openxla / xla

A machine learning compiler for GPUs, CPUs, and ML accelerators
Apache License 2.0
2.55k stars 394 forks source link

Use `xla::ifrt::AttributeMap` for cost analysis results #14872

Closed copybara-service[bot] closed 1 month ago

copybara-service[bot] commented 1 month ago

Use xla::ifrt::AttributeMap for cost analysis results

This is for consistency with other attribute maps in IFRT. Also, using AttributeMap makes it possible to provide stable serialization, e.g., for IFRT Proxy.

JAX Python binding still uses xla::PjRtValueType since pybind/nanobind doesn't recognize xla::ifrt::AttributeMap::Value as Python native types.