segformer-b0-finetuned-batch3-26May

This model is a fine-tuned version of PushkarA07/segformer-b0-finetuned-batch2w5-15Dec on the PushkarA07/batch3-tiles_second dataset. It achieves the following results on the evaluation set:

• Loss: 0.0014
• Mean Iou: 0.8838
• Mean Accuracy: 0.9271
• Overall Accuracy: 0.9994
• Accuracy Abnormality: 0.8545
• Iou Abnormality: 0.7682

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: 8
• eval_batch_size: 8
• 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: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Abnormality	Iou Abnormality
0.0035	0.7143	10	0.0024	0.8344	0.8791	0.9992	0.7585	0.6697
0.0021	1.4286	20	0.0022	0.8422	0.8942	0.9992	0.7888	0.6852
0.0033	2.1429	30	0.0020	0.8474	0.8942	0.9992	0.7887	0.6956
0.007	2.8571	40	0.0020	0.8510	0.8943	0.9992	0.7889	0.7028
0.0036	3.5714	50	0.0019	0.8553	0.8983	0.9993	0.7968	0.7113
0.0032	4.2857	60	0.0018	0.8583	0.8969	0.9993	0.7940	0.7173
0.0026	5.0	70	0.0018	0.8594	0.9003	0.9993	0.8009	0.7195
0.0033	5.7143	80	0.0018	0.8600	0.8999	0.9993	0.8000	0.7207
0.0048	6.4286	90	0.0018	0.8616	0.8997	0.9993	0.7997	0.7239
0.003	7.1429	100	0.0017	0.8653	0.9105	0.9993	0.8214	0.7313
0.0026	7.8571	110	0.0017	0.8664	0.9122	0.9993	0.8248	0.7335
0.0022	8.5714	120	0.0017	0.8599	0.8902	0.9993	0.7806	0.7206
0.0023	9.2857	130	0.0017	0.8668	0.9069	0.9993	0.8140	0.7342
0.0028	10.0	140	0.0017	0.8681	0.9177	0.9993	0.8357	0.7368
0.0017	10.7143	150	0.0017	0.8672	0.9107	0.9993	0.8217	0.7350
0.002	11.4286	160	0.0017	0.8665	0.9020	0.9993	0.8041	0.7336
0.0018	12.1429	170	0.0017	0.8676	0.9047	0.9993	0.8097	0.7358
0.0021	12.8571	180	0.0017	0.8695	0.9240	0.9993	0.8484	0.7397
0.001	13.5714	190	0.0016	0.8700	0.9145	0.9993	0.8293	0.7407
0.0014	14.2857	200	0.0016	0.8721	0.9123	0.9994	0.8248	0.7449
0.0016	15.0	210	0.0016	0.8704	0.9082	0.9994	0.8166	0.7414
0.0023	15.7143	220	0.0017	0.8709	0.9175	0.9993	0.8352	0.7425
0.0023	16.4286	230	0.0016	0.8732	0.9188	0.9994	0.8379	0.7470
0.0019	17.1429	240	0.0016	0.8731	0.9153	0.9994	0.8308	0.7468
0.0018	17.8571	250	0.0016	0.8726	0.9094	0.9994	0.8191	0.7459
0.0027	18.5714	260	0.0016	0.8697	0.9016	0.9994	0.8033	0.7400
0.0013	19.2857	270	0.0016	0.8758	0.9214	0.9994	0.8431	0.7523
0.0036	20.0	280	0.0016	0.8750	0.9226	0.9994	0.8456	0.7506
0.0025	20.7143	290	0.0016	0.8751	0.9293	0.9994	0.8589	0.7509
0.0016	21.4286	300	0.0016	0.8725	0.9095	0.9994	0.8192	0.7457
0.0032	22.1429	310	0.0016	0.8737	0.9102	0.9994	0.8206	0.7481
0.002	22.8571	320	0.0016	0.8772	0.9304	0.9994	0.8610	0.7550
0.0012	23.5714	330	0.0016	0.8760	0.9151	0.9994	0.8304	0.7526
0.0012	24.2857	340	0.0016	0.8767	0.9226	0.9994	0.8454	0.7541
0.0028	25.0	350	0.0015	0.8771	0.9205	0.9994	0.8413	0.7548
0.0021	25.7143	360	0.0015	0.8769	0.9176	0.9994	0.8354	0.7543
0.0016	26.4286	370	0.0015	0.8763	0.9156	0.9994	0.8315	0.7533
0.0025	27.1429	380	0.0016	0.8742	0.9171	0.9994	0.8345	0.7490
0.0029	27.8571	390	0.0016	0.8763	0.9322	0.9994	0.8647	0.7532
0.0025	28.5714	400	0.0015	0.8767	0.9194	0.9994	0.8391	0.7539
0.0022	29.2857	410	0.0015	0.8783	0.9205	0.9994	0.8413	0.7572
0.0014	30.0	420	0.0016	0.8792	0.9318	0.9994	0.8639	0.7590
0.0027	30.7143	430	0.0015	0.8786	0.9269	0.9994	0.8541	0.7578
0.0038	31.4286	440	0.0015	0.8787	0.9270	0.9994	0.8544	0.7581
0.0014	32.1429	450	0.0015	0.8781	0.9243	0.9994	0.8490	0.7569
0.0015	32.8571	460	0.0015	0.8781	0.9214	0.9994	0.8431	0.7568
0.0034	33.5714	470	0.0015	0.8769	0.9136	0.9994	0.8275	0.7544
0.0048	34.2857	480	0.0015	0.8783	0.9310	0.9994	0.8623	0.7573
0.0025	35.0	490	0.0015	0.8783	0.9210	0.9994	0.8422	0.7572
0.0029	35.7143	500	0.0015	0.8788	0.9234	0.9994	0.8470	0.7582
0.0024	36.4286	510	0.0015	0.8797	0.9286	0.9994	0.8576	0.7600
0.0013	37.1429	520	0.0015	0.8792	0.9197	0.9994	0.8396	0.7589
0.0023	37.8571	530	0.0015	0.8797	0.9240	0.9994	0.8484	0.7601
0.0017	38.5714	540	0.0015	0.8802	0.9269	0.9994	0.8541	0.7610
0.0023	39.2857	550	0.0015	0.8801	0.9248	0.9994	0.8498	0.7609
0.0027	40.0	560	0.0015	0.8806	0.9297	0.9994	0.8596	0.7618
0.0029	40.7143	570	0.0015	0.8798	0.9209	0.9994	0.8420	0.7602
0.0035	41.4286	580	0.0015	0.8776	0.9129	0.9994	0.8261	0.7557
0.0025	42.1429	590	0.0015	0.8792	0.9197	0.9994	0.8397	0.7590
0.0011	42.8571	600	0.0015	0.8813	0.9312	0.9994	0.8628	0.7632
0.0022	43.5714	610	0.0015	0.8803	0.9254	0.9994	0.8511	0.7612
0.0029	44.2857	620	0.0015	0.8796	0.9199	0.9994	0.8400	0.7597
0.0017	45.0	630	0.0015	0.8808	0.9254	0.9994	0.8511	0.7621
0.0013	45.7143	640	0.0015	0.8815	0.9276	0.9994	0.8554	0.7635
0.0026	46.4286	650	0.0015	0.8798	0.9258	0.9994	0.8518	0.7602
0.0018	47.1429	660	0.0015	0.8803	0.9307	0.9994	0.8616	0.7613
0.0016	47.8571	670	0.0015	0.8811	0.9272	0.9994	0.8546	0.7627
0.001	48.5714	680	0.0015	0.8796	0.9160	0.9994	0.8321	0.7598
0.002	49.2857	690	0.0015	0.8807	0.9314	0.9994	0.8632	0.7621
0.0021	50.0	700	0.0015	0.8797	0.9235	0.9994	0.8473	0.7600
0.0019	50.7143	710	0.0015	0.8800	0.9229	0.9994	0.8461	0.7606
0.0013	51.4286	720	0.0015	0.8794	0.9212	0.9994	0.8427	0.7593
0.0032	52.1429	730	0.0015	0.8806	0.9229	0.9994	0.8461	0.7618
0.0015	52.8571	740	0.0015	0.8813	0.9268	0.9994	0.8540	0.7632
0.0027	53.5714	750	0.0015	0.8807	0.9235	0.9994	0.8472	0.7620
0.0018	54.2857	760	0.0015	0.8806	0.9209	0.9994	0.8420	0.7618
0.0028	55.0	770	0.0015	0.8817	0.9245	0.9994	0.8493	0.7639
0.0019	55.7143	780	0.0015	0.8797	0.9154	0.9994	0.8311	0.7601
0.0017	56.4286	790	0.0015	0.8815	0.9238	0.9994	0.8479	0.7636
0.001	57.1429	800	0.0015	0.8811	0.9227	0.9994	0.8456	0.7628
0.0022	57.8571	810	0.0015	0.8827	0.9303	0.9994	0.8610	0.7660
0.0012	58.5714	820	0.0015	0.8825	0.9237	0.9994	0.8475	0.7656
0.0018	59.2857	830	0.0015	0.8831	0.9268	0.9994	0.8538	0.7667
0.0022	60.0	840	0.0015	0.8821	0.9263	0.9994	0.8530	0.7649
0.0021	60.7143	850	0.0015	0.8821	0.9246	0.9994	0.8494	0.7647
0.0033	61.4286	860	0.0015	0.8818	0.9277	0.9994	0.8556	0.7642
0.0034	62.1429	870	0.0015	0.8817	0.9231	0.9994	0.8465	0.7640
0.0037	62.8571	880	0.0015	0.8821	0.9261	0.9994	0.8524	0.7647
0.0018	63.5714	890	0.0015	0.8831	0.9334	0.9994	0.8671	0.7668
0.0018	64.2857	900	0.0015	0.8830	0.9316	0.9994	0.8635	0.7667
0.0035	65.0	910	0.0015	0.8806	0.9153	0.9994	0.8309	0.7618
0.0018	65.7143	920	0.0015	0.8814	0.9312	0.9994	0.8627	0.7635
0.0015	66.4286	930	0.0015	0.8826	0.9264	0.9994	0.8531	0.7657
0.0016	67.1429	940	0.0015	0.8836	0.9358	0.9994	0.8719	0.7677
0.0023	67.8571	950	0.0015	0.8820	0.9215	0.9994	0.8433	0.7645
0.0015	68.5714	960	0.0015	0.8816	0.9283	0.9994	0.8569	0.7639
0.0023	69.2857	970	0.0015	0.8833	0.9302	0.9994	0.8607	0.7673
0.0036	70.0	980	0.0015	0.8824	0.9256	0.9994	0.8515	0.7654
0.0011	70.7143	990	0.0015	0.8816	0.9268	0.9994	0.8538	0.7637
0.0034	71.4286	1000	0.0015	0.8818	0.9267	0.9994	0.8536	0.7643
0.0014	72.1429	1010	0.0015	0.8833	0.9303	0.9994	0.8609	0.7672
0.0011	72.8571	1020	0.0015	0.8827	0.9287	0.9994	0.8577	0.7659
0.0014	73.5714	1030	0.0015	0.8819	0.9257	0.9994	0.8516	0.7643
0.001	74.2857	1040	0.0015	0.8829	0.9294	0.9994	0.8591	0.7665
0.0026	75.0	1050	0.0015	0.8812	0.9210	0.9994	0.8423	0.7629
0.0014	75.7143	1060	0.0015	0.8823	0.9288	0.9994	0.8579	0.7653
0.0029	76.4286	1070	0.0015	0.8825	0.9244	0.9994	0.8490	0.7656
0.0007	77.1429	1080	0.0015	0.8828	0.9268	0.9994	0.8538	0.7662
0.0021	77.8571	1090	0.0015	0.8829	0.9257	0.9994	0.8517	0.7664
0.002	78.5714	1100	0.0015	0.8835	0.9270	0.9994	0.8542	0.7675
0.0025	79.2857	1110	0.0015	0.8833	0.9276	0.9994	0.8554	0.7673
0.0026	80.0	1120	0.0015	0.8831	0.9255	0.9994	0.8513	0.7669
0.0035	80.7143	1130	0.0015	0.8841	0.9306	0.9994	0.8615	0.7689
0.0016	81.4286	1140	0.0015	0.8833	0.9256	0.9994	0.8515	0.7672
0.0018	82.1429	1150	0.0015	0.8828	0.9266	0.9994	0.8535	0.7661
0.0024	82.8571	1160	0.0015	0.8831	0.9280	0.9994	0.8563	0.7668
0.0022	83.5714	1170	0.0015	0.8836	0.9309	0.9994	0.8620	0.7677
0.0018	84.2857	1180	0.0015	0.8835	0.9303	0.9994	0.8608	0.7676
0.0014	85.0	1190	0.0015	0.8832	0.9266	0.9994	0.8535	0.7669
0.0013	85.7143	1200	0.0015	0.8838	0.9273	0.9994	0.8548	0.7682
0.0033	86.4286	1210	0.0014	0.8836	0.9316	0.9994	0.8635	0.7678
0.0023	87.1429	1220	0.0015	0.8831	0.9231	0.9994	0.8465	0.7667
0.0027	87.8571	1230	0.0014	0.8834	0.9284	0.9994	0.8571	0.7675
0.0014	88.5714	1240	0.0015	0.8833	0.9285	0.9994	0.8572	0.7672
0.0025	89.2857	1250	0.0014	0.8836	0.9276	0.9994	0.8555	0.7678
0.003	90.0	1260	0.0014	0.8842	0.9299	0.9994	0.8600	0.7690
0.0022	90.7143	1270	0.0014	0.8842	0.9271	0.9994	0.8545	0.7690
0.0024	91.4286	1280	0.0014	0.8839	0.9285	0.9994	0.8572	0.7684
0.0017	92.1429	1290	0.0014	0.8835	0.9262	0.9994	0.8526	0.7676
0.0014	92.8571	1300	0.0014	0.8830	0.9243	0.9994	0.8488	0.7666
0.0018	93.5714	1310	0.0014	0.8836	0.9293	0.9994	0.8589	0.7678
0.0019	94.2857	1320	0.0014	0.8833	0.9265	0.9994	0.8532	0.7673
0.0027	95.0	1330	0.0015	0.8831	0.9250	0.9994	0.8503	0.7667
0.0008	95.7143	1340	0.0015	0.8835	0.9275	0.9994	0.8553	0.7677
0.0033	96.4286	1350	0.0015	0.8835	0.9280	0.9994	0.8563	0.7675
0.0021	97.1429	1360	0.0015	0.8834	0.9268	0.9994	0.8540	0.7675
0.0016	97.8571	1370	0.0015	0.8836	0.9271	0.9994	0.8545	0.7677
0.0039	98.5714	1380	0.0014	0.8835	0.9255	0.9994	0.8513	0.7677
0.0017	99.2857	1390	0.0014	0.8836	0.9257	0.9994	0.8517	0.7677
0.0046	100.0	1400	0.0014	0.8838	0.9271	0.9994	0.8545	0.7682

Framework versions

• Transformers 4.52.3
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1