Inspiration
We got our inspiration by combining our different expertise areas. Our team had members of various backgrounds, including ML, applied math, Python, and Data Structures.
What it does
You provide the product with a 128x128 pixel 4x4 grid puzzle and it provides the solution.
How we built it
We build it using the PyCharm IDE, Keras/Tensorflow using CUDA, pillow, and NumPy
Challenges we ran into
We ran in to the issue of having a low accuracy, but we overcame it by forging new strategies in order better design our neural network itself, testing different layouts.
Accomplishments that we're proud of
We had gotten our accuracy to nearly 60% by the morning of the final day
What we learned
We all gained a lot of experience working with Anaconda environments, using Tensorflow, and python images.
What's next for Puzzle Solver
We plan on continuing to improve our data set, which would improve our accuracy
Log in or sign up for Devpost to join the conversation.