Python + GPT-4o + Flask = AI Meal Calorie Detector (GPT-4 Vision API)

preview_player
Показать описание
In today's video I create a GPT-4o powered simple meal calorie detector where you can upload an image of a meal and it will analyze it and tell you how many calories is in the meal

00:00 Testing with ChatGPT UI
03:38 Start coding: Connect to GPT-4o API
08:45 How to set the OpenAI API key
12:44 Getting JSON output from GPT4-o
18:25 Creating the Flask server
19:19 Design UI with ChatGPT
27:47 Handling the file upload
39:32 Adding a loading spinner
46:01 Final demo
47:19 Outro
Рекомендации по теме
Комментарии
Автор

Wow. Seamless integration of Python and JavaScript at the bleeding edge of AI Computer Vision.
What a great tutorial video. Thank you.

brianmcclain
Автор

I just had this exact same idea like 5 days ago and talked with my colleagues at work about it, nice to see you implement this, conventional calorie trackers are way too much work IMO

PatrickWeygoldt
Автор

I wonder if it is accurate in terms of the calories and other nutrients.

As we know, it just has an image. No depth information, it doesn't know the proportion and other params. Additionally, we know in numbers GPTs are not that good currently and if you run the same prompt again, you will see different numbers.

On the "average calories" it is good. But specific to this image, the calories and other nutrients may not be reliable. But hey, I just shared my thoughts, please share yours!

waytolegacy
Автор

Hello, could you give me some information regarding the use of artificial intelligence image editing software in a country?

chrisder