ShawnHymel / perfect-toast-machine
This is your Edge Impulse project. From here you acquire new training data, design impulses and train models.
About this project
The Perfect Toast Machine
This project attempts to toast bread using odor and temperature data (rather than relying on a simple timer). Using this data, we attempt to predict the "time remaining before burnt" using regression. From this, we can estimate a level of "doneness" e.g. by saying that "toast will be perfect 40 seconds before being burned." By hacking a toaster to cancel the toasting process at this point, we should, in theory, be able to perfectly make toast regardless of starting temperature and bread thickness or composition.
Gas and odor data collected from various types of bread over a Black and Decker simple two-slot toaster. Data was standardized before being uploaded to Edge Impulse. The original dataset, curation script, and inference code can be found here: https://github.com/ShawnHymel/perfect-toast-machine.
A full tutorial showing how to build this AI-powered toaster can be found here: https://www.digikey.com/en/maker/projects/how-to-build-an-ai-powered-toaster/2268be5548e74ceca6830bf35f0f0f9e.
Download block output
Title | Type | Size | |
---|---|---|---|
Raw data training data | NPY file | 5260 windows | |
Raw data training labels | NPY file | 5260 windows | |
Raw data testing data | NPY file | 1315 windows | |
Raw data testing labels | NPY file | 1315 windows | |
Regression model | TensorFlow Lite (float32) | 71 KB | |
Regression model | TensorFlow Lite (int8 quantized) | 20 KB | |
Regression model | TensorFlow SavedModel | 72 KB | |
Regression model | Keras h5 model | 67 KB |
Clone project
Summary
Data collected
18h 15m 50sProject info
Project ID | 129477 |
Project version | 2 |
License | Apache 2.0 |