The DesiVisionNet is a multi-class image classification model that identifies Indian food of various types. It's implemented in PyTorch after finetuning of ResNet50 over 6700 images. The model achieved 90% accuracy on the Test set taken from our Kaggle Dataset here. For deployment, we built a Gradio app and hosted it on Hugging Face Spaces. You can use it by going to the HF Spaces page and either uploading your own image or clicking on the examples to load them.
Test out the DesiVisionNet here
A sample comparison:
- PyTorch
- Torchvision
- Gradio
- Python
- NumPy
- MatPlotLib
For deployment, we built a Gradio app and hosted it on Hugging Face Spaces.
- Official PyTorch Finetuning Docs
- Official Gradio Docs for images as input
- Deep Learning Concepts from Jovian.ai
- Our Kaggle Dataset
