This model is a fine-tuned version of nvidia/segformer-b4-finetuned-cityscapes-1024-1024 on the peldrak/coastTrain dataset. It achieves the following results on the evaluation set:

Loss: 0.2345
Mean Iou: 0.7920
Mean Accuracy: 0.8609
Overall Accuracy: 0.9360
Accuracy Water: 0.9603
Accuracy Whitewater: 0.6503
Accuracy Sediment: 0.8872
Accuracy Other Natural Terrain: 0.6902
Accuracy Vegetation: 0.9383
Accuracy Development: 0.9340
Accuracy Unknown: 0.9659
Iou Water: 0.9194
Iou Whitewater: 0.5196
Iou Sediment: 0.8231
Iou Other Natural Terrain: 0.6344
Iou Vegetation: 0.8728
Iou Development: 0.8571
Iou Unknown: 0.9179
F1 Score: 0.9354

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: 4
eval_batch_size: 4
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 Water	Accuracy Whitewater	Accuracy Sediment	Accuracy Other Natural Terrain	Accuracy Vegetation	Accuracy Development	Accuracy Unknown	Iou Water	Iou Whitewater	Iou Sediment	Iou Other Natural Terrain	Iou Vegetation	Iou Development	Iou Unknown	F1 Score
1.6229	0.16	20	1.5080	0.3461	0.4464	0.6629	0.7064	0.0117	0.3677	0.0094	0.9033	0.4311	0.6952	0.6022	0.0115	0.2505	0.0054	0.5058	0.3657	0.6818	0.6578
1.6972	0.31	40	1.1157	0.4168	0.5175	0.7412	0.7367	0.0	0.5983	0.0000	0.9738	0.5606	0.7528	0.6767	0.0	0.4343	0.0000	0.5731	0.4817	0.7519	0.7355
1.1969	0.47	60	0.8521	0.4777	0.5606	0.8157	0.8778	0.0001	0.5850	0.0000	0.9674	0.5962	0.8978	0.7865	0.0001	0.4970	0.0000	0.6699	0.4942	0.8960	0.8025
1.2209	0.62	80	0.6990	0.5024	0.6009	0.8277	0.8306	0.0	0.7406	0.0	0.9174	0.7970	0.9208	0.7942	0.0	0.5641	0.0	0.6800	0.5612	0.9175	0.8201
0.793	0.78	100	0.5543	0.5256	0.6016	0.8590	0.9409	0.0	0.6744	0.0	0.9365	0.7280	0.9312	0.8480	0.0	0.5903	0.0	0.7405	0.5737	0.9264	0.8458
1.1597	0.93	120	0.4957	0.5524	0.6231	0.8720	0.9201	0.0	0.7323	0.0	0.9560	0.8134	0.9401	0.8588	0.0	0.6428	0.0	0.7344	0.6976	0.9335	0.8602
0.689	1.09	140	0.4376	0.5692	0.6374	0.8844	0.9221	0.0	0.7930	0.0000	0.9586	0.8469	0.9411	0.8748	0.0	0.6899	0.0000	0.7542	0.7310	0.9348	0.8730
0.9423	1.24	160	0.3968	0.5715	0.6433	0.8882	0.9328	0.0000	0.8273	0.0	0.9201	0.8664	0.9564	0.8794	0.0000	0.6963	0.0	0.7687	0.7135	0.9424	0.8766
1.2176	1.4	180	0.3838	0.5673	0.6397	0.8811	0.9107	0.0000	0.7816	0.0	0.9461	0.8832	0.9561	0.8578	0.0000	0.6427	0.0	0.7728	0.7550	0.9429	0.8695
0.4714	1.55	200	0.4459	0.5380	0.6092	0.8570	0.9730	0.0	0.5500	0.0	0.8806	0.9121	0.9490	0.7998	0.0	0.4883	0.0	0.7722	0.7618	0.9442	0.8398
0.5087	1.71	220	0.4062	0.5677	0.6359	0.8827	0.9365	0.0008	0.7216	0.0	0.9511	0.8918	0.9499	0.8844	0.0008	0.6722	0.0	0.7302	0.7463	0.9402	0.8709
0.484	1.86	240	0.3121	0.5926	0.6518	0.9017	0.9688	0.0001	0.7858	0.0	0.9341	0.9243	0.9498	0.8788	0.0001	0.7236	0.0	0.8069	0.7972	0.9420	0.8886
0.443	2.02	260	0.3554	0.5811	0.6575	0.8904	0.8939	0.0001	0.8842	0.0	0.9316	0.9332	0.9599	0.8541	0.0001	0.6697	0.0	0.8205	0.7724	0.9512	0.8798
0.466	2.17	280	0.3265	0.5830	0.6553	0.8954	0.9347	0.0021	0.8131	0.0	0.9256	0.9487	0.9627	0.8786	0.0021	0.6993	0.0	0.7950	0.7566	0.9494	0.8833
0.7117	2.33	300	0.4096	0.5634	0.6494	0.8672	0.8367	0.0076	0.9133	0.0	0.9048	0.9233	0.9600	0.8000	0.0076	0.5985	0.0	0.8080	0.7798	0.9503	0.8595
0.3095	2.48	320	0.3111	0.5858	0.6494	0.8986	0.9655	0.0070	0.7517	0.0	0.9363	0.9210	0.9644	0.8850	0.0070	0.6808	0.0	0.8089	0.7790	0.9399	0.8851
0.3843	2.64	340	0.3076	0.6033	0.6589	0.9050	0.9648	0.0448	0.7872	0.0	0.9478	0.9052	0.9628	0.8786	0.0443	0.6973	0.0	0.8322	0.8187	0.9516	0.8924
0.4158	2.79	360	0.2985	0.5965	0.6553	0.9026	0.9557	0.0266	0.8130	0.0	0.9561	0.8939	0.9415	0.8820	0.0265	0.7245	0.0	0.8147	0.7892	0.9386	0.8903
0.3492	2.95	380	0.2709	0.6251	0.6863	0.9126	0.9524	0.1545	0.8818	0.0	0.9347	0.9218	0.9587	0.9003	0.1514	0.7515	0.0	0.8313	0.7895	0.9520	0.9026
0.2384	3.1	400	0.2531	0.6348	0.6950	0.9169	0.9541	0.1876	0.8731	0.0015	0.9446	0.9407	0.9632	0.9061	0.1830	0.7663	0.0015	0.8395	0.7944	0.9524	0.9071
0.2227	3.26	420	0.2772	0.6388	0.6939	0.9186	0.9648	0.1767	0.9028	0.0013	0.9230	0.9247	0.9637	0.9015	0.1735	0.7597	0.0013	0.8503	0.8312	0.9544	0.9087
0.6677	3.41	440	0.2861	0.6306	0.6874	0.9091	0.9640	0.1876	0.7780	0.0559	0.9635	0.9022	0.9610	0.8965	0.1838	0.7228	0.0558	0.8215	0.7825	0.9509	0.8996
0.3552	3.57	460	0.2795	0.6408	0.6945	0.9130	0.9641	0.2606	0.8234	0.0001	0.9487	0.8958	0.9690	0.8932	0.2536	0.7201	0.0001	0.8427	0.8204	0.9553	0.9032
0.3258	3.72	480	0.3075	0.6306	0.6891	0.9042	0.9730	0.2962	0.7091	0.0068	0.9628	0.9131	0.9625	0.8945	0.2617	0.6816	0.0068	0.8003	0.8150	0.9546	0.8939
0.4778	3.88	500	0.2449	0.6570	0.7137	0.9188	0.9583	0.2928	0.8807	0.0347	0.9383	0.9280	0.9633	0.9028	0.2768	0.7677	0.0347	0.8407	0.8223	0.9541	0.9106
0.4817	4.03	520	0.2365	0.6790	0.7381	0.9235	0.9611	0.4327	0.8871	0.0484	0.9393	0.9328	0.9651	0.9121	0.3866	0.7842	0.0483	0.8459	0.8230	0.9529	0.9165
0.3363	4.19	540	0.2273	0.6783	0.7315	0.9243	0.9635	0.4529	0.8805	0.0058	0.9546	0.8945	0.9685	0.9131	0.4101	0.7847	0.0058	0.8451	0.8339	0.9554	0.9163
0.4825	4.34	560	0.2615	0.6791	0.7482	0.9180	0.9406	0.5008	0.9124	0.0441	0.9206	0.9512	0.9675	0.8996	0.4385	0.7400	0.0441	0.8552	0.8198	0.9562	0.9119
0.3482	4.5	580	0.2336	0.6965	0.7537	0.9276	0.9695	0.4845	0.8611	0.1015	0.9492	0.9455	0.9648	0.9193	0.4258	0.7901	0.1006	0.8498	0.8349	0.9550	0.9215
0.5311	4.65	600	0.2592	0.6858	0.7484	0.9200	0.9477	0.4867	0.8797	0.0974	0.9463	0.9109	0.9703	0.9136	0.4329	0.7687	0.0970	0.8438	0.8232	0.9215	0.9142
0.3754	4.81	620	0.2345	0.7039	0.7629	0.9265	0.9641	0.5201	0.8692	0.1376	0.9387	0.9338	0.9769	0.9233	0.4557	0.7998	0.1369	0.8395	0.8404	0.9315	0.9211
0.236	4.96	640	0.2342	0.7061	0.7604	0.9268	0.9669	0.4754	0.8856	0.1790	0.9329	0.9039	0.9790	0.9218	0.4330	0.8088	0.1763	0.8374	0.8329	0.9327	0.9218
0.3496	5.12	660	0.2061	0.7264	0.7870	0.9313	0.9622	0.5779	0.9263	0.2075	0.9189	0.9404	0.9757	0.9243	0.5092	0.8247	0.2026	0.8525	0.8402	0.9315	0.9273
0.1729	5.27	680	0.2289	0.7086	0.7793	0.9245	0.9541	0.6032	0.8708	0.1744	0.9351	0.9392	0.9781	0.9202	0.4789	0.7981	0.1706	0.8377	0.8265	0.9282	0.9203
0.2636	5.43	700	0.2071	0.7739	0.8389	0.9335	0.9623	0.6448	0.8970	0.5348	0.9262	0.9396	0.9676	0.9245	0.5537	0.8125	0.4941	0.8551	0.8473	0.9304	0.9324
0.1594	5.58	720	0.2175	0.7447	0.8114	0.9284	0.9632	0.6275	0.8570	0.3850	0.9306	0.9383	0.9783	0.9172	0.5130	0.7999	0.3765	0.8467	0.8207	0.9389	0.9262
0.6799	5.74	740	0.1965	0.7704	0.8330	0.9379	0.9650	0.6576	0.9113	0.4469	0.9289	0.9430	0.9783	0.9303	0.5574	0.8353	0.4280	0.8695	0.8384	0.9338	0.9364
0.4955	5.89	760	0.2184	0.7480	0.8065	0.9318	0.9569	0.5099	0.9085	0.4344	0.9266	0.9255	0.9841	0.9146	0.4472	0.8109	0.4160	0.8654	0.8514	0.9308	0.9297
0.218	6.05	780	0.2025	0.7870	0.8508	0.9364	0.9678	0.5859	0.8690	0.6870	0.9368	0.9360	0.9732	0.9282	0.4918	0.8198	0.6437	0.8651	0.8324	0.9282	0.9355
0.4344	6.2	800	0.2128	0.7816	0.8399	0.9361	0.9620	0.6012	0.8908	0.5972	0.9446	0.9092	0.9743	0.9243	0.5193	0.8213	0.5626	0.8684	0.8520	0.9233	0.9350
0.5841	6.36	820	0.2412	0.7965	0.8614	0.9378	0.9587	0.6847	0.8888	0.6398	0.9360	0.9408	0.9808	0.9254	0.5838	0.8259	0.6002	0.8758	0.8440	0.9204	0.9371
0.3048	6.51	840	0.2336	0.7869	0.8580	0.9344	0.9576	0.6990	0.8928	0.6175	0.9281	0.9381	0.9728	0.9160	0.5612	0.8190	0.5724	0.8720	0.8484	0.9193	0.9337
0.2002	6.67	860	0.2373	0.7929	0.8605	0.9343	0.9512	0.6492	0.8691	0.6968	0.9424	0.9290	0.9861	0.9216	0.5520	0.8087	0.6401	0.8685	0.8437	0.9155	0.9337
0.2093	6.82	880	0.2335	0.7918	0.8528	0.9364	0.9649	0.6226	0.8682	0.6653	0.9414	0.9311	0.9758	0.9210	0.5272	0.8166	0.6269	0.8720	0.8550	0.9235	0.9355
0.1581	6.98	900	0.2279	0.7995	0.8786	0.9368	0.9509	0.7124	0.8959	0.7476	0.9347	0.9273	0.9812	0.9159	0.5266	0.8244	0.6672	0.8800	0.8627	0.9198	0.9367
0.1209	7.13	920	0.2432	0.7832	0.8475	0.9327	0.9548	0.5902	0.8897	0.6726	0.9357	0.9173	0.9725	0.9104	0.5047	0.8096	0.6187	0.8725	0.8463	0.9201	0.9319
0.1492	7.29	940	0.2345	0.7920	0.8609	0.9360	0.9603	0.6503	0.8872	0.6902	0.9383	0.9340	0.9659	0.9194	0.5196	0.8231	0.6344	0.8728	0.8571	0.9179	0.9354

Framework versions

Transformers 4.37.0
Pytorch 2.1.2
Datasets 2.18.0
Tokenizers 0.15.1

peldrak/segformer-b4-cityscapes-finetuned-coastTrain

segformer-b4-cityscapes-finetuned-coastTrain

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

peldrak/
segformer-b4-cityscapes-finetuned-coastTrain