README.md

---
license: mit
pipeline_tag: image-classification
---
# CanineNet: Dog Breed Classification for 'Le Refuge'

**Github:** [Here](https://github.com/amaye15/CanineNet/tree/main)

**Huggingface Model:** [Here](https://huggingface.co/amaye15/ViT-Standford-Dogs)

**Streamlit App:** [Here](https://caninenet.streamlit.app/)

**DagsHub & Mlflow Experiement Tracking:** [Here](https://dagshub.com/andrewmayes14/mlflow/experiments/#/)

## Project Overview
As a volunteer at 'Le Refuge,' a local animal protection association, I embarked on a mission to develop a machine learning tool that could aid in the classification of dog breeds from their vast image database. The project was inspired by my own experience finding my beloved pet, Snooky, through this association. To give back, I aimed to streamline their data management process by implementing an advanced breed classification algorithm.

```python
from transformers import pipeline

# Specify the model you want to use
model_name = "amaye15/ViT-Standford-Dogs"  # Example: Vision Transformer

# Load the image classification pipeline with the specified model
image_classifier = pipeline("image-classification", model=model_name)

# Path or URL to your image
image_path_or_url = "path_or_url_to_your_image"

# Get predictions
predictions = image_classifier(image_path_or_url)

# Define the number of top predictions you want to see (top k)
top_k = 5

# Print the top k predictions
for prediction in predictions[:top_k]:
    print(f"Class: {prediction['label']}, Confidence: {prediction['score']:.2f}")
```