segformer-finetuned-coastal

peldrak/segformer-finetuned-coastal

segformer-finetuned-coastal is a image segmentation model from peldrak. Use it for the image segmentation task on the model card, and read the license before you ship it in a product. It is set up for transformers. The card lists the license as other.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

1% of all-time downloads

All-time downloads

621

Public

Repo size

29.8 MB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin14.9 MB · 100%

At a glance

Task: Image Segmentation
Library: transformers
License: other
Model type: segformer
Access: Public
Created: Nov 5, 2023
Updated: Nov 7, 2023
SHA: d4ba311e

Base models

nvidia/segformer-b0-finetuned-ade-512-512

Model card

From the Hugging Face model README

segformer-finetuned-coastal

This model is a fine-tuned version of nvidia/segformer-b0-finetuned-ade-512-512 on the peldrak/coastal3 dataset. It achieves the following results on the evaluation set:

Loss: 1.7456
Mean Iou: 0.0841
Mean Accuracy: 0.1629
Overall Accuracy: 0.3309
Accuracy Water: 0.2728
Accuracy Whitewater: 0.0
Accuracy Sediment: 0.0194
Accuracy Other Natural Terrain: 0.0
Accuracy Vegetation: 0.7325
Accuracy Development: 0.0055
Accuracy Unknown: 0.1102
Iou Water: 0.1767
Iou Whitewater: 0.0
Iou Sediment: 0.0157
Iou Other Natural Terrain: 0.0
Iou Vegetation: 0.3002
Iou Development: 0.0052
Iou Unknown: 0.0908

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
num_epochs: 5

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
1.8429	0.12	20	1.8742	0.0623	0.1432	0.2198	0.5172	0.0044	0.3137	0.0050	0.1165	0.0019	0.0435	0.2514	0.0037	0.0521	0.0034	0.0889	0.0019	0.0348
1.6623	0.25	40	1.8297	0.0731	0.1572	0.2536	0.4349	0.0007	0.3336	0.0010	0.3158	0.0011	0.0135	0.2462	0.0006	0.0569	0.0009	0.1940	0.0011	0.0120
1.6173	0.37	60	1.7856	0.0812	0.1585	0.3080	0.3151	0.0007	0.1179	0.0005	0.6339	0.0012	0.0405	0.2056	0.0007	0.0357	0.0005	0.2859	0.0012	0.0388
1.5078	0.49	80	1.7805	0.0764	0.1537	0.3038	0.2792	0.0	0.1036	0.0000	0.6639	0.0003	0.0290	0.1839	0.0	0.0338	0.0000	0.2895	0.0003	0.0274
1.3927	0.62	100	1.7549	0.0725	0.1521	0.3227	0.2589	0.0	0.0222	0.0001	0.7643	0.0005	0.0187	0.1703	0.0	0.0133	0.0001	0.3055	0.0005	0.0179
1.6024	0.74	120	1.7348	0.0835	0.1598	0.3271	0.3514	0.0	0.0204	0.0001	0.6538	0.0005	0.0926	0.2021	0.0	0.0147	0.0001	0.2873	0.0005	0.0796
1.3287	0.86	140	1.7352	0.0776	0.1556	0.3267	0.3309	0.0	0.0083	0.0	0.6910	0.0000	0.0591	0.1935	0.0	0.0076	0.0	0.2849	0.0000	0.0573
1.4153	0.99	160	1.7024	0.0746	0.1567	0.3356	0.2947	0.0	0.0011	0.0	0.7683	0.0006	0.0320	0.1879	0.0	0.0011	0.0	0.3021	0.0006	0.0309
1.3334	1.11	180	1.7262	0.0744	0.1567	0.3374	0.3130	0.0	0.0018	0.0	0.7620	0.0001	0.0203	0.1988	0.0	0.0017	0.0	0.3004	0.0001	0.0198
1.3956	1.23	200	1.7304	0.0858	0.1622	0.3326	0.4838	0.0	0.0127	0.0	0.5432	0.0015	0.0944	0.2373	0.0	0.0112	0.0	0.2731	0.0015	0.0777
1.5776	1.36	220	1.7300	0.0791	0.1622	0.3411	0.2581	0.0	0.0010	0.0	0.8012	0.0003	0.0748	0.1734	0.0	0.0010	0.0	0.3144	0.0003	0.0647
1.1656	1.48	240	1.7248	0.0831	0.1657	0.3440	0.2687	0.0	0.0026	0.0	0.7879	0.0014	0.0995	0.1775	0.0	0.0025	0.0	0.3183	0.0014	0.0822
1.4429	1.6	260	1.7308	0.0764	0.1616	0.3408	0.2091	0.0	0.0029	0.0	0.8518	0.0037	0.0637	0.1507	0.0	0.0028	0.0	0.3200	0.0036	0.0578
1.6649	1.73	280	1.7282	0.0743	0.1564	0.3372	0.3261	0.0	0.0033	0.0	0.7514	0.0015	0.0128	0.1994	0.0	0.0031	0.0	0.3037	0.0015	0.0125
1.3634	1.85	300	1.7216	0.0847	0.1653	0.3413	0.3196	0.0	0.0273	0.0	0.7402	0.0036	0.0665	0.2012	0.0	0.0204	0.0	0.3101	0.0035	0.0579
1.5224	1.98	320	1.7343	0.0822	0.1626	0.3410	0.3793	0.0	0.0277	0.0	0.6985	0.0017	0.0311	0.2236	0.0	0.0210	0.0	0.3000	0.0016	0.0293
1.2527	2.1	340	1.7149	0.0759	0.1559	0.3344	0.4108	0.0	0.0119	0.0	0.6629	0.0021	0.0034	0.2244	0.0	0.0105	0.0	0.2913	0.0021	0.0033
1.5931	2.22	360	1.7170	0.0838	0.1619	0.3365	0.4061	0.0	0.0098	0.0	0.6414	0.0028	0.0730	0.2228	0.0	0.0090	0.0	0.2882	0.0027	0.0641
1.2434	2.35	380	1.7437	0.0844	0.1619	0.3364	0.4734	0.0	0.0118	0.0	0.5785	0.0028	0.0667	0.2391	0.0	0.0105	0.0	0.2777	0.0027	0.0607
1.4071	2.47	400	1.7316	0.0823	0.1639	0.3433	0.3054	0.0	0.0053	0.0	0.7633	0.0057	0.0674	0.1933	0.0	0.0051	0.0	0.3125	0.0054	0.0596
1.2177	2.59	420	1.7195	0.0848	0.1657	0.3459	0.3442	0.0	0.0030	0.0	0.7315	0.0110	0.0705	0.2084	0.0	0.0029	0.0	0.3111	0.0102	0.0613
1.3724	2.72	440	1.7359	0.0843	0.1660	0.3455	0.3091	0.0	0.0089	0.0	0.7620	0.0057	0.0761	0.1946	0.0	0.0084	0.0	0.3147	0.0054	0.0669
1.3973	2.84	460	1.7469	0.0827	0.1617	0.3352	0.3153	0.0	0.0101	0.0	0.7231	0.0109	0.0724	0.1922	0.0	0.0088	0.0	0.3045	0.0099	0.0638
1.3098	2.96	480	1.7193	0.0852	0.1658	0.3447	0.3412	0.0	0.0032	0.0	0.7240	0.0038	0.0887	0.2039	0.0	0.0031	0.0	0.3076	0.0037	0.0779
0.9545	3.09	500	1.7256	0.0840	0.1627	0.3359	0.3366	0.0	0.0026	0.0	0.6959	0.0077	0.0960	0.2007	0.0	0.0025	0.0	0.2969	0.0072	0.0808
1.176	3.21	520	1.7334	0.0827	0.1616	0.3357	0.3396	0.0	0.0024	0.0	0.6963	0.0016	0.0915	0.1999	0.0	0.0024	0.0	0.2966	0.0015	0.0785
1.5622	3.33	540	1.7790	0.0689	0.1528	0.3311	0.2393	0.0	0.0015	0.0	0.8186	0.0009	0.0091	0.1638	0.0	0.0015	0.0	0.3073	0.0009	0.0089
1.2673	3.46	560	1.7339	0.0803	0.1585	0.3302	0.3228	0.0	0.0113	0.0	0.7035	0.0036	0.0686	0.1930	0.0	0.0104	0.0	0.2942	0.0034	0.0613
1.418	3.58	580	1.7648	0.0760	0.1563	0.3325	0.3074	0.0	0.0022	0.0	0.7416	0.0041	0.0386	0.1902	0.0	0.0022	0.0	0.2994	0.0038	0.0363
1.3578	3.7	600	1.7338	0.0845	0.1619	0.3327	0.3548	0.0	0.0119	0.0	0.6693	0.0076	0.0898	0.2039	0.0	0.0108	0.0	0.2940	0.0070	0.0758
1.1991	3.83	620	1.7711	0.0761	0.1546	0.3285	0.3473	0.0	0.0049	0.0	0.6906	0.0018	0.0379	0.1995	0.0	0.0046	0.0	0.2912	0.0017	0.0356
1.3699	3.95	640	1.7421	0.0829	0.1595	0.3312	0.4290	0.0	0.0048	0.0	0.5994	0.0032	0.0804	0.2223	0.0	0.0045	0.0	0.2819	0.0031	0.0687
1.308	4.07	660	1.7769	0.0709	0.1518	0.3250	0.2512	0.0	0.0035	0.0	0.7784	0.0034	0.0260	0.1656	0.0	0.0033	0.0	0.2996	0.0033	0.0245
1.3746	4.2	680	1.7811	0.0749	0.1538	0.3283	0.3497	0.0	0.0077	0.0	0.6953	0.0037	0.0200	0.2022	0.0	0.0070	0.0	0.2928	0.0035	0.0191
1.2085	4.32	700	1.7401	0.0825	0.1600	0.3319	0.3632	0.0	0.0106	0.0	0.6663	0.0042	0.0759	0.2064	0.0	0.0096	0.0	0.2912	0.0040	0.0664
0.8119	4.44	720	1.7638	0.0755	0.1539	0.3284	0.3660	0.0	0.0060	0.0	0.6784	0.0034	0.0237	0.2073	0.0	0.0053	0.0	0.2902	0.0033	0.0226
1.1547	4.57	740	1.7581	0.0795	0.1573	0.3289	0.3410	0.0	0.0140	0.0	0.6879	0.0030	0.0550	0.1996	0.0	0.0116	0.0	0.2929	0.0029	0.0497
1.2229	4.69	760	1.7817	0.0730	0.1550	0.3243	0.1861	0.0	0.0124	0.0	0.8198	0.0036	0.0631	0.1365	0.0	0.0103	0.0	0.3050	0.0034	0.0557
1.3332	4.81	780	1.7580	0.0769	0.1563	0.3276	0.2656	0.0	0.0044	0.0	0.7524	0.0042	0.0677	0.1721	0.0	0.0041	0.0	0.2998	0.0040	0.0581
1.1668	4.94	800	1.7456	0.0841	0.1629	0.3309	0.2728	0.0	0.0194	0.0	0.7325	0.0055	0.1102	0.1767	0.0	0.0157	0.0	0.3002	0.0052	0.0908

Framework versions

Transformers 4.33.0
Pytorch 2.0.0
Datasets 2.1.0
Tokenizers 0.13.3

Task: Image Segmentation
Library: transformers
Type: segformer
License: other
Created: Nov 5, 2023
Updated: Nov 7, 2023