- Remove images from repo, allow GitHub to handle
- Better docs for both usage and training (includes style numbering)
-
HandwriteWrapper forhand(includes textwrap.wrap, basic decisions for user) - Capitilize class names
- Add type hints
- Web GUI
- Text align (left, center, right) | for left select any:
strokes[:, 0] += (view_width - strokes[:, 0].max()) / 2 # Commentstrokes[:, 0] += (view_width + random.randint(-15, 15) - strokes[:, 0].min())strokes[:, 0] += (random.randint(0, 30) - strokes[:, 0].min())
- CLI and argparser
- CI/CD:
- maybe python package
- Docker image
- Better structure for model training (
prepare_data.py,data/,rnn.__init__)
- Migrate to v2
- Move code under submodules