README.md

metadata

license: other
tags:
  - generated_from_trainer
model-index:
  - name: segformer-b0-DeepCrack
    results: []

segformer-b0-DeepCrack

This model is a fine-tuned version of nvidia/mit-b4 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0121
• Mean Iou: 0.0096
• Mean Accuracy: 0.0192
• Overall Accuracy: 0.0192
• Accuracy Background: nan
• Accuracy Cracked: 0.0192
• Iou Background: 0.0
• Iou Cracked: 0.0192

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 Cracked	Iou Background	Iou Cracked
0.2068	0.13	20	0.1942	0.0041	0.0082	0.0082	nan	0.0082	0.0	0.0082
0.0778	0.27	40	0.0679	0.0306	0.0612	0.0612	nan	0.0612	0.0	0.0612
0.0325	0.4	60	0.0328	0.0167	0.0334	0.0334	nan	0.0334	0.0	0.0334
0.0203	0.53	80	0.0191	0.0074	0.0148	0.0148	nan	0.0148	0.0	0.0148
0.015	0.67	100	0.0152	0.0144	0.0288	0.0288	nan	0.0288	0.0	0.0288
0.0139	0.8	120	0.0130	0.0074	0.0148	0.0148	nan	0.0148	0.0	0.0148
0.016	0.93	140	0.0121	0.0096	0.0192	0.0192	nan	0.0192	0.0	0.0192

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu118
• Datasets 2.13.1
• Tokenizers 0.13.3