onnxruntime
Avoid small MatMul batch parameter heap allocations
#29085
Merged
Overview
Commits
3
Changes
View On GitHub
Commits
  • Avoid small MatMul batch parameter heap allocations
    Gopalakrishnan Nallasamy committed 142 days ago
  • MatMul: use inlined_containers_fwd.h and shared constexpr for inline batch cutoff
    Gopalakrishnan Nallasamy committed 123 days ago
  • MatMul: clarify inline-batch comment for DISABLE_ABSEIL build
    Gopalakrishnan Nallasamy committed 123 days ago
FAQ Terms Privacy Refunds Impressum
Loading