A basic Python script doing the following:
- Training an auto-encoder with the minst dataset.
- Saving the trained auto-encoder (.json/.h5 in models directory) for future use.
- Making predictions (testing or aka decoding with auto-encoders).
- Writing the decoded images in decoded directory.
- Converting the png images (decoded) into svg using potrace then writing them in decoded_to_svg directory.
Installing dependencies: pip install -r requirements.txt
requirements.txt compiled through the command: pip-compile requirements.in