tf_segformer_for_semantic_segmentation_v3

ArtemShyshko/tf_segformer_for_semantic_segmentation_v3

tf_segformer_for_semantic_segmentation_v3 is a machine learning model from ArtemShyshko. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as other.

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

Downloads · 30 days

19% of all-time downloads

All-time downloads

Public

Repo size

15.1 MB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.h515.1 MB · 100%

At a glance

License: other
Model type: segformer
Access: Public
Created: Feb 24, 2025
Updated: Feb 24, 2025
SHA: a3a2974f

Base models

nvidia/mit-b0

Model card

From the Hugging Face model README

tf_segformer_for_semantic_segmentation_v3

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

Train Loss: 0.2414
Validation Loss: 0.2879
Validation Mean Iou: 0.6578
Validation Mean Accuracy: 0.8810
Validation Overall Accuracy: 0.9074
Validation Accuracy Unlabeled: nan
Validation Accuracy Building: 0.9097
Validation Accuracy Land: 0.9396
Validation Accuracy Road: 0.8000
Validation Accuracy Vegetation: 0.8165
Validation Accuracy Water: 0.9392
Validation Iou Unlabeled: 0.0
Validation Iou Building: 0.7772
Validation Iou Land: 0.9028
Validation Iou Road: 0.6828
Validation Iou Vegetation: 0.6714
Validation Iou Water: 0.9125
Epoch: 49

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:

optimizer: {'name': 'Adam', 'weight_decay': None, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': False, 'is_legacy_optimizer': False, 'learning_rate': 6e-05, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
training_precision: float32

Training results

Train Loss	Validation Loss	Validation Mean Iou	Validation Mean Accuracy	Validation Overall Accuracy	Validation Accuracy Unlabeled	Validation Accuracy Building	Validation Accuracy Land	Validation Accuracy Road	Validation Accuracy Vegetation	Validation Accuracy Water	Validation Iou Unlabeled	Validation Iou Building	Validation Iou Land	Validation Iou Road	Validation Iou Vegetation	Validation Iou Water	Epoch
1.4526	0.8367	0.4904	0.7407	0.7665	nan	0.8805	0.8041	0.4586	0.6741	0.8860	0.0	0.5858	0.7562	0.3475	0.4603	0.7926	0
0.9142	0.8189	0.4817	0.7306	0.7891	nan	0.8061	0.8595	0.5010	0.7384	0.7477	0.0	0.6034	0.8033	0.3891	0.4754	0.6193	1
0.8439	0.6237	0.5236	0.7804	0.8064	nan	0.8191	0.8573	0.6912	0.6527	0.8817	0.0	0.6321	0.8009	0.4659	0.5053	0.7374	2
0.7108	0.4985	0.5541	0.7800	0.8242	nan	0.8262	0.9105	0.6019	0.6743	0.8872	0.0	0.6613	0.7989	0.4979	0.5621	0.8042	3
0.6197	0.5273	0.5573	0.7976	0.8330	nan	0.8263	0.9030	0.7527	0.5885	0.9176	0.0	0.6898	0.8121	0.5267	0.5201	0.7950	4
0.6202	0.4948	0.5699	0.8110	0.8453	nan	0.8221	0.8921	0.7786	0.6705	0.8913	0.0	0.6880	0.8297	0.5585	0.5145	0.8285	5
0.5454	0.4675	0.5851	0.8227	0.8442	nan	0.8582	0.8823	0.8103	0.6061	0.9564	0.0	0.6922	0.8086	0.5854	0.5328	0.8918	6
0.5566	0.4371	0.5819	0.8162	0.8512	nan	0.8316	0.9073	0.7572	0.6785	0.9065	0.0	0.6899	0.8392	0.5751	0.5413	0.8461	7
0.5456	0.4071	0.6949	0.8177	0.8394	nan	0.8253	0.8993	0.7893	0.6329	0.9419	nan	0.6964	0.8189	0.5692	0.5349	0.8549	8
0.4690	0.4223	0.7227	0.8386	0.8695	nan	0.8670	0.9101	0.7523	0.7607	0.9030	nan	0.6947	0.8547	0.5985	0.6205	0.8450	9
0.4529	0.4232	0.7150	0.8162	0.8821	nan	0.8292	0.9446	0.7629	0.5798	0.9642	nan	0.7179	0.8763	0.5976	0.5008	0.8824	10
0.4981	0.4245	0.7228	0.8380	0.8533	nan	0.8670	0.8952	0.7708	0.7406	0.9164	nan	0.7185	0.8274	0.6025	0.6131	0.8525	11
0.4228	0.4220	0.7169	0.8353	0.8518	nan	0.8611	0.8920	0.7923	0.6990	0.9322	nan	0.7122	0.8267	0.5888	0.5889	0.8679	12
0.4255	0.3802	0.6233	0.8495	0.8762	nan	0.8466	0.9237	0.7867	0.7465	0.9441	0.0	0.7233	0.8528	0.6567	0.6183	0.8886	13
0.4457	0.4365	0.7254	0.8327	0.8693	nan	0.8500	0.9243	0.8040	0.6128	0.9722	nan	0.7290	0.8538	0.6080	0.5445	0.8917	14
0.4173	0.3841	0.6104	0.8470	0.8702	nan	0.8736	0.9048	0.7940	0.7135	0.9493	0.0	0.7304	0.8491	0.6228	0.5789	0.8812	15
0.3853	0.3850	0.7315	0.8385	0.8794	nan	0.8490	0.9274	0.8012	0.6797	0.9349	nan	0.7278	0.8684	0.6104	0.5668	0.8840	16
0.3878	0.3584	0.6125	0.8333	0.8776	nan	0.8501	0.9372	0.7605	0.6868	0.9320	0.0	0.7171	0.8644	0.6342	0.5848	0.8748	17
0.3712	0.3656	0.7346	0.8374	0.8701	nan	0.8502	0.9288	0.8080	0.6459	0.9539	nan	0.7293	0.8599	0.6076	0.5570	0.9193	18
0.3599	0.3579	0.6272	0.8596	0.8796	nan	0.8910	0.9121	0.8196	0.7660	0.9093	0.0	0.7358	0.8709	0.6404	0.6403	0.8758	19
0.3553	0.3772	0.7371	0.8366	0.8702	nan	0.8727	0.9261	0.7609	0.7066	0.9165	nan	0.7314	0.8508	0.6217	0.6103	0.8715	20
0.3575	0.3076	0.7509	0.8477	0.8782	nan	0.8486	0.9322	0.7724	0.7382	0.9472	nan	0.7268	0.8575	0.6197	0.6427	0.9080	21
0.3574	0.3587	0.6171	0.8376	0.8730	nan	0.8116	0.9386	0.8013	0.7115	0.9252	0.0	0.7260	0.8539	0.6531	0.5915	0.8783	22
0.3441	0.3546	0.6244	0.8566	0.8773	nan	0.8605	0.9106	0.8081	0.7666	0.9372	0.0	0.7214	0.8583	0.6324	0.6287	0.9054	23
0.3272	0.3879	0.6356	0.8645	0.8830	nan	0.8536	0.9187	0.7823	0.8268	0.9412	0.0	0.7377	0.8671	0.6367	0.6561	0.9160	24
0.3349	0.3440	0.6224	0.8474	0.8776	nan	0.8202	0.9365	0.7851	0.7852	0.9101	0.0	0.7372	0.8732	0.6418	0.6020	0.8802	25
0.3297	0.3825	0.6282	0.8582	0.8791	nan	0.8002	0.9170	0.7873	0.8529	0.9334	0.0	0.7102	0.8569	0.6205	0.6699	0.9117	26
0.3612	0.3368	0.6213	0.8429	0.8749	nan	0.8787	0.9287	0.7290	0.7225	0.9557	0.0	0.7496	0.8513	0.6246	0.6033	0.8989	27
0.3062	0.3057	0.6322	0.8566	0.8824	nan	0.9034	0.9204	0.7822	0.7184	0.9585	0.0	0.7574	0.8629	0.6293	0.6130	0.9304	28
0.3014	0.3295	0.6323	0.8528	0.8867	nan	0.8591	0.9423	0.7520	0.7583	0.9520	0.0	0.7497	0.8783	0.6324	0.6216	0.9117	29
0.3040	0.2963	0.6321	0.8604	0.8738	nan	0.8959	0.9064	0.7845	0.7703	0.9451	0.0	0.7565	0.8413	0.6466	0.6575	0.8904	30
0.2993	0.3428	0.6398	0.8676	0.8839	nan	0.8966	0.9122	0.8000	0.8014	0.9280	0.0	0.7597	0.8636	0.6597	0.6551	0.9005	31
0.3216	0.3200	0.6399	0.8620	0.8920	nan	0.8787	0.9374	0.7806	0.7470	0.9660	0.0	0.7557	0.8794	0.6513	0.6330	0.9199	32
0.3160	0.3261	0.6219	0.8580	0.8768	nan	0.8980	0.9022	0.8299	0.7000	0.9600	0.0	0.7370	0.8564	0.6536	0.5784	0.9062	33
0.3045	0.3550	0.6462	0.8639	0.9011	nan	0.8732	0.9450	0.8075	0.7252	0.9686	0.0	0.7674	0.8872	0.6693	0.6262	0.9273	34
0.2839	0.3102	0.6421	0.8618	0.8895	nan	0.9091	0.9314	0.7588	0.7451	0.9646	0.0	0.7695	0.8729	0.6418	0.6417	0.9264	35
0.3003	0.2942	0.6253	0.8502	0.8824	nan	0.9090	0.9207	0.8180	0.6335	0.9701	0.0	0.7566	0.8691	0.6745	0.5433	0.9081	36
0.2726	0.3068	0.6369	0.8642	0.8880	nan	0.9063	0.9213	0.8244	0.7194	0.9498	0.0	0.7534	0.8755	0.6625	0.6225	0.9077	37
0.2700	0.3095	0.6421	0.8645	0.8818	nan	0.8651	0.9209	0.7934	0.7898	0.9532	0.0	0.7606	0.8526	0.6613	0.6499	0.9279	38
0.2835	0.2986	0.6391	0.8598	0.8906	nan	0.8870	0.9343	0.7945	0.7351	0.9481	0.0	0.7588	0.8776	0.6573	0.6247	0.9160	39
0.2746	0.2812	0.6401	0.8652	0.8895	nan	0.8654	0.9247	0.7898	0.7820	0.9642	0.0	0.7451	0.8659	0.6704	0.6198	0.9394	40
0.2866	0.3194	0.6506	0.8679	0.8943	nan	0.8787	0.9420	0.8137	0.7722	0.9330	0.0	0.7775	0.8787	0.6729	0.6738	0.9005	41
0.2641	0.3372	0.6415	0.8745	0.8794	nan	0.8758	0.9017	0.7996	0.8507	0.9446	0.0	0.7665	0.8619	0.6275	0.6932	0.9001	42
0.2396	0.3112	0.6466	0.8604	0.8889	nan	0.8734	0.9432	0.7511	0.7938	0.9403	0.0	0.7722	0.8680	0.6424	0.6795	0.9173	43
0.2566	0.3336	0.6475	0.8762	0.8850	nan	0.8915	0.9044	0.7995	0.8637	0.9222	0.0	0.7743	0.8640	0.6572	0.6878	0.9016	44
0.2612	0.3152	0.6522	0.8776	0.8978	nan	0.8580	0.9306	0.8161	0.8183	0.9651	0.0	0.7516	0.8810	0.6798	0.6604	0.9404	45
0.2442	0.2982	0.6471	0.8749	0.8923	nan	0.8858	0.9214	0.8028	0.8284	0.9363	0.0	0.7642	0.8778	0.6724	0.6602	0.9078	46
0.2418	0.3066	0.6460	0.8689	0.8871	nan	0.8999	0.9207	0.7802	0.8186	0.9248	0.0	0.7754	0.8710	0.6689	0.6587	0.9020	47
0.2410	0.2876	0.6493	0.8697	0.8940	nan	0.8915	0.9324	0.7913	0.7950	0.9382	0.0	0.7792	0.8757	0.6665	0.6756	0.8986	48
0.2414	0.2879	0.6578	0.8810	0.9074	nan	0.9097	0.9396	0.8000	0.8165	0.9392	0.0	0.7772	0.9028	0.6828	0.6714	0.9125	49

Framework versions

Transformers 4.37.2
TensorFlow 2.11.0
Datasets 3.3.1
Tokenizers 0.15.2

Type: segformer
License: other
Created: Feb 24, 2025
Updated: Feb 24, 2025