Examples

Hands-on Jupyter notebooks for learning Nabla from basics to distributed training and custom kernels.

Example Notebooks

• Example 1: Tensors and Operations
• Example 2: Automatic Differentiation
• Example 3: Graph Tracing — Under the Hood
• Example 4a: MLP Training — PyTorch-Style (Imperative)
• Example 4b: MLP Training (JAX-Style / Functional)
• Example 5: Transforms and @nb.compile
• Example 6a: Transformer Training — PyTorch-Style (Imperative)
• Example 6b: Transformer Training (JAX-Style / Functional)
• Example 7: Pipeline Parallelism (GPipe)
• Example 8: 2D Parallel Training (Pipeline + Data Parallelism)
• Example 9: Pipeline Parallel Inference
• Example 10: Compiled vs Eager vs JAX
• Example 11: LoRA & QLoRA Fine-Tuning
• Example 12: Custom Mojo Kernels
• Example 13: CNN Training