segformer-finetuned-segments-riceleafdisease-dec-18

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

• Loss: 0.0654
• Mean Iou: 0.8022
• Mean Accuracy: 0.8495
• Overall Accuracy: 0.9793
• Accuracy Unlabelled: nan
• Accuracy Healthy: 0.9379
• Accuracy Disease: 0.6144
• Accuracy Background: 0.9962
• Iou Unlabelled: nan
• Iou Healthy: 0.8897
• Iou Disease: 0.5318
• Iou Background: 0.9851

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Unlabelled	Accuracy Healthy	Accuracy Disease	Accuracy Background	Iou Unlabelled	Iou Healthy	Iou Disease	Iou Background
1.0053	0.33	20	1.2067	0.3883	0.6563	0.8803	nan	0.9886	0.0986	0.8816	0.0	0.6303	0.0418	0.8810
0.8256	0.67	40	0.8465	0.4445	0.7343	0.9192	nan	0.9352	0.3360	0.9317	0.0	0.7090	0.1390	0.9300
0.4785	1.0	60	0.5391	0.4488	0.6814	0.9466	nan	0.9824	0.0991	0.9626	0.0	0.7520	0.0852	0.9579
0.5875	1.33	80	0.6045	0.6251	0.7087	0.9510	nan	0.9790	0.1807	0.9664	nan	0.7718	0.1407	0.9627
0.5354	1.67	100	0.3814	0.6633	0.7242	0.9596	nan	0.9313	0.2584	0.9831	nan	0.7937	0.2267	0.9694
0.5146	2.0	120	0.3455	0.6755	0.7307	0.9650	nan	0.9507	0.2552	0.9862	nan	0.8187	0.2321	0.9756
0.5212	2.33	140	0.4321	0.6471	0.7169	0.9598	nan	0.9798	0.1946	0.9764	nan	0.7934	0.1752	0.9727
0.4044	2.67	160	0.1609	0.6607	0.7065	0.9607	nan	0.9067	0.2230	0.9898	nan	0.7994	0.2145	0.9681
0.4572	3.0	180	0.2172	0.6882	0.7268	0.9657	nan	0.9050	0.2808	0.9946	nan	0.8254	0.2667	0.9724
0.2676	3.33	200	0.2950	0.6926	0.7451	0.9690	nan	0.9674	0.2806	0.9874	nan	0.8332	0.2641	0.9804
0.2991	3.67	220	0.2523	0.6976	0.7411	0.9695	nan	0.9337	0.2959	0.9937	nan	0.8401	0.2734	0.9792
0.5168	4.0	240	0.1013	0.6599	0.6969	0.9651	nan	0.9065	0.1880	0.9961	nan	0.8226	0.1835	0.9735
0.3117	4.33	260	0.2323	0.6951	0.7424	0.9696	nan	0.9496	0.2863	0.9912	nan	0.8409	0.2644	0.9800
0.2888	4.67	280	0.1264	0.7466	0.7901	0.9736	nan	0.9277	0.4470	0.9957	nan	0.8582	0.3998	0.9817
0.1684	5.0	300	0.1291	0.7646	0.8119	0.9758	nan	0.9373	0.5033	0.9951	nan	0.8693	0.4406	0.9837
0.2041	5.33	320	0.1804	0.7720	0.8213	0.9773	nan	0.9462	0.5229	0.9948	nan	0.8732	0.4564	0.9863
0.179	5.67	340	0.1381	0.7937	0.8560	0.9773	nan	0.9280	0.6454	0.9948	nan	0.8727	0.5235	0.9849
0.1444	6.0	360	0.1671	0.7393	0.7972	0.9727	nan	0.9750	0.4301	0.9866	nan	0.8485	0.3862	0.9832
0.2365	6.33	380	0.1272	0.7813	0.8275	0.9771	nan	0.9354	0.5511	0.9958	nan	0.8722	0.4870	0.9848
0.2216	6.67	400	0.0907	0.7923	0.8358	0.9775	nan	0.9224	0.5875	0.9976	nan	0.8761	0.5171	0.9837
0.1437	7.0	420	0.0782	0.7715	0.8148	0.9732	nan	0.9167	0.5329	0.9948	nan	0.8591	0.4777	0.9778
0.1065	7.33	440	0.0877	0.7537	0.7917	0.9725	nan	0.9187	0.4610	0.9955	nan	0.8549	0.4281	0.9781
0.2535	7.67	460	0.0784	0.7457	0.7810	0.9723	nan	0.9078	0.4374	0.9979	nan	0.8593	0.4003	0.9776
0.108	8.0	480	0.1003	0.7544	0.7984	0.9759	nan	0.9563	0.4455	0.9934	nan	0.8703	0.4084	0.9844
0.0884	8.33	500	0.0744	0.7689	0.8108	0.9753	nan	0.9335	0.5039	0.9952	nan	0.8694	0.4557	0.9816
0.1935	8.67	520	0.1047	0.7954	0.8373	0.9800	nan	0.9482	0.5672	0.9965	nan	0.8878	0.5111	0.9874
0.268	9.0	540	0.1281	0.8058	0.8559	0.9816	nan	0.9563	0.6158	0.9957	nan	0.8945	0.5328	0.9900
0.0779	9.33	560	0.1393	0.7680	0.8126	0.9774	nan	0.9645	0.4806	0.9928	nan	0.8763	0.4418	0.9859
0.1295	9.67	580	0.0749	0.7433	0.7861	0.9741	nan	0.9572	0.4092	0.9920	nan	0.8595	0.3877	0.9826
0.6322	10.0	600	0.0792	0.7825	0.8233	0.9776	nan	0.9419	0.5322	0.9957	nan	0.8757	0.4869	0.9849
0.1491	10.33	620	0.0685	0.7805	0.8265	0.9760	nan	0.9320	0.5525	0.9950	nan	0.8753	0.4845	0.9816
0.0876	10.67	640	0.1347	0.7672	0.8134	0.9783	nan	0.9713	0.4762	0.9928	nan	0.8825	0.4317	0.9873
0.3076	11.0	660	0.0989	0.7737	0.8120	0.9745	nan	0.9152	0.5239	0.9970	nan	0.8700	0.4723	0.9788
0.08	11.33	680	0.0923	0.7914	0.8350	0.9766	nan	0.9212	0.5869	0.9968	nan	0.8780	0.5150	0.9813
0.0816	11.67	700	0.0878	0.7948	0.8392	0.9790	nan	0.9446	0.5773	0.9957	nan	0.8891	0.5101	0.9851
0.057	12.0	720	0.0824	0.8032	0.8537	0.9792	nan	0.9496	0.6175	0.9940	nan	0.8874	0.5368	0.9855
0.0881	12.33	740	0.0802	0.7694	0.8051	0.9766	nan	0.9349	0.4833	0.9970	nan	0.8792	0.4466	0.9825
0.1084	12.67	760	0.1033	0.8113	0.8662	0.9793	nan	0.9392	0.6646	0.9947	nan	0.8858	0.5626	0.9854
0.0703	13.0	780	0.0889	0.7795	0.8234	0.9765	nan	0.9385	0.5367	0.9949	nan	0.8775	0.4785	0.9824
0.1332	13.33	800	0.0803	0.7859	0.8332	0.9779	nan	0.9494	0.5562	0.9940	nan	0.8857	0.4881	0.9840
0.0872	13.67	820	0.1034	0.7741	0.8173	0.9788	nan	0.9666	0.4913	0.9939	nan	0.8860	0.4492	0.9871
0.0475	14.0	840	0.0728	0.7826	0.8295	0.9771	nan	0.9400	0.5536	0.9948	nan	0.8820	0.4829	0.9829
0.0569	14.33	860	0.0940	0.7794	0.8236	0.9786	nan	0.9671	0.5108	0.9930	nan	0.8824	0.4690	0.9867
0.7	14.67	880	0.0753	0.8024	0.8459	0.9797	nan	0.9443	0.5974	0.9961	nan	0.8905	0.5309	0.9858
0.0805	15.0	900	0.0738	0.8145	0.8720	0.9810	nan	0.9542	0.6679	0.9940	nan	0.8929	0.5622	0.9884
0.601	15.33	920	0.0970	0.8117	0.8614	0.9822	nan	0.9685	0.6216	0.9941	nan	0.8990	0.5461	0.9900
0.0844	15.67	940	0.0732	0.7691	0.8128	0.9755	nan	0.9479	0.4975	0.9930	nan	0.8731	0.4529	0.9814
0.1097	16.0	960	0.0622	0.8170	0.8721	0.9798	nan	0.9321	0.6881	0.9960	nan	0.8893	0.5761	0.9855
0.1446	16.33	980	0.0675	0.7983	0.8491	0.9797	nan	0.9621	0.5921	0.9930	nan	0.8900	0.5180	0.9868
0.3657	16.67	1000	0.0696	0.7900	0.8335	0.9777	nan	0.9371	0.5676	0.9957	nan	0.8836	0.5032	0.9831
0.2767	17.0	1020	0.0883	0.7599	0.8009	0.9715	nan	0.9063	0.5008	0.9955	nan	0.8580	0.4468	0.9750
0.1155	17.33	1040	0.0720	0.7929	0.8337	0.9797	nan	0.9527	0.5526	0.9958	nan	0.8882	0.5034	0.9870
0.0765	17.67	1060	0.0733	0.7843	0.8297	0.9767	nan	0.9395	0.5553	0.9944	nan	0.8798	0.4911	0.9821
0.1484	18.0	1080	0.0833	0.7904	0.8384	0.9788	nan	0.9595	0.5626	0.9932	nan	0.8870	0.4985	0.9856
0.1017	18.33	1100	0.0928	0.7866	0.8304	0.9799	nan	0.9707	0.5269	0.9935	nan	0.8888	0.4829	0.9881
0.0745	18.67	1120	0.0662	0.8084	0.8545	0.9806	nan	0.9472	0.6203	0.9961	nan	0.8947	0.5434	0.9870
0.093	19.0	1140	0.0673	0.8026	0.8510	0.9799	nan	0.9517	0.6065	0.9947	nan	0.8926	0.5290	0.9863
0.0475	19.33	1160	0.0818	0.7982	0.8447	0.9805	nan	0.9632	0.5768	0.9942	nan	0.8939	0.5130	0.9878
0.051	19.67	1180	0.0690	0.7931	0.8368	0.9800	nan	0.9578	0.5573	0.9951	nan	0.8924	0.4996	0.9873
0.0432	20.0	1200	0.0654	0.8022	0.8495	0.9793	nan	0.9379	0.6144	0.9962	nan	0.8897	0.5318	0.9851

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.0+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0