Augmented-MIT-b5

This model is a fine-tuned version of nvidia/mit-b5 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0371
• Mean Iou: 0.3355
• Mean Accuracy: 0.6711
• Overall Accuracy: 0.6711
• Accuracy Background: nan
• Accuracy Crack: 0.6711
• Iou Background: 0.0
• Iou Crack: 0.6711

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
0.0365	0.14	1000	0.0446	0.3813	0.7627	0.7627	nan	0.7627	0.0	0.7627
0.0114	0.27	2000	0.0411	0.3691	0.7381	0.7381	nan	0.7381	0.0	0.7381
0.0148	0.41	3000	0.0400	0.3224	0.6448	0.6448	nan	0.6448	0.0	0.6448
0.0134	0.54	4000	0.0413	0.2819	0.5638	0.5638	nan	0.5638	0.0	0.5638
0.013	0.68	5000	0.0392	0.3618	0.7235	0.7235	nan	0.7235	0.0	0.7235
0.0532	0.81	6000	0.0373	0.3355	0.6710	0.6710	nan	0.6710	0.0	0.6710
0.0508	0.95	7000	0.0371	0.3355	0.6711	0.6711	nan	0.6711	0.0	0.6711

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3