segformer-b0-finetuned-ade-512-512_necrosis

This model is a fine-tuned version of nvidia/segformer-b0-finetuned-ade-512-512 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0608
• Mean Iou: 0.8722
• Mean Accuracy: 0.9242
• Overall Accuracy: 0.9813
• Accuracy Background: 0.9949
• Accuracy Necrosis: 0.8211
• Accuracy Root: 0.9564
• Iou Background: 0.9895
• Iou Necrosis: 0.7138
• Iou Root: 0.9132

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Necrosis	Accuracy Root	Iou Background	Iou Necrosis	Iou Root
0.8601	0.625	20	0.8393	0.5508	0.6494	0.8874	0.9288	0.1647	0.8546	0.9164	0.0706	0.6654
0.6009	1.25	40	0.5086	0.5870	0.6565	0.9438	0.9765	0.0084	0.9847	0.9727	0.0078	0.7805
0.4497	1.875	60	0.3599	0.5953	0.6654	0.9438	0.9750	0.0350	0.9862	0.9698	0.0333	0.7826
0.3922	2.5	80	0.2861	0.6267	0.6957	0.9469	0.9752	0.1275	0.9846	0.9717	0.1183	0.7902
0.2496	3.125	100	0.2361	0.7322	0.7918	0.9622	0.9846	0.4183	0.9726	0.9804	0.3765	0.8398
0.2184	3.75	120	0.1989	0.7824	0.8508	0.9664	0.9840	0.6063	0.9621	0.9805	0.5121	0.8546
0.2193	4.375	140	0.1700	0.8123	0.8765	0.9721	0.9900	0.6864	0.9529	0.9849	0.5768	0.8752
0.1705	5.0	160	0.1500	0.8121	0.8731	0.9723	0.9889	0.6684	0.9621	0.9851	0.5749	0.8763
0.1611	5.625	180	0.1420	0.8381	0.9065	0.9753	0.9942	0.7919	0.9333	0.9863	0.6416	0.8863
0.128	6.25	200	0.1293	0.8420	0.9101	0.9763	0.9938	0.7972	0.9393	0.9873	0.6473	0.8914
0.1368	6.875	220	0.1115	0.8385	0.8990	0.9763	0.9914	0.7476	0.9581	0.9874	0.6362	0.8920
0.1459	7.5	240	0.1074	0.8411	0.8985	0.9771	0.9929	0.7457	0.9568	0.9881	0.6397	0.8955
0.1066	8.125	260	0.1026	0.8505	0.9127	0.9776	0.9947	0.8020	0.9415	0.9877	0.6676	0.8963
0.0973	8.75	280	0.0959	0.8558	0.9189	0.9787	0.9931	0.8118	0.9517	0.9885	0.6769	0.9020
0.1286	9.375	300	0.0883	0.8544	0.9024	0.9792	0.9944	0.7529	0.9598	0.9885	0.6704	0.9043
0.0824	10.0	320	0.0872	0.8614	0.9190	0.9796	0.9934	0.8078	0.9559	0.9887	0.6896	0.9058
0.083	10.625	340	0.0868	0.8641	0.9205	0.9796	0.9955	0.8207	0.9453	0.9882	0.6990	0.9051
0.0794	11.25	360	0.0816	0.8612	0.9198	0.9796	0.9943	0.8142	0.9510	0.9889	0.6893	0.9054
0.0979	11.875	380	0.0816	0.8575	0.9062	0.9796	0.9929	0.7582	0.9675	0.9888	0.6770	0.9066
0.0734	12.5	400	0.0785	0.8584	0.9033	0.9799	0.9949	0.7537	0.9612	0.9889	0.6790	0.9073
0.108	13.125	420	0.0749	0.8642	0.9161	0.9802	0.9949	0.7983	0.9551	0.9889	0.6954	0.9084
0.0803	13.75	440	0.0758	0.8691	0.9265	0.9804	0.9949	0.8359	0.9488	0.9887	0.7099	0.9086
0.0812	14.375	460	0.0734	0.8683	0.9235	0.9805	0.9949	0.8238	0.9517	0.9889	0.7067	0.9094
0.0715	15.0	480	0.0696	0.8683	0.9239	0.9806	0.9931	0.8180	0.9605	0.9892	0.7054	0.9104
0.0673	15.625	500	0.0675	0.8698	0.9275	0.9808	0.9938	0.8328	0.9560	0.9893	0.7091	0.9109
0.072	16.25	520	0.0696	0.8699	0.9231	0.9809	0.9948	0.8195	0.9550	0.9892	0.7094	0.9112
0.0681	16.875	540	0.0696	0.8696	0.9235	0.9806	0.9955	0.8255	0.9496	0.9889	0.7105	0.9096
0.0641	17.5	560	0.0671	0.8618	0.9063	0.9805	0.9944	0.7587	0.9657	0.9894	0.6860	0.9101
0.0842	18.125	580	0.0681	0.8692	0.9211	0.9808	0.9948	0.8128	0.9558	0.9892	0.7073	0.9111
0.0738	18.75	600	0.0661	0.8693	0.9214	0.9809	0.9942	0.8109	0.9591	0.9893	0.7070	0.9116
0.0629	19.375	620	0.0640	0.8685	0.9177	0.9810	0.9937	0.7946	0.9648	0.9895	0.7037	0.9122
0.064	20.0	640	0.0637	0.8705	0.9238	0.9811	0.9936	0.8162	0.9616	0.9896	0.7093	0.9128
0.0599	20.625	660	0.0638	0.8704	0.9221	0.9811	0.9950	0.8153	0.9561	0.9894	0.7098	0.9121
0.0645	21.25	680	0.0644	0.8715	0.9257	0.9811	0.9939	0.8243	0.9588	0.9894	0.7126	0.9126
0.0843	21.875	700	0.0643	0.8670	0.9131	0.9810	0.9949	0.7827	0.9619	0.9895	0.6995	0.9119
0.0578	22.5	720	0.0629	0.8716	0.9255	0.9809	0.9958	0.8319	0.9486	0.9890	0.7151	0.9107
0.0586	23.125	740	0.0616	0.8681	0.9178	0.9810	0.9937	0.7949	0.9647	0.9896	0.7023	0.9123
0.0678	23.75	760	0.0614	0.8732	0.9318	0.9812	0.9944	0.8481	0.9528	0.9895	0.7176	0.9124
0.0757	24.375	780	0.0627	0.8680	0.9151	0.9811	0.9949	0.7891	0.9613	0.9896	0.7019	0.9125
0.081	25.0	800	0.0621	0.8721	0.9248	0.9813	0.9950	0.8242	0.9553	0.9895	0.7138	0.9129
0.0628	25.625	820	0.0604	0.8718	0.9239	0.9814	0.9941	0.8173	0.9604	0.9896	0.7121	0.9136
0.0515	26.25	840	0.0612	0.8720	0.9233	0.9813	0.9945	0.8162	0.9591	0.9896	0.7131	0.9134
0.0735	26.875	860	0.0605	0.8719	0.9224	0.9813	0.9953	0.8159	0.9559	0.9895	0.7132	0.9131
0.06	27.5	880	0.0610	0.8729	0.9254	0.9814	0.9951	0.8259	0.9551	0.9895	0.7160	0.9133
0.0525	28.125	900	0.0610	0.8716	0.9227	0.9813	0.9946	0.8147	0.9588	0.9896	0.7118	0.9134
0.0738	28.75	920	0.0610	0.8713	0.9217	0.9813	0.9949	0.8120	0.9584	0.9896	0.7111	0.9133
0.0632	29.375	940	0.0606	0.8718	0.9228	0.9813	0.9951	0.8166	0.9566	0.9895	0.7129	0.9131
0.0547	30.0	960	0.0608	0.8722	0.9242	0.9813	0.9949	0.8211	0.9564	0.9895	0.7138	0.9132

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0