Mozilla/docornot

This model is a fine-tuned version of facebook/deit-tiny-distilled-patch16-224 on the docornot dataset.

It achieves the following results on the evaluation set:

• Loss: 0.0000
• Accuracy: 1.0

CO2 emissions

This model was trained on an M1 and took 0.322 g of CO2 (measured with CodeCarbon)

Model description

This model is distilled Vision Transformer (ViT) model. Images are presented to the model as a sequence of fixed-size patches (resolution 16x16), which are linearly embedded.

Intended uses & limitations

You can use this model to detect if an image is a picture or a document.

Training procedure

Source code used to generate this model : https://github.com/mozilla/docornot

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• 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	Accuracy
0.0	1.0	1600	0.0000	1.0

Framework versions

• Transformers 4.39.2
• Pytorch 2.2.2
• Datasets 2.18.0
• Tokenizers 0.15.2