segformer-b1-miic-tl

This model is a fine-tuned version of nvidia/mit-b1 on the yijisuk/ic-chip-sample dataset. It achieves the following results on the evaluation set:

• Loss: 0.2212
• Mean Iou: 0.4723
• Mean Accuracy: 0.9446
• Overall Accuracy: 0.9446
• Accuracy Unlabeled: nan
• Accuracy Circuit: 0.9446
• Iou Unlabeled: 0.0
• Iou Circuit: 0.9446
• Dice Coefficient: 0.8541

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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Unlabeled	Accuracy Circuit	Iou Unlabeled	Iou Circuit	Dice Coefficient
0.3419	3.12	250	0.2745	0.4850	0.9701	0.9701	nan	0.9701	0.0	0.9701	0.8149
0.2785	6.25	500	0.2789	0.4828	0.9657	0.9657	nan	0.9657	0.0	0.9657	0.8285
0.2549	9.38	750	0.2888	0.4721	0.9443	0.9443	nan	0.9443	0.0	0.9443	0.8372
0.2728	12.5	1000	0.2426	0.4699	0.9397	0.9397	nan	0.9397	0.0	0.9397	0.8424
0.2625	15.62	1250	0.1990	0.4632	0.9264	0.9264	nan	0.9264	0.0	0.9264	0.8520
0.2449	18.75	1500	0.2121	0.4706	0.9412	0.9412	nan	0.9412	0.0	0.9412	0.8508
0.2173	21.88	1750	0.2768	0.4780	0.9559	0.9559	nan	0.9559	0.0	0.9559	0.8485
0.2158	25.0	2000	0.2772	0.4643	0.9287	0.9287	nan	0.9287	0.0	0.9287	0.8383
0.1843	28.12	2250	0.1818	0.4671	0.9343	0.9343	nan	0.9343	0.0	0.9343	0.8685
0.1608	31.25	2500	0.1794	0.4591	0.9182	0.9182	nan	0.9182	0.0	0.9182	0.8618
0.1504	34.38	2750	0.1805	0.4586	0.9172	0.9172	nan	0.9172	0.0	0.9172	0.8647
0.1495	37.5	3000	0.2090	0.4773	0.9545	0.9545	nan	0.9545	0.0	0.9545	0.8595
0.142	40.62	3250	0.2048	0.4750	0.9500	0.9500	nan	0.9500	0.0	0.9500	0.8666
0.1401	43.75	3500	0.2131	0.4756	0.9512	0.9512	nan	0.9512	0.0	0.9512	0.8580
0.1339	46.88	3750	0.2469	0.4773	0.9546	0.9546	nan	0.9546	0.0	0.9546	0.8481
0.1303	50.0	4000	0.2212	0.4723	0.9446	0.9446	nan	0.9446	0.0	0.9446	0.8541

Framework versions

• Transformers 4.36.2
• Pytorch 1.11.0+cu115
• Datasets 2.15.0
• Tokenizers 0.15.0