segformer-b0-finetuned-net-4Sep

This model is a fine-tuned version of PushkarA07/segformer-b0-finetuned-net-4Sep on the PushkarA07/batch2-tiles_W5 dataset. It achieves the following results on the evaluation set:

• Loss: 0.0043
• Mean Iou: 0.8489
• Mean Accuracy: 0.8926
• Overall Accuracy: 0.9984
• Accuracy Abnormality: 0.7857
• Iou Abnormality: 0.6994

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: 6e-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: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Abnormality	Iou Abnormality
0.0105	0.5556	10	0.0057	0.8215	0.8853	0.9979	0.7715	0.6451
0.0053	1.1111	20	0.0052	0.8205	0.8583	0.9981	0.7172	0.6430
0.0123	1.6667	30	0.0052	0.8345	0.8931	0.9981	0.7870	0.6710
0.0048	2.2222	40	0.0048	0.8348	0.8720	0.9982	0.7445	0.6713
0.0102	2.7778	50	0.0049	0.8308	0.8649	0.9982	0.7303	0.6635
0.0037	3.3333	60	0.0048	0.8355	0.8789	0.9982	0.7585	0.6728
0.0059	3.8889	70	0.0048	0.8404	0.8860	0.9983	0.7726	0.6825
0.0071	4.4444	80	0.0046	0.8400	0.8780	0.9983	0.7565	0.6817
0.0029	5.0	90	0.0046	0.8422	0.8825	0.9983	0.7655	0.6861
0.008	5.5556	100	0.0045	0.8429	0.8834	0.9983	0.7673	0.6876
0.0054	6.1111	110	0.0047	0.8426	0.8909	0.9983	0.7825	0.6870
0.0059	6.6667	120	0.0047	0.8322	0.8613	0.9982	0.7230	0.6661
0.0097	7.2222	130	0.0049	0.8431	0.9026	0.9982	0.8060	0.6879
0.0063	7.7778	140	0.0045	0.8377	0.8707	0.9983	0.7419	0.6771
0.0102	8.3333	150	0.0047	0.8346	0.8706	0.9982	0.7418	0.6711
0.0065	8.8889	160	0.0046	0.8438	0.8927	0.9983	0.7861	0.6893
0.004	9.4444	170	0.0046	0.8427	0.8805	0.9983	0.7615	0.6870
0.0057	10.0	180	0.0046	0.8469	0.9000	0.9983	0.8007	0.6956
0.0097	10.5556	190	0.0044	0.8430	0.8789	0.9983	0.7582	0.6877
0.0059	11.1111	200	0.0045	0.8467	0.8955	0.9983	0.7918	0.6951
0.0124	11.6667	210	0.0044	0.8438	0.8820	0.9983	0.7645	0.6893
0.0105	12.2222	220	0.0044	0.8470	0.8905	0.9983	0.7817	0.6956
0.0043	12.7778	230	0.0044	0.8472	0.8877	0.9984	0.7760	0.6961
0.0041	13.3333	240	0.0044	0.8438	0.8772	0.9983	0.7549	0.6892
0.0083	13.8889	250	0.0044	0.8427	0.8752	0.9983	0.7509	0.6870
0.0089	14.4444	260	0.0044	0.8424	0.8774	0.9983	0.7554	0.6864
0.0034	15.0	270	0.0044	0.8455	0.8885	0.9983	0.7776	0.6927
0.0044	15.5556	280	0.0044	0.8456	0.8879	0.9983	0.7764	0.6928
0.0052	16.1111	290	0.0044	0.8450	0.8859	0.9983	0.7724	0.6916
0.0055	16.6667	300	0.0045	0.8470	0.8964	0.9983	0.7935	0.6957
0.0045	17.2222	310	0.0044	0.8462	0.8896	0.9983	0.7797	0.6941
0.0048	17.7778	320	0.0045	0.8449	0.8902	0.9983	0.7811	0.6914
0.0048	18.3333	330	0.0044	0.8449	0.8833	0.9983	0.7671	0.6914
0.0044	18.8889	340	0.0044	0.8466	0.8888	0.9983	0.7783	0.6948
0.0061	19.4444	350	0.0045	0.8483	0.9034	0.9983	0.8075	0.6983
0.0026	20.0	360	0.0044	0.8461	0.8838	0.9984	0.7682	0.6939
0.008	20.5556	370	0.0044	0.8488	0.8936	0.9984	0.7879	0.6993
0.0047	21.1111	380	0.0044	0.8482	0.8955	0.9983	0.7916	0.6980
0.0109	21.6667	390	0.0045	0.8465	0.8964	0.9983	0.7935	0.6947
0.007	22.2222	400	0.0045	0.8440	0.8788	0.9983	0.7581	0.6897
0.0101	22.7778	410	0.0044	0.8464	0.8914	0.9983	0.7835	0.6945
0.0038	23.3333	420	0.0044	0.8425	0.8802	0.9983	0.7608	0.6867
0.0049	23.8889	430	0.0045	0.8463	0.8916	0.9983	0.7838	0.6943
0.0074	24.4444	440	0.0044	0.8455	0.8861	0.9983	0.7727	0.6927
0.0065	25.0	450	0.0044	0.8466	0.8861	0.9984	0.7728	0.6948
0.0056	25.5556	460	0.0046	0.8490	0.9117	0.9983	0.8243	0.6997
0.0047	26.1111	470	0.0043	0.8469	0.8839	0.9984	0.7682	0.6955
0.0043	26.6667	480	0.0045	0.8475	0.8983	0.9983	0.7973	0.6967
0.0051	27.2222	490	0.0044	0.8430	0.8810	0.9983	0.7625	0.6877
0.0131	27.7778	500	0.0044	0.8477	0.8903	0.9984	0.7812	0.6971
0.0042	28.3333	510	0.0044	0.8489	0.9029	0.9983	0.8066	0.6996
0.0074	28.8889	520	0.0043	0.8454	0.8827	0.9983	0.7660	0.6924
0.002	29.4444	530	0.0044	0.8461	0.8892	0.9983	0.7789	0.6939
0.0125	30.0	540	0.0043	0.8470	0.8896	0.9983	0.7797	0.6957
0.0054	30.5556	550	0.0044	0.8471	0.8928	0.9983	0.7862	0.6959
0.009	31.1111	560	0.0044	0.8399	0.8690	0.9983	0.7383	0.6814
0.0074	31.6667	570	0.0044	0.8493	0.8937	0.9984	0.7880	0.7002
0.0048	32.2222	580	0.0045	0.8463	0.8897	0.9983	0.7801	0.6943
0.004	32.7778	590	0.0044	0.8479	0.8955	0.9983	0.7917	0.6975
0.0047	33.3333	600	0.0045	0.8488	0.9040	0.9983	0.8088	0.6993
0.0039	33.8889	610	0.0044	0.8476	0.8909	0.9983	0.7825	0.6968
0.0047	34.4444	620	0.0044	0.8483	0.8937	0.9983	0.7880	0.6983
0.0038	35.0	630	0.0044	0.8496	0.8980	0.9984	0.7966	0.7009
0.0028	35.5556	640	0.0043	0.8486	0.8928	0.9984	0.7861	0.6988
0.0129	36.1111	650	0.0043	0.8491	0.8935	0.9984	0.7877	0.6998
0.0055	36.6667	660	0.0044	0.8499	0.9056	0.9983	0.8119	0.7014
0.0089	37.2222	670	0.0044	0.8481	0.8899	0.9984	0.7804	0.6979
0.0042	37.7778	680	0.0043	0.8501	0.8935	0.9984	0.7877	0.7018
0.0068	38.3333	690	0.0044	0.8451	0.8826	0.9983	0.7657	0.6919
0.0036	38.8889	700	0.0044	0.8506	0.8978	0.9984	0.7963	0.7028
0.0129	39.4444	710	0.0043	0.8488	0.8904	0.9984	0.7814	0.6993
0.0043	40.0	720	0.0043	0.8491	0.8934	0.9984	0.7874	0.6999
0.0036	40.5556	730	0.0044	0.8410	0.8734	0.9983	0.7473	0.6837
0.0103	41.1111	740	0.0045	0.8497	0.9069	0.9983	0.8145	0.7011
0.0071	41.6667	750	0.0043	0.8448	0.8783	0.9984	0.7571	0.6913
0.0072	42.2222	760	0.0044	0.8501	0.8945	0.9984	0.7897	0.7018
0.0038	42.7778	770	0.0043	0.8495	0.8916	0.9984	0.7839	0.7007
0.0049	43.3333	780	0.0043	0.8449	0.8777	0.9984	0.7558	0.6915
0.0062	43.8889	790	0.0044	0.8495	0.8958	0.9984	0.7922	0.7006
0.0044	44.4444	800	0.0043	0.8492	0.8913	0.9984	0.7832	0.7001
0.0034	45.0	810	0.0044	0.8480	0.8854	0.9984	0.7713	0.6976
0.0033	45.5556	820	0.0043	0.8486	0.8903	0.9984	0.7813	0.6989
0.0053	46.1111	830	0.0044	0.8481	0.8982	0.9983	0.7970	0.6980
0.0025	46.6667	840	0.0044	0.8460	0.8855	0.9983	0.7716	0.6937
0.0069	47.2222	850	0.0044	0.8463	0.8827	0.9984	0.7660	0.6942
0.0031	47.7778	860	0.0043	0.8498	0.8917	0.9984	0.7840	0.7013
0.006	48.3333	870	0.0043	0.8462	0.8820	0.9984	0.7645	0.6941
0.0025	48.8889	880	0.0044	0.8498	0.8956	0.9984	0.7918	0.7011
0.0075	49.4444	890	0.0043	0.8498	0.8915	0.9984	0.7835	0.7013
0.0064	50.0	900	0.0043	0.8500	0.8931	0.9984	0.7867	0.7016
0.0067	50.5556	910	0.0043	0.8518	0.9003	0.9984	0.8013	0.7052
0.0046	51.1111	920	0.0043	0.8496	0.8913	0.9984	0.7832	0.7008
0.0075	51.6667	930	0.0044	0.8483	0.8884	0.9984	0.7775	0.6983
0.0102	52.2222	940	0.0044	0.8506	0.8991	0.9984	0.7988	0.7028
0.0056	52.7778	950	0.0043	0.8488	0.8909	0.9984	0.7824	0.6991
0.0066	53.3333	960	0.0043	0.8451	0.8800	0.9984	0.7605	0.6918
0.0073	53.8889	970	0.0044	0.8519	0.9031	0.9984	0.8070	0.7055
0.0054	54.4444	980	0.0043	0.8445	0.8772	0.9984	0.7549	0.6906
0.0066	55.0	990	0.0043	0.8506	0.8955	0.9984	0.7916	0.7028
0.0047	55.5556	1000	0.0043	0.8523	0.9037	0.9984	0.8082	0.7062
0.0083	56.1111	1010	0.0043	0.8492	0.8892	0.9984	0.7789	0.7001
0.0022	56.6667	1020	0.0043	0.8483	0.8866	0.9984	0.7737	0.6983
0.011	57.2222	1030	0.0043	0.8439	0.8757	0.9984	0.7519	0.6894
0.0033	57.7778	1040	0.0043	0.8496	0.8967	0.9984	0.7940	0.7008
0.0051	58.3333	1050	0.0044	0.8504	0.8978	0.9984	0.7963	0.7025
0.0029	58.8889	1060	0.0043	0.8491	0.8895	0.9984	0.7796	0.6999
0.0052	59.4444	1070	0.0043	0.8465	0.8812	0.9984	0.7628	0.6946
0.0036	60.0	1080	0.0043	0.8511	0.8999	0.9984	0.8005	0.7038
0.0026	60.5556	1090	0.0043	0.8498	0.8899	0.9984	0.7804	0.7012
0.0023	61.1111	1100	0.0043	0.8498	0.8906	0.9984	0.7818	0.7013
0.0087	61.6667	1110	0.0043	0.8464	0.8817	0.9984	0.7638	0.6945
0.0057	62.2222	1120	0.0043	0.8495	0.8914	0.9984	0.7833	0.7006
0.004	62.7778	1130	0.0043	0.8482	0.8865	0.9984	0.7735	0.6980
0.0087	63.3333	1140	0.0043	0.8506	0.8979	0.9984	0.7964	0.7029
0.0087	63.8889	1150	0.0043	0.8488	0.8900	0.9984	0.7805	0.6993
0.004	64.4444	1160	0.0044	0.8502	0.8983	0.9984	0.7972	0.7020
0.0037	65.0	1170	0.0044	0.8501	0.8994	0.9984	0.7996	0.7018
0.0084	65.5556	1180	0.0043	0.8485	0.8884	0.9984	0.7773	0.6986
0.0034	66.1111	1190	0.0044	0.8496	0.8953	0.9984	0.7912	0.7008
0.0063	66.6667	1200	0.0044	0.8498	0.9003	0.9983	0.8013	0.7012
0.0091	67.2222	1210	0.0043	0.8482	0.8885	0.9984	0.7775	0.6980
0.009	67.7778	1220	0.0043	0.8475	0.8857	0.9984	0.7719	0.6967
0.0054	68.3333	1230	0.0043	0.8493	0.8939	0.9984	0.7885	0.7003
0.0043	68.8889	1240	0.0043	0.8477	0.8904	0.9984	0.7814	0.6972
0.0046	69.4444	1250	0.0043	0.8476	0.8884	0.9984	0.7773	0.6968
0.0053	70.0	1260	0.0043	0.8445	0.8777	0.9984	0.7559	0.6907
0.0066	70.5556	1270	0.0043	0.8491	0.8903	0.9984	0.7811	0.6999
0.0044	71.1111	1280	0.0043	0.8480	0.8874	0.9984	0.7752	0.6976
0.0073	71.6667	1290	0.0043	0.8490	0.8921	0.9984	0.7848	0.6995
0.0035	72.2222	1300	0.0043	0.8486	0.8910	0.9984	0.7825	0.6989
0.0074	72.7778	1310	0.0044	0.8481	0.8896	0.9984	0.7798	0.6979
0.006	73.3333	1320	0.0044	0.8487	0.8949	0.9983	0.7904	0.6990
0.0062	73.8889	1330	0.0044	0.8492	0.8984	0.9983	0.7975	0.7001
0.0113	74.4444	1340	0.0043	0.8480	0.8908	0.9984	0.7821	0.6976
0.004	75.0	1350	0.0043	0.8452	0.8807	0.9984	0.7619	0.6920
0.0066	75.5556	1360	0.0043	0.8464	0.8844	0.9984	0.7693	0.6945
0.0059	76.1111	1370	0.0043	0.8486	0.8941	0.9984	0.7889	0.6989
0.0021	76.6667	1380	0.0044	0.8489	0.8997	0.9983	0.8000	0.6995
0.0034	77.2222	1390	0.0044	0.8476	0.8947	0.9983	0.7901	0.6969
0.0115	77.7778	1400	0.0044	0.8471	0.8894	0.9983	0.7794	0.6959
0.0064	78.3333	1410	0.0044	0.8480	0.8959	0.9983	0.7924	0.6976
0.0041	78.8889	1420	0.0044	0.8474	0.8926	0.9983	0.7858	0.6965
0.0062	79.4444	1430	0.0044	0.8487	0.8921	0.9984	0.7848	0.6991
0.0064	80.0	1440	0.0044	0.8486	0.8910	0.9984	0.7825	0.6989
0.0066	80.5556	1450	0.0044	0.8481	0.8896	0.9984	0.7799	0.6978
0.0039	81.1111	1460	0.0044	0.8490	0.8964	0.9983	0.7934	0.6998
0.0061	81.6667	1470	0.0044	0.8497	0.9002	0.9983	0.8011	0.7011
0.006	82.2222	1480	0.0044	0.8494	0.8962	0.9984	0.7930	0.7005
0.0036	82.7778	1490	0.0044	0.8492	0.8937	0.9984	0.7880	0.7001
0.0154	83.3333	1500	0.0044	0.8472	0.8868	0.9984	0.7741	0.6960
0.0063	83.8889	1510	0.0044	0.8465	0.8842	0.9984	0.7690	0.6946
0.0085	84.4444	1520	0.0044	0.8468	0.8878	0.9984	0.7761	0.6954
0.005	85.0	1530	0.0044	0.8491	0.8948	0.9984	0.7902	0.6999
0.0061	85.5556	1540	0.0044	0.8502	0.8995	0.9984	0.7996	0.7021
0.0144	86.1111	1550	0.0044	0.8488	0.8938	0.9984	0.7883	0.6992
0.0075	86.6667	1560	0.0044	0.8484	0.8923	0.9984	0.7851	0.6984
0.0119	87.2222	1570	0.0044	0.8482	0.8926	0.9984	0.7859	0.6981
0.0038	87.7778	1580	0.0044	0.8486	0.8952	0.9983	0.7910	0.6990
0.0047	88.3333	1590	0.0044	0.8486	0.8971	0.9983	0.7949	0.6989
0.0059	88.8889	1600	0.0044	0.8487	0.8956	0.9983	0.7918	0.6990
0.0035	89.4444	1610	0.0044	0.8479	0.8910	0.9984	0.7826	0.6974
0.0058	90.0	1620	0.0044	0.8477	0.8908	0.9984	0.7822	0.6971
0.0067	90.5556	1630	0.0044	0.8483	0.8943	0.9983	0.7891	0.6982
0.009	91.1111	1640	0.0044	0.8479	0.8917	0.9984	0.7840	0.6975
0.0042	91.6667	1650	0.0044	0.8476	0.8912	0.9983	0.7829	0.6969
0.0024	92.2222	1660	0.0044	0.8480	0.8923	0.9984	0.7851	0.6977
0.0055	92.7778	1670	0.0044	0.8474	0.8905	0.9983	0.7816	0.6966
0.0041	93.3333	1680	0.0044	0.8484	0.8932	0.9984	0.7870	0.6985
0.0041	93.8889	1690	0.0044	0.8491	0.8957	0.9984	0.7920	0.6998
0.0086	94.4444	1700	0.0044	0.8496	0.8977	0.9984	0.7962	0.7008
0.0086	95.0	1710	0.0044	0.8499	0.8994	0.9983	0.7995	0.7014
0.0081	95.5556	1720	0.0044	0.8497	0.8970	0.9984	0.7946	0.7010
0.0066	96.1111	1730	0.0044	0.8494	0.8962	0.9984	0.7931	0.7005
0.0056	96.6667	1740	0.0044	0.8494	0.8949	0.9984	0.7903	0.7005
0.0071	97.2222	1750	0.0043	0.8494	0.8944	0.9984	0.7894	0.7005
0.013	97.7778	1760	0.0043	0.8489	0.8923	0.9984	0.7852	0.6995
0.0044	98.3333	1770	0.0043	0.8496	0.8950	0.9984	0.7906	0.7008
0.0049	98.8889	1780	0.0043	0.8490	0.8929	0.9984	0.7865	0.6996
0.0079	99.4444	1790	0.0043	0.8487	0.8924	0.9984	0.7854	0.6991
0.0065	100.0	1800	0.0043	0.8489	0.8926	0.9984	0.7857	0.6994

Framework versions

• Transformers 4.56.0
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.0