segformer-b0-finetuned-morphpadver1-no_ckp-2

This model is a fine-tuned version of nvidia/mit-b0 on the NICOPOI-9/morphpad_512_4class dataset. It achieves the following results on the evaluation set:

• Loss: 0.0023
• Mean Iou: 0.9987
• Mean Accuracy: 0.9994
• Overall Accuracy: 0.9994
• Accuracy 0-0: 0.9992
• Accuracy 0-90: 0.9991
• Accuracy 90-0: 0.9994
• Accuracy 90-90: 0.9998
• Iou 0-0: 0.9991
• Iou 0-90: 0.9984
• Iou 90-0: 0.9984
• Iou 90-90: 0.9991

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: 1
• eval_batch_size: 1
• 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: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy 0-0	Accuracy 0-90	Accuracy 90-0	Accuracy 90-90	Iou 0-0	Iou 0-90	Iou 90-0	Iou 90-90
0.2037	0.1112	1000	0.1512	0.8947	0.9437	0.9447	0.9614	0.9508	0.8858	0.9769	0.9353	0.8554	0.8502	0.9381
0.0768	0.2224	2000	0.0772	0.9450	0.9715	0.9720	0.9772	0.9613	0.9599	0.9878	0.9643	0.9249	0.9246	0.9664
0.1051	0.3337	3000	0.0555	0.9622	0.9807	0.9809	0.9885	0.9656	0.9794	0.9893	0.9798	0.9440	0.9446	0.9804
0.0426	0.4449	4000	0.0438	0.9706	0.9851	0.9852	0.9908	0.9785	0.9805	0.9905	0.9852	0.9557	0.9558	0.9858
0.03	0.5561	5000	0.0386	0.9735	0.9866	0.9867	0.9897	0.9893	0.9758	0.9916	0.9866	0.9589	0.9602	0.9886
0.0255	0.6673	6000	0.0314	0.9788	0.9892	0.9894	0.9926	0.9796	0.9880	0.9968	0.9902	0.9671	0.9664	0.9914
0.0446	0.7786	7000	0.0353	0.9822	0.9910	0.9911	0.9935	0.9873	0.9890	0.9944	0.9915	0.9727	0.9725	0.9921
0.0188	0.8898	8000	0.0323	0.9836	0.9917	0.9919	0.9964	0.9861	0.9882	0.9962	0.9932	0.9741	0.9730	0.9941
0.017	1.0010	9000	0.0167	0.9895	0.9947	0.9948	0.9966	0.9916	0.9931	0.9977	0.9946	0.9844	0.9842	0.9951
0.0135	1.1122	10000	0.0191	0.9890	0.9945	0.9946	0.9975	0.9882	0.9949	0.9974	0.9952	0.9829	0.9825	0.9955
0.0164	1.2234	11000	0.0147	0.9912	0.9956	0.9956	0.9979	0.9942	0.9922	0.9980	0.9957	0.9867	0.9863	0.9961
0.0602	1.3347	12000	0.0195	0.9913	0.9956	0.9957	0.9972	0.9929	0.9937	0.9986	0.9960	0.9866	0.9864	0.9962
0.0095	1.4459	13000	0.0229	0.9899	0.9949	0.9950	0.9970	0.9941	0.9899	0.9987	0.9959	0.9840	0.9832	0.9966
0.0434	1.5571	14000	0.0094	0.9942	0.9971	0.9971	0.9985	0.9947	0.9965	0.9987	0.9970	0.9913	0.9914	0.9971
0.0108	1.6683	15000	0.0110	0.9935	0.9967	0.9968	0.9983	0.9931	0.9967	0.9989	0.9973	0.9897	0.9896	0.9974
0.0091	1.7796	16000	0.0094	0.9946	0.9973	0.9973	0.9981	0.9954	0.9973	0.9983	0.9972	0.9919	0.9918	0.9973
0.0417	1.8908	17000	0.0123	0.9930	0.9965	0.9966	0.9969	0.9931	0.9972	0.9988	0.9961	0.9887	0.9898	0.9976
0.0101	2.0020	18000	0.0080	0.9954	0.9977	0.9977	0.9980	0.9959	0.9982	0.9987	0.9975	0.9933	0.9930	0.9978
0.0072	2.1132	19000	0.0193	0.9943	0.9971	0.9972	0.9981	0.9939	0.9978	0.9987	0.9975	0.9911	0.9906	0.9978
0.0077	2.2244	20000	0.0056	0.9967	0.9984	0.9984	0.9986	0.9975	0.9984	0.9990	0.9980	0.9956	0.9952	0.9981
0.0048	2.3357	21000	0.0052	0.9969	0.9985	0.9985	0.9991	0.9973	0.9981	0.9994	0.9983	0.9956	0.9955	0.9983
0.0052	2.4469	22000	0.0155	0.9962	0.9981	0.9981	0.9992	0.9948	0.9989	0.9994	0.9982	0.9941	0.9941	0.9983
0.0042	2.5581	23000	0.0042	0.9976	0.9988	0.9988	0.9989	0.9979	0.9990	0.9994	0.9984	0.9967	0.9967	0.9985
0.004	2.6693	24000	0.0041	0.9976	0.9988	0.9988	0.9995	0.9977	0.9986	0.9994	0.9984	0.9967	0.9968	0.9986
0.0044	2.7806	25000	0.0056	0.9971	0.9986	0.9986	0.9992	0.9968	0.9987	0.9995	0.9987	0.9956	0.9956	0.9986
0.0036	2.8918	26000	0.0236	0.9960	0.9980	0.9980	0.9988	0.9946	0.9991	0.9995	0.9985	0.9935	0.9932	0.9987
0.0038	3.0030	27000	0.0156	0.9969	0.9985	0.9985	0.9992	0.9963	0.9989	0.9996	0.9988	0.9952	0.9951	0.9987
0.0028	3.1142	28000	0.0037	0.9979	0.9989	0.9990	0.9992	0.9978	0.9992	0.9995	0.9988	0.9970	0.9969	0.9988
0.0044	3.2254	29000	0.0030	0.9983	0.9992	0.9992	0.9993	0.9987	0.9993	0.9994	0.9989	0.9978	0.9978	0.9989
0.0051	3.3367	30000	0.0111	0.9975	0.9987	0.9988	0.9993	0.9969	0.9990	0.9998	0.9989	0.9961	0.9961	0.9989
0.0027	3.4479	31000	0.0176	0.9964	0.9982	0.9982	0.9995	0.9945	0.9992	0.9996	0.9990	0.9939	0.9939	0.9989
0.0126	3.5591	32000	0.0095	0.9973	0.9986	0.9987	0.9996	0.9963	0.9992	0.9995	0.9990	0.9956	0.9956	0.9989
0.0097	3.6703	33000	0.0044	0.9973	0.9986	0.9987	0.9997	0.9962	0.9989	0.9997	0.9989	0.9956	0.9958	0.9988
0.0066	3.7816	34000	0.0030	0.9982	0.9991	0.9991	0.9996	0.9977	0.9993	0.9997	0.9990	0.9974	0.9976	0.9987
0.0024	3.8928	35000	0.0025	0.9987	0.9993	0.9993	0.9996	0.9990	0.9991	0.9996	0.9991	0.9983	0.9982	0.9991
0.0032	4.0040	36000	0.0026	0.9985	0.9992	0.9992	0.9996	0.9983	0.9994	0.9996	0.9990	0.9979	0.9979	0.9990
0.003	4.1152	37000	0.0025	0.9986	0.9993	0.9993	0.9995	0.9987	0.9993	0.9997	0.9991	0.9981	0.9981	0.9990
0.0023	4.2264	38000	0.0023	0.9987	0.9994	0.9994	0.9996	0.9989	0.9992	0.9997	0.9991	0.9984	0.9984	0.9991
0.0068	4.3377	39000	0.0025	0.9987	0.9993	0.9993	0.9993	0.9988	0.9995	0.9997	0.9991	0.9983	0.9982	0.9990
0.0058	4.4489	40000	0.0030	0.9980	0.9990	0.9990	0.9996	0.9970	0.9996	0.9997	0.9991	0.9968	0.9971	0.9989
0.0039	4.5601	41000	0.0022	0.9989	0.9994	0.9994	0.9995	0.9993	0.9993	0.9997	0.9992	0.9985	0.9985	0.9992
0.0021	4.6713	42000	0.0022	0.9988	0.9994	0.9994	0.9995	0.9990	0.9994	0.9997	0.9992	0.9984	0.9984	0.9991
0.0016	4.7826	43000	0.0023	0.9987	0.9993	0.9993	0.9995	0.9987	0.9995	0.9996	0.9992	0.9982	0.9981	0.9992
0.0283	4.8938	44000	0.0023	0.9987	0.9994	0.9994	0.9992	0.9991	0.9994	0.9998	0.9991	0.9984	0.9984	0.9991

Framework versions

• Transformers 4.48.3
• Pytorch 2.1.0
• Datasets 3.2.0
• Tokenizers 0.21.0