detr_finetuned_kitti_mots

This model is a fine-tuned version of microsoft/conditional-detr-resnet-50 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5468
• Map: 0.5754
• Map 50: 0.8688
• Map 75: 0.625
• Map Small: 0.3516
• Map Medium: 0.5857
• Map Large: 0.783
• Mar 1: 0.1695
• Mar 10: 0.6195
• Mar 100: 0.6862
• Mar Small: 0.5362
• Mar Medium: 0.6977
• Mar Large: 0.8372
• Map Car: 0.6938
• Mar 100 Car: 0.769
• Map Pedestrian: 0.4569
• Mar 100 Pedestrian: 0.6034

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: 8
• 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: cosine
• 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 Car	Mar 100 Car	Map Pedestrian	Mar 100 Pedestrian
1.5904	1.0	743	1.2276	0.1644	0.3572	0.1419	0.0253	0.1267	0.3984	0.08	0.2466	0.3977	0.1364	0.4022	0.7009	0.277	0.4472	0.0518	0.3482
1.3176	2.0	1486	1.1301	0.2261	0.5107	0.1714	0.0479	0.1929	0.5	0.1024	0.3066	0.4201	0.201	0.4196	0.6825	0.3095	0.4709	0.1426	0.3694
1.1187	3.0	2229	1.0565	0.2699	0.5705	0.2147	0.0416	0.2365	0.5833	0.1142	0.332	0.4328	0.1875	0.4434	0.7003	0.3674	0.4924	0.1725	0.3733
1.0835	4.0	2972	1.0326	0.2851	0.5903	0.2363	0.0572	0.2557	0.6048	0.116	0.3481	0.4474	0.2339	0.4447	0.7149	0.3781	0.5115	0.1921	0.3832
1.0681	5.0	3715	1.0095	0.3017	0.5998	0.2728	0.0538	0.2878	0.6051	0.1205	0.3625	0.4664	0.2237	0.4787	0.7298	0.4003	0.5288	0.2031	0.404
1.042	6.0	4458	1.0315	0.271	0.6062	0.1981	0.0628	0.2597	0.5656	0.1073	0.3465	0.4393	0.2258	0.4469	0.6762	0.3601	0.4929	0.182	0.3857
1.0051	7.0	5201	0.9796	0.2875	0.6189	0.2331	0.0695	0.2789	0.5719	0.1122	0.367	0.473	0.2403	0.4912	0.7076	0.3936	0.5423	0.1814	0.4036
0.9798	8.0	5944	0.9567	0.3178	0.6257	0.2915	0.0823	0.3041	0.6211	0.124	0.379	0.485	0.2552	0.5059	0.7105	0.4214	0.5633	0.2142	0.4068
0.9814	9.0	6687	1.0122	0.2799	0.579	0.239	0.0493	0.2585	0.5948	0.1124	0.3488	0.4607	0.2195	0.4725	0.7226	0.3877	0.5279	0.1721	0.3934
1.0057	10.0	7430	0.9464	0.3156	0.6365	0.2797	0.081	0.3083	0.6042	0.1201	0.3814	0.4929	0.275	0.5011	0.736	0.4326	0.5579	0.1986	0.428
0.9267	11.0	8173	0.9137	0.3364	0.6634	0.2955	0.118	0.326	0.6291	0.1276	0.3995	0.5073	0.3059	0.5147	0.7301	0.4369	0.5728	0.2359	0.4417
0.8938	12.0	8916	0.8808	0.3622	0.6811	0.3348	0.1147	0.3497	0.6639	0.1358	0.4187	0.5212	0.2956	0.536	0.7572	0.4784	0.5986	0.246	0.4438
0.8575	13.0	9659	0.8632	0.3614	0.6888	0.34	0.1005	0.3647	0.6565	0.1305	0.4254	0.5258	0.2989	0.5449	0.7523	0.4629	0.5893	0.2599	0.4624
0.8442	14.0	10402	0.8544	0.3746	0.6922	0.3628	0.1129	0.3759	0.6419	0.1356	0.4284	0.5352	0.3325	0.5545	0.7302	0.4878	0.6007	0.2614	0.4696
0.8213	15.0	11145	0.8441	0.3792	0.7034	0.3561	0.1321	0.3742	0.6534	0.1365	0.4317	0.5378	0.324	0.56	0.7363	0.5039	0.6128	0.2544	0.4628
0.8115	16.0	11888	0.8210	0.3888	0.709	0.3807	0.128	0.392	0.6675	0.1399	0.4413	0.5468	0.3479	0.5616	0.7491	0.5037	0.6212	0.2739	0.4723
0.8093	17.0	12631	0.8127	0.385	0.7092	0.3751	0.1383	0.3858	0.6626	0.1381	0.4399	0.5493	0.3673	0.556	0.7529	0.5141	0.6257	0.2559	0.4729
0.7893	18.0	13374	0.8204	0.3901	0.7194	0.3675	0.1295	0.3842	0.6851	0.1378	0.4404	0.5416	0.3486	0.5481	0.7602	0.5048	0.6092	0.2755	0.474
0.7459	19.0	14117	0.7880	0.4058	0.729	0.4085	0.1582	0.4043	0.6709	0.1407	0.46	0.5627	0.3686	0.5794	0.7528	0.53	0.6421	0.2816	0.4832
0.7383	20.0	14860	0.7477	0.4264	0.7551	0.4264	0.1563	0.4328	0.7092	0.1441	0.48	0.5888	0.3989	0.6004	0.788	0.538	0.6577	0.3149	0.5199
0.7207	21.0	15603	0.7688	0.4188	0.7599	0.4087	0.1591	0.419	0.7057	0.1419	0.4691	0.5641	0.3832	0.5686	0.772	0.5306	0.6361	0.307	0.492
0.7127	22.0	16346	0.7450	0.4368	0.7622	0.4379	0.1733	0.4368	0.7099	0.1483	0.4828	0.5834	0.3873	0.5959	0.7875	0.5601	0.6532	0.3136	0.5136
0.698	23.0	17089	0.7429	0.4401	0.7626	0.4515	0.1739	0.4456	0.7194	0.1498	0.4841	0.5901	0.3878	0.6079	0.7878	0.5536	0.6524	0.3266	0.5278
0.6836	24.0	17832	0.7642	0.422	0.7565	0.4217	0.154	0.428	0.7066	0.1433	0.47	0.5743	0.3941	0.5791	0.7801	0.5423	0.6404	0.3017	0.5081
0.6684	25.0	18575	0.7016	0.4599	0.7889	0.4717	0.1873	0.4706	0.7334	0.1497	0.5093	0.6116	0.4221	0.6271	0.8006	0.5767	0.6837	0.3431	0.5395
0.6471	26.0	19318	0.6890	0.4724	0.791	0.4869	0.2053	0.4825	0.7304	0.1552	0.516	0.6068	0.4245	0.6179	0.7994	0.6004	0.6895	0.3443	0.5241
0.6259	27.0	20061	0.6788	0.4726	0.7816	0.4914	0.2182	0.481	0.7202	0.1555	0.5212	0.6199	0.4374	0.6387	0.7919	0.6137	0.7058	0.3316	0.5339
0.6038	28.0	20804	0.6664	0.4844	0.8057	0.5206	0.221	0.4985	0.7352	0.1539	0.5336	0.6236	0.4423	0.6382	0.8047	0.6049	0.6987	0.3639	0.5486
0.5963	29.0	21547	0.6562	0.4952	0.8143	0.5265	0.2398	0.5087	0.7436	0.1549	0.54	0.6299	0.4495	0.6445	0.8094	0.6177	0.7083	0.3728	0.5514
0.5821	30.0	22290	0.6533	0.5019	0.8238	0.5289	0.2385	0.5151	0.7553	0.1579	0.5441	0.6328	0.4478	0.6471	0.8184	0.6045	0.6965	0.3992	0.5691
0.5642	31.0	23033	0.6434	0.506	0.8291	0.532	0.2451	0.5199	0.7421	0.1575	0.5505	0.6342	0.4598	0.6482	0.8086	0.6226	0.7103	0.3893	0.558
0.5547	32.0	23776	0.6382	0.5041	0.8261	0.5454	0.2484	0.5193	0.7529	0.1566	0.551	0.6357	0.4587	0.6488	0.8154	0.6221	0.7123	0.3861	0.5591
0.536	33.0	24519	0.6175	0.5188	0.8382	0.5556	0.2732	0.5318	0.753	0.1607	0.565	0.644	0.4731	0.659	0.8112	0.6352	0.7259	0.4024	0.5621
0.5231	34.0	25262	0.6037	0.531	0.8421	0.5703	0.2879	0.5412	0.765	0.1636	0.5774	0.6578	0.4848	0.6731	0.827	0.6485	0.7353	0.4134	0.5803
0.5042	35.0	26005	0.5947	0.5373	0.846	0.5869	0.3039	0.5495	0.7621	0.1622	0.583	0.6614	0.507	0.6715	0.8218	0.6552	0.7398	0.4194	0.5829
0.4956	36.0	26748	0.5955	0.5379	0.8496	0.5835	0.2997	0.5484	0.7641	0.1651	0.5818	0.6599	0.498	0.6726	0.8222	0.6606	0.7391	0.4152	0.5808
0.4838	37.0	27491	0.5849	0.549	0.8514	0.591	0.3076	0.5609	0.7749	0.1669	0.5937	0.6703	0.5038	0.6858	0.8311	0.6618	0.7439	0.4362	0.5967
0.4586	38.0	28234	0.5708	0.5568	0.8591	0.6044	0.3296	0.5658	0.7782	0.1678	0.6007	0.6776	0.5205	0.6896	0.8356	0.6762	0.7563	0.4374	0.5989
0.455	39.0	28977	0.5749	0.5525	0.8581	0.6	0.3221	0.5597	0.7801	0.1669	0.5965	0.6695	0.5138	0.6786	0.834	0.6722	0.7511	0.4327	0.5879
0.4426	40.0	29720	0.5670	0.5605	0.8626	0.6086	0.3276	0.573	0.7795	0.1678	0.6052	0.6765	0.5188	0.6894	0.8331	0.6771	0.7574	0.4439	0.5957
0.4337	41.0	30463	0.5652	0.5621	0.8631	0.6117	0.3368	0.5715	0.7822	0.1677	0.6074	0.6774	0.5267	0.686	0.8367	0.6783	0.7547	0.4459	0.6001
0.4181	42.0	31206	0.5612	0.5625	0.8635	0.6096	0.3284	0.5754	0.782	0.1677	0.6074	0.678	0.5179	0.6906	0.8381	0.6785	0.7569	0.4466	0.5992
0.4198	43.0	31949	0.5575	0.5692	0.8651	0.6197	0.3385	0.5803	0.7836	0.1698	0.6139	0.6811	0.5308	0.6905	0.8384	0.6867	0.7632	0.4516	0.599
0.4058	44.0	32692	0.5524	0.5706	0.8675	0.6212	0.3436	0.5816	0.7814	0.1692	0.6154	0.6847	0.5345	0.6962	0.8357	0.6871	0.7652	0.4541	0.6042
0.4026	45.0	33435	0.5508	0.5722	0.8664	0.6223	0.3448	0.5838	0.7817	0.1686	0.618	0.6842	0.5337	0.6959	0.8356	0.6902	0.767	0.4541	0.6015
0.398	46.0	34178	0.5498	0.5733	0.8701	0.6257	0.348	0.5843	0.7826	0.169	0.6185	0.6856	0.5347	0.6974	0.8371	0.6898	0.7665	0.4569	0.6048
0.3954	47.0	34921	0.5471	0.5749	0.8687	0.6261	0.3515	0.5856	0.784	0.1695	0.6195	0.6866	0.5355	0.6977	0.8395	0.6925	0.7683	0.4573	0.6049
0.3967	48.0	35664	0.5491	0.5736	0.87	0.6228	0.3481	0.5848	0.7816	0.169	0.6181	0.6848	0.5329	0.697	0.8362	0.6918	0.7674	0.4554	0.6022
0.3874	49.0	36407	0.5468	0.5756	0.8689	0.6253	0.3525	0.5857	0.783	0.1697	0.62	0.6863	0.5366	0.6977	0.8372	0.694	0.7691	0.4571	0.6035
0.3861	50.0	37150	0.5468	0.5754	0.8688	0.625	0.3516	0.5857	0.783	0.1695	0.6195	0.6862	0.5362	0.6977	0.8372	0.6938	0.769	0.4569	0.6034

Framework versions

• Transformers 4.49.0
• Pytorch 2.5.1+cu121
• Datasets 3.3.2
• Tokenizers 0.21.0