rust-ml
Installation
SKILL.md
Machine Learning Development
Domain Constraints
| Domain Rule | Design Constraint | Rust Implication |
|---|---|---|
| Large data | Efficient memory | Zero-copy, streaming |
| GPU acceleration | CUDA/Metal support | candle, tch-rs |
| Model portability | Standard formats | ONNX |
| Batch processing | Throughput over latency | Batched inference |
| Numerical precision | Float handling | ndarray, careful f32/f64 |
| Reproducibility | Deterministic | Seeded random, versioning |
Critical Rules
- Avoid copying large tensors — memory bandwidth is the bottleneck. Use references, views, in-place ops.
- Batch operations for GPU efficiency — GPU has overhead per kernel launch, batch to amortize.
- Use standard model formats — train in Python, deploy in Rust via ONNX.