segformer-b-finetuned-segments-sidewalk-2

dwang-LI/segformer-b-finetuned-segments-sidewalk-2

segformer-b-finetuned-segments-sidewalk-2 is a image segmentation model from dwang-LI. 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

7% of all-time downloads

All-time downloads

Public

Repo size

29.9 MB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin15 MB · 100%

At a glance

Task: Image Segmentation
Library: transformers
License: other
Model type: segformer
Access: Public
Created: Sep 16, 2023
Updated: Sep 17, 2023
SHA: 750d36e2

Base models

nvidia/mit-b0

Model card

From the Hugging Face model README

segformer-b-finetuned-segments-sidewalk-2

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

Loss: 0.7733
Mean Iou: 0.2394
Mean Accuracy: 0.2885
Overall Accuracy: 0.8145
Accuarcy Unlabeled: nan
Accuarcy Flat-road: 0.9002
Accuarcy Flat-sidewalk: 0.9256
Accuarcy Flat-crosswalk: 0.6731
Accuarcy Flat-cyclinglane: 0.7624
Accuarcy Flat-parkingdriveway: 0.3720
Accuarcy Flat-railtrack: nan
Accuarcy Flat-curb: 0.3753
Accuarcy Human-person: 0.0482
Accuarcy Human-rider: 0.0
Accuarcy Vehicle-car: 0.9125
Accuarcy Vehicle-truck: 0.0
Accuarcy Vehicle-bus: 0.0
Accuarcy Vehicle-tramtrain: nan
Accuarcy Vehicle-motorcycle: 0.0
Accuarcy Vehicle-bicycle: 0.0
Accuarcy Vehicle-caravan: 0.0
Accuarcy Vehicle-cartrailer: 0.0
Accuarcy Construction-building: 0.8988
Accuarcy Construction-door: 0.0
Accuarcy Construction-wall: 0.3240
Accuarcy Construction-fenceguardrail: 0.0009
Accuarcy Construction-bridge: 0.0
Accuarcy Construction-tunnel: nan
Accuarcy Construction-stairs: 0.0
Accuarcy Object-pole: 0.0228
Accuarcy Object-trafficsign: 0.0
Accuarcy Object-trafficlight: 0.0
Accuarcy Nature-vegetation: 0.9283
Accuarcy Nature-terrain: 0.8528
Accuarcy Sky: 0.9460
Accuarcy Void-ground: 0.0
Accuarcy Void-dynamic: 0.0
Accuarcy Void-static: 0.0014
Accuarcy Void-unclear: 0.0
Iou Unlabeled: nan
Iou Flat-road: 0.7132
Iou Flat-sidewalk: 0.8399
Iou Flat-crosswalk: 0.5677
Iou Flat-cyclinglane: 0.6711
Iou Flat-parkingdriveway: 0.2585
Iou Flat-railtrack: nan
Iou Flat-curb: 0.3157
Iou Human-person: 0.0474
Iou Human-rider: 0.0
Iou Vehicle-car: 0.7025
Iou Vehicle-truck: 0.0
Iou Vehicle-bus: 0.0
Iou Vehicle-tramtrain: nan
Iou Vehicle-motorcycle: 0.0
Iou Vehicle-bicycle: 0.0
Iou Vehicle-caravan: 0.0
Iou Vehicle-cartrailer: 0.0
Iou Construction-building: 0.6194
Iou Construction-door: 0.0
Iou Construction-wall: 0.2615
Iou Construction-fenceguardrail: 0.0009
Iou Construction-bridge: 0.0
Iou Construction-tunnel: nan
Iou Construction-stairs: 0.0
Iou Object-pole: 0.0227
Iou Object-trafficsign: 0.0
Iou Object-trafficlight: 0.0
Iou Nature-vegetation: 0.7851
Iou Nature-terrain: 0.7352
Iou Sky: 0.8791
Iou Void-ground: 0.0
Iou Void-dynamic: 0.0
Iou Void-static: 0.0014
Iou Void-unclear: 0.0

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: 0.0001
train_batch_size: 24
eval_batch_size: 24
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 12

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuarcy Unlabeled	Accuarcy Flat-road	Accuarcy Flat-sidewalk	Accuarcy Flat-crosswalk	Accuarcy Flat-cyclinglane	Accuarcy Flat-parkingdriveway	Accuarcy Flat-railtrack	Accuarcy Flat-curb	Accuarcy Human-person	Accuarcy Vehicle-car	Accuarcy Vehicle-tramtrain	Accuarcy Construction-building	Accuarcy Construction-wall	Accuarcy Construction-fenceguardrail	Accuarcy Construction-tunnel	Accuarcy Object-pole	Accuarcy Nature-vegetation	Accuarcy Nature-terrain	Accuarcy Sky	Accuarcy Void-static	Iou Unlabeled	Iou Flat-road	Iou Flat-sidewalk	Iou Flat-crosswalk	Iou Flat-cyclinglane	Iou Flat-parkingdriveway	Iou Flat-railtrack	Iou Flat-curb	Iou Human-person	Iou Vehicle-car	Iou Vehicle-tramtrain	Iou Construction-building	Iou Construction-wall	Iou Construction-fenceguardrail	Iou Construction-tunnel	Iou Object-pole	Iou Nature-vegetation	Iou Nature-terrain	Iou Sky	Iou Void-static
2.2728	0.59	20	2.3946	0.1035	0.1549	0.6540	nan	0.6440	0.9384	0.0	0.0006	0.0001	nan	0.0001	0.0	0.9243	nan	0.6269	0.0000	0.0002	nan	0.0	0.9320	0.0116	0.7234	0.0	nan	0.4920	0.6851	0.0	0.0006	0.0001	nan	0.0001	0.0	0.3557	nan	0.4837	0.0000	0.0002	0.0	0.0	0.5828	0.0115	0.7007	0.0
1.9006	1.18	40	1.7230	0.1153	0.1706	0.6814	nan	0.8635	0.8762	0.0	0.0003	0.0003	nan	0.0	0.0	0.8614	nan	0.8115	0.0	0.0	nan	0.0	0.9315	0.0405	0.9034	0.0	nan	0.4876	0.7405	0.0	0.0003	0.0003	nan	0.0	0.0	0.5225	nan	0.5210	0.0	0.0	0.0	0.0	0.6069	0.0399	0.7696	0.0
1.6721	1.76	60	1.4574	0.1289	0.1783	0.6968	nan	0.8799	0.8822	0.0	0.0528	0.0003	nan	0.0	0.0	0.8812	nan	0.8573	0.0000	0.0	nan	0.0	0.9298	0.1473	0.8959	0.0	nan	0.4937	0.7555	0.0	0.0519	0.0003	nan	0.0	0.0	0.5454	nan	0.5547	0.0000	0.0	nan	0.0	0.6303	0.1427	0.8205	0.0
1.4066	2.35	80	1.3422	0.1589	0.2055	0.7457	nan	0.8230	0.9475	0.0	0.3015	0.0047	nan	0.0000	0.0	0.8977	nan	0.8695	0.0000	0.0	nan	0.0	0.9091	0.6841	0.9322	0.0	nan	0.6093	0.7599	0.0	0.2787	0.0046	nan	0.0000	0.0	0.5489	nan	0.5596	0.0000	0.0	nan	0.0	0.7275	0.6092	0.8285	0.0
1.3429	2.94	100	1.1795	0.1653	0.2103	0.7562	nan	0.8569	0.9495	0.0	0.3507	0.0066	nan	0.0000	0.0	0.8981	nan	0.8869	0.0001	0.0	nan	0.0	0.9026	0.7728	0.8950	0.0	nan	0.6153	0.7730	0.0	0.3326	0.0065	nan	0.0000	0.0	0.5899	nan	0.5742	0.0001	0.0	nan	0.0	0.7403	0.6481	0.8448	0.0
1.2661	3.53	120	1.1008	0.1712	0.2174	0.7629	nan	0.8484	0.9495	0.0	0.4917	0.0181	nan	0.0001	0.0	0.8996	nan	0.9043	0.0001	0.0	nan	0.0	0.8869	0.8036	0.9371	0.0	nan	0.6100	0.7894	0.0	0.4346	0.0175	nan	0.0001	0.0	0.6153	nan	0.5608	0.0001	0.0	nan	0.0	0.7533	0.6752	0.8508	0.0
1.2166	4.12	140	1.0514	0.1771	0.2232	0.7695	nan	0.8815	0.9342	0.0	0.5539	0.0713	nan	0.0030	0.0	0.9014	nan	0.9029	0.0016	0.0	nan	0.0	0.9068	0.8398	0.9225	0.0	nan	0.6195	0.7981	0.0	0.5017	0.0642	nan	0.0030	0.0	0.6222	nan	0.5694	0.0016	0.0	nan	0.0	0.7585	0.6979	0.8546	0.0
1.0262	4.71	160	1.0025	0.1782	0.2236	0.7665	nan	0.9188	0.9111	0.0	0.5462	0.1006	nan	0.0031	0.0	0.8814	nan	0.8881	0.0027	0.0	nan	0.0	0.9260	0.8130	0.9404	0.0	nan	0.5776	0.8071	0.0	0.5005	0.0888	nan	0.0031	0.0	0.6651	nan	0.5803	0.0027	0.0	nan	0.0	0.7415	0.7028	0.8558	0.0
1.0928	5.29	180	0.9698	0.1852	0.2308	0.7778	nan	0.8513	0.9428	0.0	0.6760	0.1497	nan	0.0419	0.0	0.8856	nan	0.9132	0.0056	0.0	nan	0.0002	0.9134	0.8535	0.9219	0.0	nan	0.6410	0.8062	0.0	0.5617	0.1228	nan	0.0405	0.0	0.6597	nan	0.5705	0.0056	0.0	nan	0.0002	0.7603	0.7081	0.8642	0.0
0.8736	5.88	200	0.9250	0.1906	0.2370	0.7850	nan	0.9149	0.9249	0.0001	0.7226	0.1944	nan	0.0715	0.0027	0.8853	nan	0.8917	0.0153	0.0	nan	0.0005	0.9353	0.8470	0.9402	0.0	nan	0.6511	0.8250	0.0001	0.5978	0.1516	nan	0.0682	0.0027	0.6817	nan	0.5862	0.0152	0.0	nan	0.0005	0.7477	0.7159	0.8635	0.0
0.7832	6.47	220	0.8852	0.1961	0.2421	0.7875	nan	0.8962	0.9385	0.0642	0.6975	0.2064	nan	0.1581	0.0003	0.8995	nan	0.9011	0.0392	0.0	nan	0.0009	0.8974	0.8728	0.9342	0.0	nan	0.6576	0.8222	0.0624	0.6239	0.1577	nan	0.1421	0.0003	0.6802	nan	0.5989	0.0383	0.0	nan	0.0009	0.7547	0.6706	0.8700	0.0
0.7822	7.06	240	0.8621	0.2145	0.2598	0.7992	nan	0.8827	0.9398	0.4415	0.7426	0.2656	nan	0.2218	0.0023	0.8967	nan	0.9092	0.0558	0.0000	nan	0.0020	0.9249	0.8259	0.9429	0.0	nan	0.6911	0.8250	0.3902	0.6320	0.2017	nan	0.1950	0.0023	0.6915	nan	0.5886	0.0540	0.0000	nan	0.0020	0.7732	0.7329	0.8703	0.0
0.6742	7.65	260	0.8371	0.2193	0.2667	0.8027	nan	0.8766	0.9312	0.3983	0.7724	0.2975	nan	0.2975	0.0055	0.9111	nan	0.9086	0.1602	0.0001	nan	0.0034	0.9371	0.8321	0.9353	0.0000	nan	0.6894	0.8388	0.3591	0.6398	0.2119	nan	0.2519	0.0055	0.6754	nan	0.6033	0.1492	0.0001	nan	0.0034	0.7671	0.7293	0.8750	0.0000
0.8116	8.24	280	0.8277	0.2314	0.2819	0.8087	nan	0.8894	0.9207	0.6812	0.7773	0.3594	nan	0.3120	0.0109	0.9016	nan	0.8885	0.2424	0.0005	nan	0.0107	0.9398	0.8575	0.9461	0.0003	nan	0.7112	0.8407	0.5738	0.6399	0.2424	nan	0.2666	0.0108	0.6924	nan	0.6145	0.2148	0.0005	nan	0.0106	0.7579	0.7244	0.8738	0.0003
0.7791	8.82	300	0.8059	0.2255	0.2723	0.8077	nan	0.8684	0.9414	0.4680	0.7998	0.2901	nan	0.3174	0.0107	0.8846	nan	0.9111	0.2193	0.0000	nan	0.0099	0.9290	0.8439	0.9465	0.0000	nan	0.7039	0.8383	0.4188	0.6308	0.2131	nan	0.2698	0.0106	0.7114	nan	0.6008	0.1942	0.0000	nan	0.0099	0.7791	0.7343	0.8760	0.0000
0.7334	9.41	320	0.7962	0.2342	0.2830	0.8117	nan	0.8921	0.9332	0.6837	0.7454	0.3381	nan	0.3264	0.0298	0.9198	nan	0.9018	0.2712	0.0003	nan	0.0182	0.9194	0.8508	0.9434	0.0008	nan	0.7121	0.8388	0.5627	0.6590	0.2316	nan	0.2794	0.0296	0.6884	nan	0.6204	0.2324	0.0003	nan	0.0182	0.7820	0.7278	0.8762	0.0008
0.7645	10.0	340	0.7783	0.2342	0.2809	0.8133	nan	0.8999	0.9347	0.5997	0.7491	0.3278	nan	0.3613	0.0164	0.9043	nan	0.9156	0.2684	0.0003	nan	0.0167	0.9235	0.8454	0.9455	0.0007	nan	0.7218	0.8409	0.5162	0.6738	0.2390	nan	0.3039	0.0162	0.7015	nan	0.6019	0.2260	0.0003	nan	0.0167	0.7860	0.7381	0.8764	0.0007
0.6792	10.59	360	0.7774	0.2358	0.2841	0.8141	nan	0.8954	0.9341	0.6272	0.7826	0.3543	nan	0.3360	0.0300	0.9162	nan	0.8964	0.2909	0.0005	nan	0.0199	0.9226	0.8558	0.9443	0.0010	nan	0.7198	0.8402	0.5426	0.6699	0.2489	nan	0.2900	0.0297	0.6966	nan	0.6186	0.2450	0.0005	nan	0.0199	0.7835	0.7251	0.8784	0.0010
0.8047	11.18	380	0.7734	0.2388	0.2878	0.8147	nan	0.8924	0.9265	0.6512	0.7739	0.3846	nan	0.3762	0.0383	0.9122	nan	0.9053	0.3142	0.0005	nan	0.0216	0.9303	0.8513	0.9427	0.0014	nan	0.7171	0.8421	0.5575	0.6761	0.2609	nan	0.3165	0.0376	0.6982	nan	0.6155	0.2551	0.0005	nan	0.0215	0.7854	0.7377	0.8797	0.0014
0.7136	11.76	400	0.7733	0.2394	0.2885	0.8145	nan	0.9002	0.9256	0.6731	0.7624	0.3720	nan	0.3753	0.0482	0.9125	nan	0.8988	0.3240	0.0009	nan	0.0228	0.9283	0.8528	0.9460	0.0014	nan	0.7132	0.8399	0.5677	0.6711	0.2585	nan	0.3157	0.0474	0.7025	nan	0.6194	0.2615	0.0009	nan	0.0227	0.7851	0.7352	0.8791	0.0014

Framework versions

Transformers 4.32.1
Pytorch 2.0.1
Datasets 2.12.0
Tokenizers 0.13.2

Task: Image Segmentation
Library: transformers
Type: segformer
License: other
Created: Sep 16, 2023
Updated: Sep 17, 2023