8368	0.0	0.0
14.4593	5.0	745	11.9880	0.2282	0.309	0.2486	0.2272	0.2805	-1.0	0.0774	0.2571	0.4835	0.4812	0.531	-1.0	0.1022	0.3023	0.028	0.7914	0.7824	0.8405	0.0	0.0
12.878	6.0	894	10.9109	0.3867	0.5529	0.418	0.3955	0.3944	-1.0	0.2517	0.3723	0.4836	0.4837	0.526	-1.0	0.1759	0.3349	0.6007	0.7629	0.7704	0.8367	0.0	0.0
11.4964	7.0	1043	10.0801	0.4098	0.5788	0.4432	0.4076	0.5423	-1.0	0.2558	0.3796	0.4805	0.4774	0.5626	-1.0	0.2016	0.3227	0.6587	0.7622	0.7792	0.8372	0.0	0.0
10.938	8.0	1192	9.6962	0.4168	0.5909	0.4511	0.4136	0.5736	-1.0	0.2622	0.384	0.4829	0.479	0.5953	-1.0	0.2148	0.3045	0.6771	0.7944	0.7753	0.8327	0.0	0.0
10.2185	9.0	1341	9.2201	0.4414	0.6237	0.4939	0.4461	0.4126	-1.0	0.2975	0.4088	0.5078	0.5043	0.5659	-1.0	0.3159	0.4023	0.6628	0.7886	0.7866	0.8403	0.0	0.0
10.1417	10.0	1490	8.8414	0.4258	0.5998	0.4616	0.432	0.4459	-1.0	0.2829	0.4079	0.5097	0.5112	0.5104	-1.0	0.2659	0.4256	0.6382	0.76	0.799	0.8533	0.0	0.0
9.6417	11.0	1639	8.9382	0.4158	0.5798	0.4758	0.4134	0.5169	-1.0	0.2758	0.401	0.5016	0.499	0.5324	-1.0	0.2131	0.3791	0.6644	0.7857	0.7857	0.8414	0.0	0.0
9.3618	12.0	1788	8.5653	0.4349	0.6074	0.5026	0.4337	0.5436	-1.0	0.2857	0.4042	0.5052	0.5022	0.5646	-1.0	0.2359	0.375	0.7111	0.8	0.7927	0.8458	0.0	0.0
9.1601	13.0	1937	8.5162	0.4346	0.6159	0.4669	0.4332	0.5417	-1.0	0.2803	0.3997	0.4998	0.497	0.5623	-1.0	0.2628	0.3795	0.6862	0.7778	0.7892	0.842	0.0	0.0
8.9938	14.0	2086	8.2598	0.4403	0.6165	0.4962	0.4379	0.5494	-1.0	0.2803	0.4002	0.5017	0.4986	0.5674	-1.0	0.2661	0.3773	0.7019	0.7806	0.7931	0.8489	0.0	0.0
8.7258	15.0	2235	8.2129	0.4507	0.6142	0.5192	0.4505	0.5363	-1.0	0.2931	0.4137	0.5149	0.5123	0.5611	-1.0	0.3076	0.4114	0.7154	0.8028	0.7796	0.8453	0.0	0.0
8.1655	16.0	2384	8.1459	0.4436	0.6071	0.4985	0.4455	0.4346	-1.0	0.2891	0.4125	0.5136	0.5134	0.5115	-1.0	0.3046	0.4091	0.6794	0.7946	0.7905	0.8507	0.0	0.0
8.1808	17.0	2533	8.1069	0.4426	0.6231	0.4971	0.4436	0.5167	-1.0	0.2868	0.4078	0.5074	0.5057	0.5667	-1.0	0.2985	0.4023	0.6942	0.7886	0.7776	0.8388	0.0	0.0
8.45	18.0	2682	7.9322	0.4443	0.6157	0.4739	0.4427	0.5757	-1.0	0.2839	0.4082	0.5104	0.5066	0.5963	-1.0	0.2715	0.4023	0.704	0.7833	0.8018	0.856	0.0	0.0
8.0344	19.0	2831	7.9556	0.4464	0.6133	0.4963	0.4456	0.5388	-1.0	0.2866	0.4113	0.5129	0.5111	0.5584	-1.0	0.2678	0.4093	0.722	0.7914	0.7956	0.851	0.0	0.0
7.9178	20.0	2980	7.7323	0.448	0.6146	0.4994	0.4453	0.5526	-1.0	0.286	0.4045	0.5073	0.5039	0.571	-1.0	0.2858	0.4068	0.7008	0.7649	0.8053	0.8576	0.0	0.0

Framework versions

• Transformers 4.49.0
• Pytorch 2.5.1+cu124
• Datasets 3.3.1
• Tokenizers 0.21.0