dfine-small-remia-cartons

This model is a fine-tuned version of ustc-community/dfine-small-coco on the Nevico/remia-ned1-cam1-cartons dataset. It achieves the following results on the evaluation set:

• Loss: 0.6602
• Map: 0.9785
• Map 50: 0.99
• Map 75: 0.9793
• Map Small: 0.835
• Map Medium: 0.9765
• Map Large: 0.9847
• Mar 1: 0.1806
• Mar 10: 0.9926
• Mar 100: 0.9964
• Mar Small: 1.0
• Mar Medium: 0.9941
• Mar Large: 0.9996
• Map Carton: 0.9785
• Mar 100 Carton: 0.9964

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
• num_epochs: 40.0

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 Carton	Mar 100 Carton
No log	1.0	131	3.2220	0.8931	0.9245	0.9101	0.0	0.9035	0.892	0.1685	0.9586	0.9803	0.0	0.9768	0.9929	0.8931	0.9803
No log	2.0	262	1.5777	0.9448	0.9707	0.9601	0.9505	0.9473	0.9449	0.1783	0.9719	0.9885	0.95	0.9822	0.9976	0.9448	0.9885
No log	3.0	393	1.2223	0.9549	0.976	0.9662	1.0	0.9551	0.9594	0.1805	0.9767	0.9901	1.0	0.9858	0.9961	0.9549	0.9901
15.7996	4.0	524	1.0484	0.9609	0.9774	0.9691	1.0	0.9639	0.9599	0.1801	0.9833	0.9951	1.0	0.9929	0.998	0.9609	0.9951
15.7996	5.0	655	0.9448	0.9627	0.9802	0.9706	1.0	0.966	0.9618	0.1805	0.9782	0.9943	1.0	0.9921	0.9972	0.9627	0.9943
15.7996	6.0	786	0.9046	0.9665	0.9818	0.9739	0.6667	0.9688	0.9672	0.1821	0.9836	0.9947	1.0	0.9924	0.998	0.9665	0.9947
15.7996	7.0	917	0.9322	0.9636	0.9831	0.9714	0.6667	0.9645	0.9648	0.1806	0.9829	0.9936	1.0	0.9898	0.9988	0.9636	0.9936
8.6003	8.0	1048	0.8879	0.9682	0.9851	0.9757	0.835	0.9714	0.9676	0.1821	0.9846	0.9961	1.0	0.9938	0.9992	0.9682	0.9961
8.6003	9.0	1179	0.8388	0.9641	0.9825	0.9711	0.835	0.9654	0.9654	0.1806	0.9814	0.9946	1.0	0.9921	0.998	0.9641	0.9946
8.6003	10.0	1310	0.7875	0.9703	0.9861	0.9747	0.835	0.9706	0.9725	0.1805	0.9867	0.9952	1.0	0.9938	0.9972	0.9703	0.9952
8.6003	11.0	1441	0.7985	0.9673	0.9838	0.9742	0.6667	0.9655	0.973	0.1805	0.9834	0.9947	1.0	0.9912	0.9996	0.9673	0.9947
8.1331	12.0	1572	0.8073	0.9687	0.9843	0.9738	0.835	0.9711	0.9695	0.1805	0.9842	0.9959	1.0	0.9941	0.9984	0.9687	0.9959
8.1331	13.0	1703	0.7851	0.9697	0.9841	0.9741	1.0	0.9727	0.9687	0.1823	0.9857	0.9959	1.0	0.9932	0.9996	0.9697	0.9959
8.1331	14.0	1834	0.7909	0.9693	0.9852	0.9732	0.835	0.9685	0.9743	0.1803	0.9857	0.9946	1.0	0.9926	0.9972	0.9693	0.9946
8.1331	15.0	1965	0.7394	0.9713	0.9856	0.9748	1.0	0.9719	0.9744	0.179	0.9854	0.9962	1.0	0.9935	1.0	0.9713	0.9962
7.7791	16.0	2096	0.7798	0.9737	0.9873	0.976	0.6667	0.9741	0.9764	0.1823	0.9895	0.9964	1.0	0.9943	0.9992	0.9737	0.9964
7.7791	17.0	2227	0.7578	0.9717	0.9868	0.9746	1.0	0.9706	0.9764	0.1803	0.9869	0.9962	1.0	0.9935	1.0	0.9717	0.9962
7.7791	18.0	2358	0.7157	0.9757	0.9866	0.9809	0.835	0.9767	0.9773	0.1823	0.9869	0.9974	1.0	0.9955	1.0	0.9757	0.9974
7.7791	19.0	2489	0.7317	0.9736	0.9862	0.9752	0.835	0.9748	0.9764	0.1823	0.9862	0.9967	1.0	0.9943	1.0	0.9736	0.9967
7.6618	20.0	2620	0.7203	0.9727	0.986	0.9752	0.835	0.9759	0.9741	0.1805	0.989	0.9964	1.0	0.9941	0.9996	0.9727	0.9964
7.6618	21.0	2751	0.7176	0.9738	0.9881	0.9761	0.6667	0.9744	0.9767	0.1821	0.9882	0.9952	1.0	0.9932	0.998	0.9738	0.9952
7.6618	22.0	2882	0.7231	0.9719	0.9856	0.9749	0.835	0.975	0.973	0.1823	0.9874	0.9961	1.0	0.9935	0.9996	0.9719	0.9961
7.4825	23.0	3013	0.7240	0.9742	0.9868	0.9753	0.6667	0.9749	0.9787	0.1823	0.988	0.9967	1.0	0.9943	1.0	0.9742	0.9967
7.4825	24.0	3144	0.7183	0.9754	0.9866	0.9764	0.835	0.9744	0.9804	0.1806	0.989	0.9969	1.0	0.9946	1.0	0.9754	0.9969
7.4825	25.0	3275	0.7135	0.9768	0.9889	0.9785	0.835	0.9773	0.9801	0.1823	0.9883	0.9964	1.0	0.9943	0.9992	0.9768	0.9964
7.4825	26.0	3406	0.7105	0.9773	0.9883	0.9781	0.835	0.9765	0.9831	0.1821	0.99	0.997	1.0	0.9949	1.0	0.9773	0.997
7.3980	27.0	3537	0.6886	0.9773	0.9897	0.9783	0.835	0.9759	0.9822	0.1823	0.9895	0.9961	1.0	0.9943	0.9984	0.9773	0.9961
7.3980	28.0	3668	0.6826	0.9761	0.9887	0.9776	0.6667	0.9757	0.9819	0.1823	0.9926	0.9962	1.0	0.9938	0.9996	0.9761	0.9962
7.3980	29.0	3799	0.6752	0.9778	0.9891	0.9779	0.835	0.9762	0.9848	0.1823	0.9913	0.9966	1.0	0.9941	1.0	0.9778	0.9966
7.3980	30.0	3930	0.6758	0.9757	0.9858	0.9795	1.0	0.9769	0.9772	0.1821	0.9882	0.9962	1.0	0.9946	0.9984	0.9757	0.9962
7.3320	31.0	4061	0.6754	0.978	0.9889	0.9789	0.835	0.9778	0.9834	0.1823	0.9916	0.9967	1.0	0.9946	0.9996	0.978	0.9967
7.3320	32.0	4192	0.6749	0.9782	0.9896	0.9784	1.0	0.9766	0.9846	0.1823	0.9906	0.9972	1.0	0.9952	1.0	0.9782	0.9972
7.3320	33.0	4323	0.6595	0.9785	0.99	0.9791	0.835	0.9765	0.9854	0.1823	0.9928	0.9966	1.0	0.9941	1.0	0.9785	0.9966
7.3320	34.0	4454	0.6726	0.9779	0.9888	0.9793	0.835	0.9781	0.9816	0.1823	0.9895	0.9961	1.0	0.9946	0.998	0.9779	0.9961
7.2814	35.0	4585	0.6961	0.9774	0.9895	0.9782	0.835	0.9755	0.984	0.1823	0.989	0.9964	1.0	0.9941	0.9996	0.9774	0.9964
7.2814	36.0	4716	0.6575	0.9774	0.9886	0.9781	0.835	0.9778	0.9831	0.1823	0.9906	0.9966	1.0	0.9943	0.9996	0.9774	0.9966
7.2814	37.0	4847	0.6977	0.9767	0.9885	0.9781	0.6667	0.9768	0.9817	0.1823	0.9901	0.9964	1.0	0.9943	0.9992	0.9767	0.9964
7.2814	38.0	4978	0.6628	0.9771	0.9882	0.9789	0.835	0.9776	0.9813	0.1823	0.9903	0.9962	1.0	0.9941	0.9992	0.9771	0.9962
7.1800	39.0	5109	0.6696	0.9767	0.9887	0.9788	0.835	0.977	0.9818	0.1823	0.9901	0.9964	1.0	0.9941	0.9996	0.9767	0.9964
7.1800	40.0	5240	0.6845	0.9771	0.9889	0.9786	0.835	0.9776	0.9819	0.1823	0.989	0.9961	1.0	0.9938	0.9992	0.9771	0.9961

Framework versions

• Transformers 5.13.1
• Pytorch 2.13.0+cu130
• Datasets 5.0.0
• Tokenizers 0.22.2