tf_segformer_for_semantic_segmentation_v4

This model is a fine-tuned version of ArtemShyshko/tf_segformer_for_semantic_segmentation_v3 on an unknown dataset. It achieves the following results on the evaluation set:

• Train Loss: 0.1650
• Validation Loss: 0.2153
• Validation Mean Iou: 0.6876
• Validation Mean Accuracy: 0.9048
• Validation Overall Accuracy: 0.9199
• Validation Accuracy Unlabeled: nan
• Validation Accuracy Building: 0.9185
• Validation Accuracy Land: 0.9449
• Validation Accuracy Road: 0.8327
• Validation Accuracy Vegetation: 0.8659
• Validation Accuracy Water: 0.9617
• Validation Iou Unlabeled: 0.0
• Validation Iou Building: 0.8219
• Validation Iou Land: 0.9131
• Validation Iou Road: 0.7215
• Validation Iou Vegetation: 0.7242
• Validation Iou Water: 0.9446
• 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
0.2467	0.2755	0.6468	0.8761	0.8850	nan	0.9232	0.9014	0.8291	0.7739	0.9529	0.0	0.7716	0.8658	0.6733	0.6420	0.9280	0
0.2367	0.2769	0.6553	0.8753	0.8970	nan	0.9238	0.9310	0.7846	0.7702	0.9671	0.0	0.7730	0.8842	0.6751	0.6636	0.9356	1
0.2406	0.3519	0.6367	0.8676	0.8789	nan	0.8361	0.9081	0.7676	0.9095	0.9169	0.0	0.7527	0.8729	0.6552	0.6379	0.9013	2
0.2348	0.2933	0.6570	0.8832	0.9020	nan	0.8996	0.9290	0.8104	0.8389	0.9380	0.0	0.7799	0.8954	0.6720	0.6820	0.9125	3
0.2336	0.2938	0.6564	0.8851	0.8900	nan	0.9009	0.9052	0.8029	0.8619	0.9546	0.0	0.7787	0.8662	0.6791	0.6865	0.9279	4
0.2376	0.2802	0.6517	0.8767	0.8889	nan	0.8726	0.9189	0.7901	0.8472	0.9548	0.0	0.7803	0.8701	0.6728	0.6518	0.9351	5
0.2367	0.2769	0.6584	0.8787	0.9003	nan	0.9103	0.9317	0.8161	0.8120	0.9234	0.0	0.7769	0.8935	0.6936	0.6841	0.9023	6
0.2507	0.2647	0.6606	0.8887	0.8902	nan	0.9096	0.8991	0.8177	0.8655	0.9517	0.0	0.7848	0.8550	0.6809	0.7212	0.9216	7
0.2220	0.2647	0.6520	0.8755	0.8888	nan	0.9019	0.9207	0.8034	0.8054	0.9459	0.0	0.7771	0.8728	0.6830	0.6600	0.9194	8
0.2384	0.2602	0.6627	0.8821	0.8941	nan	0.9051	0.9254	0.8178	0.7913	0.9711	0.0	0.7900	0.8708	0.6862	0.6855	0.9437	9
0.2066	0.2682	0.6584	0.8797	0.8947	nan	0.8881	0.9245	0.7584	0.8780	0.9495	0.0	0.7857	0.8756	0.6560	0.7086	0.9246	10
0.2247	0.2975	0.6574	0.8844	0.8961	nan	0.8884	0.9223	0.7936	0.8745	0.9432	0.0	0.7915	0.8862	0.6822	0.6724	0.9123	11
0.2261	0.2969	0.6598	0.8815	0.9060	nan	0.9053	0.9372	0.7996	0.8116	0.9539	0.0	0.7777	0.8957	0.6733	0.6864	0.9259	12
0.2089	0.3076	0.6623	0.8983	0.9017	nan	0.9136	0.9111	0.8503	0.8682	0.9484	0.0	0.7756	0.8870	0.6867	0.7028	0.9220	13
0.2105	0.2626	0.6672	0.8844	0.9068	nan	0.8999	0.9426	0.7893	0.8336	0.9565	0.0	0.7941	0.8967	0.6793	0.6950	0.9383	14
0.2076	0.2772	0.6623	0.8839	0.8983	nan	0.9070	0.9260	0.7668	0.8572	0.9624	0.0	0.7831	0.8851	0.6612	0.7057	0.9388	15
0.2083	0.2664	0.6675	0.8814	0.9068	nan	0.8879	0.9491	0.8269	0.7909	0.9520	0.0	0.7876	0.8973	0.7071	0.6836	0.9295	16
0.2248	0.2665	0.6701	0.8895	0.9022	nan	0.9080	0.9308	0.8070	0.8489	0.9527	0.0	0.7985	0.8846	0.6915	0.7227	0.9233	17
0.2047	0.2568	0.6634	0.8846	0.9119	nan	0.9280	0.9419	0.7913	0.8303	0.9316	0.0	0.7889	0.9093	0.6815	0.6865	0.9142	18
0.2068	0.2512	0.6711	0.8978	0.8992	nan	0.9060	0.9058	0.8462	0.8982	0.9330	0.0	0.7949	0.8736	0.6881	0.7596	0.9101	19
0.1991	0.2676	0.6721	0.8920	0.9029	nan	0.9295	0.9232	0.7967	0.8581	0.9524	0.0	0.7934	0.8871	0.6849	0.7333	0.9341	20
0.2191	0.2659	0.6677	0.8894	0.9005	nan	0.9404	0.9180	0.8045	0.8251	0.9591	0.0	0.8064	0.8839	0.6820	0.6952	0.9384	21
0.1970	0.2853	0.6735	0.8995	0.9099	nan	0.9171	0.9268	0.8364	0.8649	0.9520	0.0	0.8005	0.9006	0.7030	0.7028	0.9338	22
0.1915	0.2762	0.6668	0.8788	0.9035	nan	0.9152	0.9446	0.8138	0.7688	0.9519	0.0	0.7978	0.8899	0.7077	0.6770	0.9286	23
0.2052	0.2811	0.6643	0.8813	0.9016	nan	0.8935	0.9396	0.7971	0.8198	0.9563	0.0	0.7984	0.8828	0.6914	0.6915	0.9218	24
0.1869	0.2700	0.6673	0.8848	0.9033	nan	0.9119	0.9345	0.7530	0.8789	0.9458	0.0	0.7954	0.8913	0.6736	0.7233	0.9201	25
0.1792	0.2457	0.6718	0.8948	0.9091	nan	0.9382	0.9292	0.8370	0.8277	0.9422	0.0	0.7998	0.9052	0.6937	0.7086	0.9236	26
0.1800	0.2548	0.6763	0.8980	0.9066	nan	0.9288	0.9219	0.8142	0.8690	0.9562	0.0	0.8098	0.8899	0.6983	0.7221	0.9378	27
0.1817	0.2376	0.6722	0.8916	0.9119	nan	0.9263	0.9397	0.8103	0.8333	0.9483	0.0	0.8046	0.9095	0.6872	0.7042	0.9275	28
0.1819	0.2529	0.6830	0.9084	0.9146	nan	0.9167	0.9243	0.8381	0.8899	0.9731	0.0	0.8121	0.9013	0.6987	0.7376	0.9484	29
0.1886	0.2466	0.6713	0.8867	0.9029	nan	0.9136	0.9348	0.7895	0.8383	0.9571	0.0	0.7964	0.8824	0.6828	0.7288	0.9374	30
0.2018	0.2568	0.6678	0.8849	0.9069	nan	0.9198	0.9402	0.8335	0.7745	0.9564	0.0	0.7950	0.8920	0.7093	0.6994	0.9114	31
0.1954	0.2296	0.6734	0.8940	0.9163	nan	0.8895	0.9466	0.8254	0.8546	0.9538	0.0	0.7993	0.9096	0.6943	0.7081	0.9292	32
0.1849	0.2215	0.6684	0.8923	0.9023	nan	0.9052	0.9234	0.7964	0.8880	0.9486	0.0	0.8043	0.8871	0.6884	0.7043	0.9263	33
0.1941	0.2405	0.6742	0.8964	0.9164	nan	0.9040	0.9440	0.8431	0.8437	0.9471	0.0	0.7987	0.9139	0.7265	0.6853	0.9210	34
0.1881	0.2605	0.6642	0.8888	0.9023	nan	0.8699	0.9318	0.8009	0.8862	0.9553	0.0	0.7893	0.8879	0.6939	0.6788	0.9353	35
0.1943	0.2699	0.6777	0.8968	0.9079	nan	0.9107	0.9318	0.8230	0.8593	0.9592	0.0	0.8178	0.8914	0.7058	0.7155	0.9358	36
0.1872	0.2382	0.6859	0.9060	0.9130	nan	0.9350	0.9265	0.8462	0.8647	0.9575	0.0	0.8186	0.8951	0.7148	0.7501	0.9367	37
0.1887	0.2526	0.6795	0.8990	0.9080	nan	0.9147	0.9278	0.8445	0.8660	0.9418	0.0	0.8114	0.8844	0.7228	0.7462	0.9125	38
0.1865	0.2354	0.6789	0.8966	0.9103	nan	0.9246	0.9360	0.8430	0.8216	0.9576	0.0	0.8185	0.8953	0.7082	0.7200	0.9316	39
0.1828	0.2307	0.6804	0.8986	0.9101	nan	0.8899	0.9370	0.8666	0.8531	0.9464	0.0	0.8115	0.8902	0.7156	0.7484	0.9166	40
0.1804	0.2335	0.6764	0.8951	0.9114	nan	0.9012	0.9405	0.8371	0.8464	0.9507	0.0	0.8047	0.8982	0.7301	0.7010	0.9245	41
0.1795	0.2624	0.6826	0.8998	0.9151	nan	0.9294	0.9372	0.8116	0.8699	0.9510	0.0	0.8257	0.9003	0.7014	0.7381	0.9304	42
0.1777	0.2251	0.6863	0.9010	0.9164	nan	0.9130	0.9427	0.8211	0.8754	0.9528	0.0	0.8200	0.9004	0.7167	0.7506	0.9301	43
0.1750	0.2294	0.6769	0.8958	0.9019	nan	0.9304	0.9174	0.8202	0.8608	0.9502	0.0	0.8213	0.8742	0.7147	0.7251	0.9258	44
0.1669	0.2141	0.6816	0.9014	0.9107	nan	0.9209	0.9305	0.8525	0.8350	0.9681	0.0	0.8216	0.8964	0.7116	0.7166	0.9432	45
0.1622	0.2484	0.6736	0.8954	0.9093	nan	0.9411	0.9309	0.8257	0.8152	0.9639	0.0	0.8113	0.8988	0.7015	0.6999	0.9302	46
0.1859	0.2145	0.6757	0.8942	0.9134	nan	0.9466	0.9325	0.8332	0.7897	0.9689	0.0	0.8232	0.9005	0.7140	0.6672	0.9491	47
0.1693	0.2243	0.6817	0.9000	0.9120	nan	0.9198	0.9342	0.8038	0.8742	0.9682	0.0	0.8232	0.8962	0.6932	0.7314	0.9461	48
0.1650	0.2153	0.6876	0.9048	0.9199	nan	0.9185	0.9449	0.8327	0.8659	0.9617	0.0	0.8219	0.9131	0.7215	0.7242	0.9446	49

Framework versions

• Transformers 4.37.2
• TensorFlow 2.11.0
• Datasets 3.3.1
• Tokenizers 0.15.2