Vit Snacks

🚀 vit-snacks

This is a fine - tuned image classification model based on the ViT architecture, trained on the snacks dataset to classify different types of snacks.

🚀 Quick Start

This model is a fine - tuned version of [google/vit - base - patch16 - 224 - in21k](https://huggingface.co/google/vit - base - patch16 - 224 - in21k) on the Matthijs/snacks dataset. It achieves the following results on the evaluation set:

• Loss: 0.2754
• Accuracy: 0.9393

✨ Features

• Image Classification: Specialized in classifying different varieties of snacks.

📚 Documentation

Model description

Upload any image of your fave yummy snack.

Intended uses & limitations

There are only 20 different varieties of snacks.

Training and evaluation data

More information needed.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Framework versions

• Transformers 4.20.1
• Pytorch 1.11.0+cu113
• Datasets 2.3.2
• Tokenizers 0.12.1

📄 License

This project is licensed under the Apache - 2.0 license.