PyPTO-Lib¶
PyPTO-Lib is a collection of tensor-level kernels, model implementations, and validation workflows built with PyPTO for Ascend NPUs.
Use this documentation to move from a first simulator run to validated model kernels and systematic precision or performance tuning.
Choose a path¶
Run a kernel¶
Start with installation and environment setup, then run your first kernel.
Write a kernel¶
Use the PyPTO Coding chapter for the canonical kernel style and hand-written CCE extern-kernel conventions.
Examples¶
Use the example catalog for focused, self-contained kernels organized by learning level.
Models¶
Use the model pages for end-to-end and component-level model implementations.
Run and Validate¶
Read the compile and runtime workflow and the Golden Harness overview to understand how a script compiles, executes, and checks its result.
Diagnose and optimize¶
Begin with the debugging playbook, then choose the precision or performance workflow for the problem at hand.
Ecosystem¶
- PyPTO documentation covers the programming model, language semantics, and compiler.
- simpler documentation covers the runtime.
- PyPTO-Lib on GitHub contains the source code and issue tracker.