Torch Spyre, the PyTorch backend for the IBM Spyre Accelerator, has reached L2 integration with PyTorch's Cross-Repository CI Relay (CRCR), which triggers downstream CI on upstream changes and surfaces results to reviewers through PyTorch CI CRCR HUD1. The integration uses an agentic four-stage test-selection pipeline, a declarative YAML framework that adapts upstream tests without patching them, and workflow patterns for reliability at scale. Upgrading from PyTorch 2.13 to 2.14 required evaluating roughly 4,000 changed tests instead of tens of thousands. The team indicated most of the framework targets the generic privateuse1 device and could be upstreamed.
Torch Spyre Reaches L2 Integration with PyTorch Cross-Repository CI Relay
Torch Spyre, the PyTorch backend for IBM's Spyre Accelerator, achieved L2 integration with PyTorch's Cross-Repository CI Relay using agentic test…