segformer-b0-finetuned-segments-docboundary-nov-13

This model is a fine-tuned version of nvidia/mit-b0 on the vigneshgs7/doc-boundary dataset. It achieves the following results on the evaluation set:

• Loss: 0.2806
• Mean Iou: 0.4886
• Mean Accuracy: 0.9771
• Overall Accuracy: 0.9771
• Accuracy Page: nan
• Accuracy Surface: 0.9771
• Iou Page: 0.0
• Iou Surface: 0.9771

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 Page	Accuracy Surface	Iou Page	Iou Surface
0.4191	2.22	20	0.5643	0.4651	0.9301	0.9301	nan	0.9301	0.0	0.9301
0.3141	4.44	40	0.3959	0.4866	0.9733	0.9733	nan	0.9733	0.0	0.9733
0.2865	6.67	60	0.2889	0.4870	0.9740	0.9740	nan	0.9740	0.0	0.9740
0.3955	8.89	80	0.2806	0.4886	0.9771	0.9771	nan	0.9771	0.0	0.9771

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0
• Datasets 2.14.6
• Tokenizers 0.14.1