„Esse est percipi.“ George Berkeley (1685 - 1753)
What it does
It's about museum-relevant content, that is created by users.
Makes fun to use.
See GitHub link for more info.
How we built it
The model was trained with the help of transfer learning.
Backand: PyTorch, Flask
Frontend: Angular, React
Challenges we ran into
PyTorch was new for me.
So far I've only used TensorFlow and wanted to try PyTorch out in this hackathron.
Exporing latent space of GAN need more time.
Creating the videos isn't pretty.
Too little sleep and thus reduced problem-solving possibilities.
Accomplishments that we're proud of
I'm starting to like PyTorch.
What we learned
What's next for percipi