A FastAI model built with a finetuned RESNET50 and using cosine similarity to give user recommendations based on their favourite images of cats by their breed in a Jupyter notebook. BTW - the accuracy of the model in making correct predictions of breed is 91%.
Here're some of the project's best features:
- FastAI model trained using ResNet
- Get Recommendations on 6 different breeds!
- Cosine Similarity for recommendations
1. Just download and run the Jupyter notebook step by step. Make sure all the imports are installed.
Make Pull requests which improve the functionality of the application in any sorts. It should conform with the following conditions:
- Clear, short, crisp description of the PR.
- Should add on to the value of the application.
Technologies used in the project:
- Python
- Fast AI
- Jupyter Notebook
- Torch
- Numpy
- Pandas
- SKLearn
Distributed under the MIT License.