kelvinandreas/vit-traffic-sign-GTSRB

Model Details

This model uses the Vision Transformer (ViT) architecture for classifying traffic signs from the German Traffic Sign Recognition Benchmark (GTSRB). It provides a robust solution for image classification tasks, specifically aimed at recognizing traffic signs across 43 different classes.

Model Description

• Developed by: Kelvin Andreas
• Model type: Vision Transformer (ViT)
• Finetuned from model: google/vit-base-patch16-224-in21k
• Repository: https://huggingface.co/kelvinandreas/vit-traffic-sign-GTSRB
• Demo: https://huggingface.co/spaces/kelvinandreas/traffic-sign-classification

How to Get Started with the Model

To use the model, follow these steps:

1. Install the required dependencies:

   pip install transformers torch
   

2. Load the model and processor:

   from transformers import ViTForImageClassification, ViTImageProcessor
   import torch
   from PIL import Image
   
   processor = ViTImageProcessor.from_pretrained("kelvinandreas/vit-traffic-sign-GTSRB")
   model = ViTForImageClassification.from_pretrained("kelvinandreas/vit-traffic-sign-GTSRB")
   
   # Load and process image
   image = Image.open("path_to_image.jpg")
   inputs = processor(images=image, return_tensors="pt")
   
   # Make prediction
   outputs = model(**inputs)
   logits = outputs.logits
   predicted_class_idx = torch.argmax(logits, dim=-1)
   print(predicted_class_idx)
   

Results

The model performance on the GTSRB dataset is as follows:

• Accuracy: 0.9846
  
• Precision: 0.9853
  
• Recall: 0.9846
  
• F1 Score: 0.9846