onnxruntime
44d524ea - Avoid double-counting offline-prepacked GPU weights

Commit

8 days ago

Avoid double-counting offline-prepacked GPU weights Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 1faf9517-5062-4655-a967-3834143334d0

References

#31962 - Add workspace estimation accounting and reporting

#32071 - Integrate CUDA workspace with activation memory patterns

#32189 - Verify Level-2 workspace reservations

Author

chilo-ms

Parents

2c33db55

onnxruntime 44d524ea - Avoid double-counting offline-prepacked GPU weights

onnxruntime
44d524ea - Avoid double-counting offline-prepacked GPU weights