No Ingredient Left Behind

• 

• 

  A generated recipe

Inspiration

Food waste is a major problem throughout households; we decided to make a creative solution to combat this problem

What it does

In order to prevent food waste, you input your spare ingredients into the website. It runs through our custom trained co:here model, and recommends you a detailed recipe on how to cook your food.

How we built it

First we used public data on recipes, prune it through python to create a readible dataset. We then create a fine tuned model using co:here for training. Finally, we create a react front-end, and using a FastAPI backend to interact with our co:here model in order to suggest a proper recipe to users!

Challenges we ran into

Finding a proper dataset was difficult; we had to prune certain parts, and find a suitable dataset. That way we could successfully train an accurate model.

Accomplishments that we're proud of

Our execution of the co:here model is something we're very proud of. Having just recently been introduced to it, we have made significant progress in understanding and application

What we learned

We have advanced our knowledge on co:here and FastAPI in order to make this project work.

What's next for No Ingredient Left Behind

A potential feature would be to add CV; it would be great to take a photo of a pantry or fridge, and automatically create a list of ingredients. We could also add image generation of the recipe; output an ai-generated food of what you should end up with.

Built With

• co:here
• fastapi
• html5
• javascript
• python
• react