segformer-b0-finetuned-breastcancer-oct-1

This model is a fine-tuned version of nvidia/mit-b0 on the as-cle-bert/breastcancer-semantic-segmentation dataset. It achieves the following results on the evaluation set:

• Loss: 0.2413
• Mean Iou: 0.4091
• Mean Accuracy: 0.5395
• Overall Accuracy: 0.9467
• Accuracy Ignore: 0.1246
• Accuracy Benign Breast Cancer: 0.5579
• Accuracy Malignant Breast Cancer: 0.4873
• Accuracy Background: 0.9882
• Iou Ignore: 0.1150
• Iou Benign Breast Cancer: 0.1215
• Iou Malignant Breast Cancer: 0.4523
• Iou Background: 0.9478

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Ignore	Accuracy Benign Breast Cancer	Accuracy Malignant Breast Cancer	Accuracy Background	Iou Ignore	Iou Benign Breast Cancer	Iou Malignant Breast Cancer	Iou Background
0.074	0.625	10	0.2831	0.3265	0.5114	0.9244	0.0	0.7065	0.3650	0.9742	0.0	0.0562	0.3140	0.9359
0.0747	1.25	20	0.2687	0.3706	0.4979	0.9394	0.0	0.5673	0.4396	0.9846	0.0	0.1507	0.3934	0.9382
0.1038	1.875	30	0.2365	0.3962	0.5469	0.9464	0.0206	0.6556	0.5268	0.9845	0.0205	0.1487	0.4688	0.9469
0.0884	2.5	40	0.2424	0.3985	0.5438	0.9450	0.0784	0.6292	0.4807	0.9869	0.0752	0.1318	0.4410	0.9461
0.0591	3.125	50	0.2473	0.3919	0.5478	0.9384	0.1061	0.6262	0.4793	0.9797	0.1006	0.1115	0.4161	0.9393
0.0566	3.75	60	0.2776	0.3640	0.4351	0.9388	0.1037	0.2770	0.3692	0.9904	0.0984	0.0697	0.3485	0.9393
0.063	4.375	70	0.2262	0.4058	0.5094	0.9472	0.1072	0.4018	0.5440	0.9844	0.1015	0.0892	0.4848	0.9478
0.0358	5.0	80	0.2398	0.4065	0.5087	0.9464	0.1084	0.4191	0.5220	0.9854	0.1026	0.1078	0.4699	0.9457
0.0319	5.625	90	0.2659	0.3844	0.4737	0.9423	0.1103	0.3823	0.4122	0.9902	0.1034	0.1010	0.3917	0.9416
0.0499	6.25	100	0.2393	0.4030	0.5532	0.9423	0.1065	0.6272	0.4963	0.9826	0.1006	0.1285	0.4407	0.9423
0.0394	6.875	110	0.2415	0.4024	0.5233	0.9463	0.1086	0.5205	0.4751	0.9890	0.1026	0.1137	0.4464	0.9470
0.037	7.5	120	0.2475	0.3916	0.4752	0.9458	0.1124	0.3507	0.4465	0.9912	0.1053	0.0904	0.4245	0.9462
0.0588	8.125	130	0.2458	0.4041	0.5223	0.9455	0.1276	0.5146	0.4577	0.9895	0.1171	0.1240	0.4292	0.9460
0.0426	8.75	140	0.2463	0.4046	0.5264	0.9459	0.1225	0.5322	0.4614	0.9897	0.1134	0.1252	0.4333	0.9467
0.0848	9.375	150	0.2388	0.4078	0.5297	0.9469	0.1154	0.5298	0.4849	0.9888	0.1078	0.1253	0.4506	0.9475
0.0574	10.0	160	0.2413	0.4091	0.5395	0.9467	0.1246	0.5579	0.4873	0.9882	0.1150	0.1215	0.4523	0.9478

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1