segformer-b0-finetuned-cityscapes-1024-1024_corm

This model is a fine-tuned version of nvidia/segformer-b0-finetuned-cityscapes-1024-1024 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0506
• Mean Iou: 0.9091
• Mean Accuracy: 0.9514
• Overall Accuracy: 0.9835
• Accuracy Background: 0.9977
• Accuracy Corm: 0.9256
• Accuracy Damage: 0.9308
• Iou Background: 0.9938
• Iou Corm: 0.8393
• Iou Damage: 0.8942

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Corm	Accuracy Damage	Iou Background	Iou Corm	Iou Damage
0.8589	0.6061	20	0.8411	0.4729	0.5967	0.8533	0.9785	0.5630	0.2486	0.9291	0.2623	0.2273
0.7143	1.2121	40	0.6737	0.6855	0.8154	0.9261	0.9743	0.7133	0.7586	0.9723	0.4438	0.6403
0.5785	1.8182	60	0.5154	0.7300	0.8494	0.9393	0.9707	0.6523	0.9253	0.9704	0.4854	0.7342
0.4449	2.4242	80	0.3898	0.8138	0.9106	0.9596	0.9778	0.8207	0.9333	0.9771	0.6594	0.8048
0.3313	3.0303	100	0.3061	0.8287	0.9246	0.9633	0.9820	0.9175	0.8743	0.9809	0.7007	0.8046
0.2799	3.6364	120	0.2367	0.8545	0.9230	0.9717	0.9907	0.8453	0.9331	0.9882	0.7423	0.8330
0.2482	4.2424	140	0.2088	0.8669	0.9301	0.9746	0.9927	0.8707	0.9269	0.9899	0.7653	0.8455
0.2134	4.8485	160	0.1841	0.8637	0.9343	0.9739	0.9939	0.9403	0.8687	0.9908	0.7666	0.8337
0.1695	5.4545	180	0.1585	0.8832	0.9405	0.9780	0.9943	0.9053	0.9221	0.9916	0.7954	0.8625
0.1581	6.0606	200	0.1414	0.8895	0.9410	0.9795	0.9959	0.8998	0.9273	0.9922	0.8051	0.8711
0.1413	6.6667	220	0.1253	0.8943	0.9430	0.9805	0.9966	0.9060	0.9263	0.9926	0.8141	0.8763
0.1125	7.2727	240	0.1138	0.8955	0.9453	0.9807	0.9965	0.9193	0.9203	0.9927	0.8169	0.8770
0.1195	7.8788	260	0.1124	0.8811	0.9411	0.9779	0.9967	0.9503	0.8763	0.9926	0.7970	0.8537
0.1032	8.4848	280	0.1049	0.8912	0.9457	0.9798	0.9964	0.9413	0.8994	0.9927	0.8117	0.8692
0.104	9.0909	300	0.0912	0.9013	0.9459	0.9819	0.9971	0.9070	0.9337	0.9929	0.8255	0.8856
0.0994	9.6970	320	0.0887	0.9022	0.9473	0.9820	0.9972	0.9172	0.9275	0.9930	0.8275	0.8860
0.0914	10.3030	340	0.0866	0.8999	0.9471	0.9815	0.9974	0.9278	0.9160	0.9930	0.8249	0.8819
0.0917	10.9091	360	0.0813	0.9032	0.9473	0.9822	0.9975	0.9166	0.9279	0.9930	0.8293	0.8873
0.0822	11.5152	380	0.0774	0.9038	0.9454	0.9825	0.9972	0.8900	0.9490	0.9932	0.8280	0.8903
0.078	12.1212	400	0.0766	0.9035	0.9488	0.9823	0.9973	0.9244	0.9247	0.9932	0.8301	0.8871
0.0782	12.7273	420	0.0739	0.9027	0.9490	0.9822	0.9974	0.9302	0.9195	0.9933	0.8293	0.8856
0.0759	13.3333	440	0.0715	0.9025	0.9487	0.9821	0.9975	0.9316	0.9170	0.9933	0.8292	0.8851
0.066	13.9394	460	0.0691	0.9059	0.9480	0.9828	0.9975	0.9089	0.9377	0.9934	0.8328	0.8915
0.0774	14.5455	480	0.0674	0.9059	0.9493	0.9827	0.9976	0.9237	0.9267	0.9933	0.8339	0.8904
0.0719	15.1515	500	0.0690	0.9034	0.9488	0.9823	0.9977	0.9305	0.9183	0.9933	0.8307	0.8862
0.0713	15.7576	520	0.0666	0.9003	0.9486	0.9817	0.9974	0.9368	0.9117	0.9935	0.8262	0.8814
0.0647	16.3636	540	0.0645	0.9033	0.9498	0.9823	0.9974	0.9346	0.9173	0.9934	0.8307	0.8858
0.0576	16.9697	560	0.0637	0.9046	0.9499	0.9826	0.9975	0.9301	0.9221	0.9936	0.8323	0.8879
0.0598	17.5758	580	0.0625	0.9044	0.9501	0.9825	0.9974	0.9333	0.9197	0.9935	0.8321	0.8875
0.0676	18.1818	600	0.0644	0.8991	0.9489	0.9815	0.9974	0.9447	0.9047	0.9936	0.8243	0.8794
0.0474	18.7879	620	0.0624	0.9018	0.9503	0.9820	0.9973	0.9427	0.9108	0.9936	0.8285	0.8832
0.0611	19.3939	640	0.0606	0.9064	0.9504	0.9829	0.9976	0.9282	0.9254	0.9936	0.8350	0.8905
0.058	20.0	660	0.0596	0.9048	0.9508	0.9826	0.9973	0.9355	0.9197	0.9936	0.8330	0.8877
0.0574	20.6061	680	0.0575	0.9082	0.9484	0.9834	0.9973	0.8972	0.9507	0.9938	0.8356	0.8951
0.0562	21.2121	700	0.0576	0.9065	0.9465	0.9831	0.9973	0.8870	0.9553	0.9937	0.8319	0.8939
0.0551	21.8182	720	0.0571	0.9067	0.9515	0.9830	0.9972	0.9299	0.9274	0.9938	0.8361	0.8903
0.0498	22.4242	740	0.0564	0.9090	0.9509	0.9834	0.9976	0.9217	0.9336	0.9936	0.8388	0.8945
0.0566	23.0303	760	0.0554	0.9067	0.9511	0.9830	0.9975	0.9307	0.9251	0.9938	0.8359	0.8904
0.0436	23.6364	780	0.0567	0.9056	0.9509	0.9827	0.9976	0.9362	0.9191	0.9936	0.8344	0.8889
0.0586	24.2424	800	0.0548	0.9081	0.9515	0.9832	0.9974	0.9273	0.9296	0.9938	0.8380	0.8924
0.0497	24.8485	820	0.0549	0.9091	0.9511	0.9834	0.9976	0.9227	0.9331	0.9937	0.8390	0.8946
0.0535	25.4545	840	0.0544	0.9073	0.9510	0.9831	0.9976	0.9296	0.9256	0.9937	0.8368	0.8913
0.0514	26.0606	860	0.0539	0.9096	0.9514	0.9836	0.9975	0.9205	0.9362	0.9938	0.8399	0.8953
0.0684	26.6667	880	0.0550	0.9055	0.9511	0.9827	0.9976	0.9383	0.9174	0.9937	0.8344	0.8884
0.0542	27.2727	900	0.0524	0.9100	0.9512	0.9836	0.9976	0.9196	0.9364	0.9938	0.8403	0.8959
0.0455	27.8788	920	0.0534	0.9083	0.9518	0.9833	0.9975	0.9296	0.9282	0.9938	0.8384	0.8929
0.0512	28.4848	940	0.0525	0.9095	0.9504	0.9836	0.9977	0.9149	0.9386	0.9938	0.8393	0.8954
0.0486	29.0909	960	0.0524	0.9083	0.9516	0.9833	0.9976	0.9292	0.9280	0.9938	0.8383	0.8927
0.0486	29.6970	980	0.0517	0.9099	0.9509	0.9836	0.9976	0.9172	0.9380	0.9938	0.8401	0.8958
0.0388	30.3030	1000	0.0514	0.9101	0.9510	0.9837	0.9975	0.9147	0.9408	0.9938	0.8402	0.8962
0.0571	30.9091	1020	0.0518	0.9091	0.9513	0.9834	0.9977	0.9247	0.9314	0.9938	0.8392	0.8942
0.0515	31.5152	1040	0.0511	0.9103	0.9510	0.9837	0.9975	0.9135	0.9418	0.9939	0.8404	0.8966
0.049	32.1212	1060	0.0517	0.9100	0.9510	0.9836	0.9976	0.9171	0.9385	0.9938	0.8402	0.8958
0.0533	32.7273	1080	0.0513	0.9095	0.9514	0.9835	0.9975	0.9221	0.9346	0.9938	0.8398	0.8949
0.0443	33.3333	1100	0.0513	0.9092	0.9513	0.9835	0.9977	0.9245	0.9317	0.9938	0.8395	0.8944
0.0573	33.9394	1120	0.0516	0.9089	0.9515	0.9834	0.9976	0.9270	0.9297	0.9938	0.8390	0.8938
0.0421	34.5455	1140	0.0516	0.9082	0.9512	0.9833	0.9977	0.9294	0.9264	0.9938	0.8382	0.8927
0.0509	35.1515	1160	0.0503	0.9102	0.9508	0.9837	0.9976	0.9145	0.9403	0.9938	0.8403	0.8966
0.0854	35.7576	1180	0.0511	0.9087	0.9518	0.9834	0.9975	0.9285	0.9293	0.9938	0.8388	0.8934
0.0522	36.3636	1200	0.0508	0.9089	0.9516	0.9834	0.9976	0.9269	0.9302	0.9938	0.8392	0.8938
0.0648	36.9697	1220	0.0503	0.9103	0.9514	0.9837	0.9975	0.9175	0.9391	0.9939	0.8408	0.8964
0.0513	37.5758	1240	0.0502	0.9099	0.9511	0.9836	0.9977	0.9203	0.9353	0.9938	0.8402	0.8957
0.0494	38.1818	1260	0.0512	0.9093	0.9516	0.9835	0.9976	0.9257	0.9316	0.9938	0.8396	0.8944
0.0513	38.7879	1280	0.0510	0.9096	0.9517	0.9836	0.9975	0.9232	0.9343	0.9939	0.8400	0.8949
0.0573	39.3939	1300	0.0508	0.9092	0.9514	0.9835	0.9976	0.9249	0.9318	0.9938	0.8395	0.8943
0.0627	40.0	1320	0.0506	0.9091	0.9514	0.9835	0.9977	0.9256	0.9308	0.9938	0.8393	0.8942

Framework versions

• Transformers 4.44.1
• Pytorch 2.6.0+cpu
• Datasets 2.21.0
• Tokenizers 0.19.1