segcrack9k_conglomerate_segformer_aug

This model is a fine-tuned version of nvidia/mit-b5 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0362
• Mean Iou: 0.3412
• Mean Accuracy: 0.6823
• Overall Accuracy: 0.6823
• Accuracy Background: nan
• Accuracy Crack: 0.6823
• Iou Background: 0.0
• Iou Crack: 0.6823

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: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Crack	Iou Background	Iou Crack
0.0323	0.14	1000	0.0445	0.3573	0.7146	0.7146	nan	0.7146	0.0	0.7146
0.0222	0.27	2000	0.0394	0.3591	0.7181	0.7181	nan	0.7181	0.0	0.7181
0.0335	0.41	3000	0.0404	0.2907	0.5813	0.5813	nan	0.5813	0.0	0.5813
0.013	0.54	4000	0.0384	0.3244	0.6489	0.6489	nan	0.6489	0.0	0.6489
0.0159	0.68	5000	0.0382	0.3088	0.6176	0.6176	nan	0.6176	0.0	0.6176
0.0608	0.81	6000	0.0366	0.3251	0.6502	0.6502	nan	0.6502	0.0	0.6502
0.1738	0.95	7000	0.0362	0.3412	0.6823	0.6823	nan	0.6823	0.0	0.6823

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.3
• Tokenizers 0.13.3