Releases are displayed in reverse chronological order and each release will contain new, changed and fixed content. Since this repository is a bit unique, here's how we'll do semantic versioning.
- Major version X (X.y.z): When an entire set of notebook lessons is complete.
- Minor version Y (x.Y.z): When a notebook lesson is complete.
- Patch version X (x.y.Z): Changes made in between notebook lessons.
- all basic ML notebooks in TensorFlow 2.0 + Keras
- explanations for all concepts
- visualizations
- interpretability
- "From scratch" implementations
- 🔢 NumPy lesson is complete.
- Code and notebook complete for the NumPy lesson.
- 🐍 Python lesson is complete.
- Code and notebook complete for the python lesson.
- 📓 Notebooks lesson is complete.
- coding framework switched from PyTorch to TensorFlow 2.0 + Keras.
Using fit and callbacks, we can avoid large training scripts but still achieve fine-grained customization (LR scheduling, early stopping, custom callbacks, etc.) However for more advanced content where we need custom implementation regardless, we’ll use native TF or PyTorch.
- all notebooks in PyTorch