segformer-b0-finetuned-segments-sidewalk-3

vanniew/segformer-b0-finetuned-segments-sidewalk-3

segformer-b0-finetuned-segments-sidewalk-3 is a image segmentation model from vanniew. 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

4% of all-time downloads

All-time downloads

Public

Parameters

3.7M

29.8 MB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors14.9 MB · 100%

At a glance

Task: Image Segmentation
Library: transformers
License: other
Model type: segformer
Access: Public
Created: Jan 7, 2025
Updated: Jan 8, 2025
SHA: cfb6d523

Base models

nvidia/mit-b0

Model card

From the Hugging Face model README

segformer-b0-finetuned-segments-sidewalk-3

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: 0.5669
Mean Iou: 0.3190
Mean Accuracy: 0.3821
Overall Accuracy: 0.8574
Accuracy Unlabeled: nan
Accuracy Flat-road: 0.9199
Accuracy Flat-sidewalk: 0.9566
Accuracy Flat-crosswalk: 0.5855
Accuracy Flat-cyclinglane: 0.8187
Accuracy Flat-parkingdriveway: 0.4712
Accuracy Flat-railtrack: nan
Accuracy Flat-curb: 0.5293
Accuracy Human-person: 0.7962
Accuracy Human-rider: 0.0
Accuracy Vehicle-car: 0.9245
Accuracy Vehicle-truck: 0.0
Accuracy Vehicle-bus: 0.0
Accuracy Vehicle-tramtrain: nan
Accuracy Vehicle-motorcycle: 0.0
Accuracy Vehicle-bicycle: 0.5123
Accuracy Vehicle-caravan: 0.0
Accuracy Vehicle-cartrailer: 0.0
Accuracy Construction-building: 0.8912
Accuracy Construction-door: 0.0
Accuracy Construction-wall: 0.5405
Accuracy Construction-fenceguardrail: 0.3863
Accuracy Construction-bridge: 0.0
Accuracy Construction-tunnel: nan
Accuracy Construction-stairs: 0.0
Accuracy Object-pole: 0.4414
Accuracy Object-trafficsign: 0.0
Accuracy Object-trafficlight: 0.0
Accuracy Nature-vegetation: 0.9331
Accuracy Nature-terrain: 0.8636
Accuracy Sky: 0.9713
Accuracy Void-ground: 0.0000
Accuracy Void-dynamic: 0.0
Accuracy Void-static: 0.3037
Accuracy Void-unclear: 0.0
Iou Unlabeled: nan
Iou Flat-road: 0.7635
Iou Flat-sidewalk: 0.8715
Iou Flat-crosswalk: 0.5513
Iou Flat-cyclinglane: 0.7441
Iou Flat-parkingdriveway: 0.3799
Iou Flat-railtrack: nan
Iou Flat-curb: 0.4143
Iou Human-person: 0.5353
Iou Human-rider: 0.0
Iou Vehicle-car: 0.8080
Iou Vehicle-truck: 0.0
Iou Vehicle-bus: 0.0
Iou Vehicle-tramtrain: nan
Iou Vehicle-motorcycle: 0.0
Iou Vehicle-bicycle: 0.3478
Iou Vehicle-caravan: 0.0
Iou Vehicle-cartrailer: 0.0
Iou Construction-building: 0.7370
Iou Construction-door: 0.0
Iou Construction-wall: 0.3749
Iou Construction-fenceguardrail: 0.3203
Iou Construction-bridge: 0.0
Iou Construction-tunnel: nan
Iou Construction-stairs: 0.0
Iou Object-pole: 0.3284
Iou Object-trafficsign: 0.0
Iou Object-trafficlight: 0.0
Iou Nature-vegetation: 0.8586
Iou Nature-terrain: 0.7283
Iou Sky: 0.9243
Iou Void-ground: 0.0000
Iou Void-dynamic: 0.0
Iou Void-static: 0.2014
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: 6e-05
train_batch_size: 4
eval_batch_size: 4
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Unlabeled	Accuracy Flat-road	Accuracy Flat-sidewalk	Accuracy Flat-crosswalk	Accuracy Flat-cyclinglane	Accuracy Flat-parkingdriveway	Accuracy Flat-railtrack	Accuracy Flat-curb	Accuracy Human-person	Accuracy Vehicle-car	Accuracy Vehicle-tramtrain	Accuracy Vehicle-bicycle	Accuracy Construction-building	Accuracy Construction-wall	Accuracy Construction-fenceguardrail	Accuracy Construction-tunnel	Accuracy Object-pole	Accuracy Nature-vegetation	Accuracy Nature-terrain	Accuracy Sky	Accuracy Void-ground	Accuracy 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 Vehicle-bicycle	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-ground	Iou Void-static
1.2409	1.0	200	1.2471	0.1570	0.2044	0.7356	nan	0.8032	0.9526	0.0	0.3067	0.0000	nan	0.0	0.0	0.9054	nan	0.0	0.8352	0.0007	0.0	nan	0.0	0.9434	0.6803	0.9090	0.0	0.0	nan	0.5173	0.7510	0.0	0.2997	0.0000	nan	0.0	0.0	0.5787	nan	0.0	0.5852	0.0007	0.0	nan	0.0	0.7164	0.5939	0.8246	0.0	0.0
1.1443	2.0	400	0.9634	0.1732	0.2195	0.7607	nan	0.8231	0.9651	0.0	0.5125	0.0159	nan	0.0005	0.0	0.9316	nan	0.0	0.8419	0.0476	0.0	nan	0.0	0.9224	0.8029	0.9413	0.0	0.0	nan	0.5515	0.7650	0.0	0.4662	0.0157	nan	0.0005	0.0	0.6181	nan	0.0	0.6102	0.0440	0.0	nan	0.0	0.7835	0.6602	0.8541	0.0	0.0
0.8805	3.0	600	0.8309	0.1963	0.2429	0.7804	nan	0.8858	0.9438	0.0	0.5258	0.2765	nan	0.2373	0.0	0.8852	nan	0.0	0.9159	0.1947	0.0000	nan	0.0004	0.9174	0.8250	0.9226	0.0	0.0	nan	0.5450	0.8189	0.0	0.4853	0.2291	nan	0.1953	0.0	0.7014	nan	0.0	0.6219	0.1442	0.0000	nan	0.0004	0.7977	0.6821	0.8641	0.0	0.0
0.5657	4.0	800	0.7312	0.2177	0.2662	0.8045	nan	0.8296	0.9595	0.0118	0.7076	0.3690	nan	0.3339	0.0056	0.9077	nan	0.0	0.9036	0.4576	0.0102	nan	0.0242	0.9308	0.8451	0.9532	0.0	0.0040	nan	0.6207	0.8228	0.0118	0.6313	0.2862	nan	0.2661	0.0055	0.7236	nan	0.0	0.6724	0.2571	0.0102	nan	0.0240	0.8220	0.7057	0.8862	0.0	0.0039
0.7813	5.0	1000	0.6767	0.2476	0.2950	0.8246	nan	0.8929	0.9575	0.3536	0.7434	0.3652	nan	0.3249	0.1722	0.9367	nan	0.0191	0.9078	0.4430	0.1157	nan	0.1126	0.9239	0.8638	0.9630	0.0	0.0504	nan	0.6867	0.8456	0.3488	0.6502	0.3199	nan	0.2679	0.1656	0.7052	nan	0.0189	0.6895	0.2999	0.1092	nan	0.1055	0.8259	0.6999	0.8927	0.0	0.0447
0.7435	6.0	1200	0.6344	0.2760	0.3279	0.8357	nan	0.9055	0.9490	0.5276	0.7616	0.4465	nan	0.4423	0.6483	0.9186	nan	0.0494	0.9033	0.4453	0.1511	nan	0.2289	0.9545	0.7769	0.9534	0.0	0.1015	nan	0.7209	0.8647	0.5022	0.6648	0.3552	nan	0.3400	0.4843	0.7524	nan	0.0493	0.7024	0.3020	0.1429	nan	0.1947	0.8107	0.6785	0.9045	0.0	0.0868
0.669	7.0	1400	0.6076	0.2955	0.3517	0.8449	nan	0.9070	0.9577	0.6665	0.7372	0.4645	nan	0.4506	0.7302	0.9348	nan	0.1936	0.8990	0.5097	0.2575	nan	0.2668	0.9282	0.8178	0.9587	0.0	0.2242	nan	0.7469	0.8615	0.5876	0.6676	0.3702	nan	0.3492	0.5076	0.7583	nan	0.1678	0.7182	0.3413	0.2286	nan	0.2275	0.8431	0.7139	0.9073	0.0	0.1654
0.4628	8.0	1600	0.6011	0.2977	0.3549	0.8429	nan	0.8884	0.9527	0.5578	0.7506	0.4890	nan	0.4638	0.7406	0.9319	nan	0.3178	0.9010	0.4809	0.3284	nan	0.2448	0.9241	0.8614	0.9616	0.0	0.2085	nan	0.7334	0.8587	0.5336	0.6751	0.3767	nan	0.3613	0.5230	0.7468	nan	0.2516	0.7101	0.3365	0.2738	nan	0.2136	0.8427	0.7254	0.9085	0.0	0.1589
1.3419	9.0	1800	0.5904	0.3029	0.3709	0.8462	nan	0.9057	0.9510	0.6361	0.7788	0.4788	nan	0.4900	0.8107	0.9469	nan	0.4336	0.8507	0.5368	0.3238	nan	0.2916	0.9233	0.8610	0.9637	0.0	0.3155	nan	0.7455	0.8668	0.5717	0.7045	0.3669	nan	0.3797	0.4684	0.7556	nan	0.2745	0.7145	0.3328	0.2865	nan	0.2461	0.8508	0.7261	0.9126	0.0	0.1879
0.5734	10.0	2000	0.5937	0.3046	0.3620	0.8455	nan	0.8903	0.9576	0.4874	0.7802	0.4436	nan	0.5076	0.7526	0.9252	nan	0.3795	0.8933	0.5108	0.3422	nan	0.3250	0.9355	0.8361	0.9634	0.0	0.2901	nan	0.7358	0.8573	0.4710	0.6988	0.3621	nan	0.3880	0.5318	0.7827	nan	0.3118	0.7178	0.3584	0.2906	nan	0.2654	0.8469	0.7203	0.9165	0.0	0.1862
0.3721	11.0	2200	0.5782	0.3079	0.3675	0.8503	nan	0.9192	0.9460	0.5240	0.8234	0.4822	nan	0.5023	0.7689	0.8965	nan	0.4271	0.9199	0.5107	0.3373	nan	0.4079	0.9468	0.8152	0.9624	0.0	0.2037	nan	0.7498	0.8668	0.5037	0.7028	0.3740	nan	0.3838	0.5157	0.7946	nan	0.3237	0.7340	0.3586	0.2842	nan	0.3043	0.8469	0.7219	0.9174	0.0	0.1618
0.5698	12.0	2400	0.5601	0.3147	0.3796	0.8547	nan	0.9163	0.9502	0.6118	0.8102	0.5160	nan	0.4905	0.7891	0.9196	nan	0.4718	0.8898	0.5471	0.4070	nan	0.4158	0.9320	0.8523	0.9700	0.0	0.2787	nan	0.7765	0.8684	0.5737	0.7307	0.3860	nan	0.3917	0.5271	0.7968	nan	0.2862	0.7351	0.3521	0.3301	nan	0.3084	0.8525	0.7310	0.9171	0.0	0.1921
0.4268	13.0	2600	0.5706	0.3144	0.3758	0.8546	nan	0.9149	0.9584	0.5832	0.8127	0.4595	nan	0.5052	0.7785	0.9385	nan	0.4538	0.8974	0.5469	0.3876	nan	0.3971	0.9269	0.8398	0.9644	0.0	0.2853	nan	0.7632	0.8670	0.5583	0.7345	0.3711	nan	0.3931	0.5330	0.7866	nan	0.3082	0.7333	0.3694	0.3199	nan	0.3094	0.8565	0.7309	0.9212	0.0	0.1913
0.3612	14.0	2800	0.5615	0.3182	0.3797	0.8587	nan	0.9143	0.9554	0.6527	0.8256	0.4717	nan	0.5467	0.7521	0.9229	nan	0.4951	0.9226	0.5236	0.3953	nan	0.4144	0.9329	0.8347	0.9637	0.0	0.2473	nan	0.7803	0.8714	0.5854	0.7509	0.3843	nan	0.4085	0.5363	0.7994	nan	0.3118	0.7381	0.3688	0.3250	nan	0.3138	0.8547	0.7242	0.9233	0.0	0.1865
0.2714	15.0	3000	0.5687	0.3165	0.3819	0.8557	nan	0.9102	0.9546	0.5390	0.8252	0.4823	nan	0.5408	0.8170	0.9223	nan	0.5004	0.8929	0.5402	0.4228	nan	0.4218	0.9324	0.8585	0.9674	0.0	0.3121	nan	0.7610	0.8703	0.5177	0.7369	0.3866	nan	0.4110	0.5213	0.8051	nan	0.3223	0.7368	0.3795	0.3303	nan	0.3181	0.8558	0.7279	0.9245	0.0	0.2057
0.3566	16.0	3200	0.5746	0.3154	0.3831	0.8543	nan	0.9156	0.9537	0.5807	0.8263	0.4642	nan	0.5228	0.8130	0.9300	nan	0.5625	0.8697	0.5473	0.3708	nan	0.4250	0.9320	0.8671	0.9687	0.0	0.3279	nan	0.7622	0.8696	0.5476	0.7333	0.3773	nan	0.4062	0.5234	0.8021	nan	0.3295	0.7271	0.3621	0.3111	nan	0.3157	0.8569	0.7294	0.9246	0.0	0.1997
0.3486	17.0	3400	0.5716	0.3168	0.3810	0.8550	nan	0.9087	0.9592	0.5998	0.7778	0.4799	nan	0.5562	0.7959	0.9295	nan	0.4951	0.8892	0.5531	0.3860	nan	0.4274	0.9295	0.8539	0.9663	0.0	0.3037	nan	0.7539	0.8708	0.5570	0.7185	0.3903	nan	0.4183	0.5347	0.7991	nan	0.3381	0.7322	0.3614	0.3128	nan	0.3210	0.8602	0.7322	0.9260	0.0	0.1958
0.2541	18.0	3600	0.5641	0.3185	0.3816	0.8573	nan	0.9128	0.9560	0.6054	0.8210	0.4610	nan	0.5325	0.7977	0.9235	nan	0.5042	0.9042	0.5403	0.3783	nan	0.4368	0.9330	0.8523	0.9728	0.0	0.2982	nan	0.7718	0.8707	0.5635	0.7389	0.3730	nan	0.4092	0.5344	0.8080	nan	0.3504	0.7379	0.3729	0.3107	nan	0.3226	0.8587	0.7271	0.9226	0.0	0.2016
0.3204	19.0	3800	0.5781	0.3170	0.3798	0.8554	nan	0.9236	0.9609	0.5748	0.7706	0.4581	nan	0.5274	0.7996	0.9291	nan	0.5223	0.9019	0.5425	0.3743	nan	0.4318	0.9265	0.8595	0.9685	0.0	0.3016	nan	0.7477	0.8692	0.5451	0.7127	0.3767	nan	0.4137	0.5304	0.8061	nan	0.3493	0.7376	0.3776	0.3140	nan	0.3249	0.8609	0.7330	0.9252	0.0	0.2027
0.3648	20.0	4000	0.5669	0.3190	0.3821	0.8574	nan	0.9199	0.9566	0.5855	0.8187	0.4712	nan	0.5293	0.7962	0.9245	nan	0.5123	0.8912	0.5405	0.3863	nan	0.4414	0.9331	0.8636	0.9713	0.0000	0.3037	nan	0.7635	0.8715	0.5513	0.7441	0.3799	nan	0.4143	0.5353	0.8080	nan	0.3478	0.7370	0.3749	0.3203	nan	0.3284	0.8586	0.7283	0.9243	0.0000	0.2014

Framework versions

Transformers 4.47.1
Pytorch 2.5.1+cu124
Datasets 3.2.0
Tokenizers 0.21.0

Task: Image Segmentation
Library: transformers
Type: segformer
License: other
Created: Jan 7, 2025
Updated: Jan 8, 2025