lw-detr-medium-tray-detection-hub-init

This model is a fine-tuned version of AnnaZhang/lwdetr_medium_60e_coco on the nielsr/tray-cart-detection dataset. It achieves the following results on the evaluation set:

• Loss: 9.3240
• Map: 0.4578
• Map 50: 0.7573
• Map 75: 0.4780
• Map Small: 0.6219
• Map Medium: 0.4323
• Map Large: 0.6013
• Mar 1: 0.0666
• Mar 10: 0.3331
• Mar 100: 0.5372
• Mar Small: 0.6238
• Mar Medium: 0.5074
• Mar Large: 0.7371
• Map Per Class: -1.0
• Mar 100 Per Class: -1.0

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: 2
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• 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
• num_epochs: 300.0

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 Per Class	Mar 100 Per Class
8.1848	1.0	25	7.9812	0.1435	0.3363	0.1109	0.4139	0.1333	0.2534	0.0251	0.1157	0.2631	0.5857	0.2184	0.4740	-1.0	-1.0
5.8618	2.0	50	8.2678	0.2526	0.4962	0.2112	0.4874	0.2491	0.3527	0.0423	0.1901	0.3557	0.5119	0.3278	0.5104	-1.0	-1.0
5.1454	3.0	75	8.5504	0.2988	0.5869	0.2618	0.5210	0.2907	0.3796	0.0424	0.2381	0.4039	0.6048	0.3726	0.5627	-1.0	-1.0
4.7482	4.0	100	8.6991	0.3458	0.6480	0.3359	0.5575	0.3369	0.4219	0.0475	0.2574	0.4483	0.6024	0.4198	0.6106	-1.0	-1.0
4.5330	5.0	125	8.3971	0.3721	0.6832	0.3540	0.5741	0.3682	0.4560	0.0540	0.2843	0.4978	0.6333	0.4800	0.5878	-1.0	-1.0
4.1680	6.0	150	8.6646	0.3952	0.7328	0.3653	0.5804	0.3761	0.5275	0.0585	0.2848	0.4699	0.6119	0.4481	0.5852	-1.0	-1.0
4.2098	7.0	175	9.0848	0.3944	0.7143	0.3776	0.5974	0.3830	0.4835	0.0524	0.2734	0.4757	0.6119	0.4600	0.5490	-1.0	-1.0
3.9427	8.0	200	8.4698	0.3977	0.7213	0.3679	0.5200	0.3686	0.5834	0.0572	0.2828	0.4775	0.5548	0.4461	0.6902	-1.0	-1.0
3.8337	9.0	225	9.0207	0.3865	0.6856	0.4031	0.6180	0.3788	0.4956	0.0589	0.2715	0.4838	0.6333	0.4621	0.6002	-1.0	-1.0
3.7943	10.0	250	8.9388	0.4012	0.7208	0.3869	0.5656	0.3900	0.4790	0.0517	0.3104	0.4934	0.5833	0.4758	0.6009	-1.0	-1.0
3.7726	11.0	275	9.2610	0.3941	0.7179	0.3722	0.5896	0.3823	0.4907	0.0481	0.2932	0.4751	0.5905	0.4543	0.5923	-1.0	-1.0
3.5588	12.0	300	9.1557	0.3897	0.7229	0.3494	0.6089	0.3730	0.5256	0.0480	0.2957	0.4950	0.6238	0.4745	0.6066	-1.0	-1.0
3.4384	13.0	325	8.9558	0.4018	0.7064	0.3752	0.5532	0.3983	0.4603	0.0504	0.3014	0.4933	0.5571	0.4783	0.5924	-1.0	-1.0
3.4022	14.0	350	9.3375	0.4079	0.7364	0.4034	0.5751	0.3916	0.5041	0.0518	0.2976	0.4990	0.5857	0.4700	0.6902	-1.0	-1.0
3.3132	15.0	375	8.7625	0.4124	0.7299	0.3900	0.6249	0.3923	0.5530	0.0563	0.3020	0.4983	0.6310	0.4747	0.6368	-1.0	-1.0
3.2947	16.0	400	8.9514	0.4222	0.7418	0.4335	0.5266	0.4128	0.5114	0.0537	0.3031	0.5168	0.5690	0.5060	0.5974	-1.0	-1.0
3.3214	17.0	425	9.2974	0.3746	0.7435	0.3315	0.5135	0.3558	0.5423	0.0471	0.2816	0.4784	0.5190	0.4529	0.6639	-1.0	-1.0
3.2517	18.0	450	8.5708	0.4240	0.7304	0.4626	0.6015	0.4174	0.5148	0.0624	0.3235	0.5256	0.6024	0.4988	0.7055	-1.0	-1.0
3.0901	19.0	475	8.8775	0.4104	0.7342	0.3616	0.5639	0.3993	0.5030	0.0528	0.2901	0.5014	0.5810	0.4819	0.6279	-1.0	-1.0
3.0710	20.0	500	8.7070	0.3996	0.7125	0.3801	0.6014	0.3778	0.5304	0.0527	0.2933	0.5012	0.6190	0.4777	0.6424	-1.0	-1.0
3.0758	21.0	525	8.7404	0.4118	0.7191	0.4078	0.5816	0.4069	0.4996	0.0521	0.2981	0.5107	0.5833	0.4896	0.6473	-1.0	-1.0
2.9934	22.0	550	9.0060	0.4213	0.7281	0.4236	0.5762	0.4049	0.5164	0.0496	0.3129	0.5064	0.5857	0.4866	0.6318	-1.0	-1.0
2.8724	23.0	575	9.2333	0.4262	0.7492	0.3845	0.5668	0.4159	0.5020	0.0519	0.3046	0.5173	0.5667	0.5026	0.6175	-1.0	-1.0
2.8296	24.0	600	9.2622	0.4243	0.7391	0.4183	0.5527	0.4098	0.5511	0.0552	0.3105	0.5106	0.5548	0.4875	0.6740	-1.0	-1.0
2.8128	25.0	625	8.8469	0.4332	0.7532	0.4067	0.5649	0.4155	0.5792	0.0628	0.3154	0.5220	0.5881	0.4947	0.7082	-1.0	-1.0
2.8109	26.0	650	9.2135	0.4067	0.7041	0.4218	0.5677	0.4093	0.5233	0.0617	0.3198	0.5268	0.5833	0.5047	0.6823	-1.0	-1.0
2.8434	27.0	675	8.9676	0.4297	0.7278	0.4447	0.5748	0.4106	0.5597	0.0555	0.3107	0.5159	0.5833	0.4997	0.6161	-1.0	-1.0
2.7521	28.0	700	9.1216	0.4318	0.7418	0.4410	0.6021	0.4245	0.5261	0.0554	0.3018	0.5228	0.6071	0.5070	0.6202	-1.0	-1.0
2.7126	29.0	725	9.1676	0.4143	0.7209	0.4215	0.5990	0.3947	0.5432	0.0578	0.3165	0.5156	0.6000	0.4858	0.7115	-1.0	-1.0
2.7230	30.0	750	9.4190	0.3812	0.6908	0.3748	0.3792	0.3741	0.5315	0.0542	0.2857	0.4743	0.3810	0.4618	0.6188	-1.0	-1.0
2.8132	31.0	775	9.3240	0.4578	0.7573	0.4780	0.6219	0.4323	0.6013	0.0666	0.3331	0.5372	0.6238	0.5074	0.7371	-1.0	-1.0
2.6837	32.0	800	9.1286	0.4207	0.7357	0.4057	0.5849	0.3971	0.5545	0.0571	0.3127	0.5144	0.5952	0.4970	0.6216	-1.0	-1.0
2.5350	33.0	825	9.3508	0.4084	0.7225	0.4179	0.6298	0.3932	0.5131	0.0504	0.3142	0.5089	0.6405	0.4839	0.6532	-1.0	-1.0
2.5510	34.0	850	9.3640	0.4034	0.7085	0.3874	0.5713	0.3815	0.5396	0.0521	0.3233	0.5079	0.5714	0.4822	0.6784	-1.0	-1.0
2.6408	35.0	875	9.6026	0.4253	0.7299	0.4322	0.5708	0.4070	0.5435	0.0548	0.3173	0.5155	0.5857	0.4964	0.6386	-1.0	-1.0
2.4881	36.0	900	9.4144	0.4137	0.7247	0.4147	0.5757	0.3872	0.5581	0.0579	0.3107	0.5124	0.5762	0.4920	0.6498	-1.0	-1.0
2.5174	37.0	925	9.1971	0.4251	0.7162	0.4269	0.5874	0.4031	0.5559	0.0629	0.3126	0.5080	0.5929	0.4893	0.6263	-1.0	-1.0
2.4497	38.0	950	9.1950	0.4299	0.7362	0.4289	0.5774	0.4060	0.6253	0.0650	0.3231	0.5239	0.5786	0.4876	0.7819	-1.0	-1.0
2.3860	39.0	975	9.5015	0.3930	0.6898	0.3825	0.5594	0.3736	0.5632	0.0627	0.3136	0.5085	0.5595	0.4738	0.7579	-1.0	-1.0
2.3519	40.0	1000	9.4528	0.4169	0.7194	0.4121	0.5626	0.3970	0.5680	0.0596	0.3261	0.5174	0.5643	0.4877	0.7301	-1.0	-1.0
2.3449	41.0	1025	9.7886	0.3988	0.7021	0.3926	0.5783	0.3902	0.5016	0.0486	0.2949	0.4947	0.5881	0.4764	0.6070	-1.0	-1.0

Framework versions

• Transformers 5.3.0.dev0
• Pytorch 2.10.0+cu128
• Datasets 4.8.2
• Tokenizers 0.22.2