PyTorch for the IPU: User Guide
- 1. Introduction
- 2. Installation
- 3. From PyTorch to PopTorch
- 4. Features
- 4.1. Options
- 4.2. Model wrapping functions
- 4.3. Error handling
- 4.4. Multi-IPU execution strategies
- 4.5. Optimizers
- 4.6. PopTorch ops
- 4.7. 16-bit float support
- 4.8. Automatic mixed-precision casting
- 4.9. PyTorch buffers
- 4.10. Creating custom ops
- 4.11. Precompilation and caching
- 4.12. Environment variables
- 5. Efficient data batching
- 6. IPU supported operations
- 7. Debugging your model
- 8. Efficient IPU I/O
- 9. Examples
- 10. Experimental features
- 11. Legacy tracing frontend
- 12. API reference
- 13. Index
- 14. Legal notices
- 15. Changelog