You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
@alankelly I am using XNNPACK with Executorch to lower my model to an Android device and run it for inference.
Running the lowered model for inference fails because of a lack of dynamic shape support in XNNPACK.
I saw that back in October dynamic shape support was being worked on (#5656).
What is the current status?
By the way, I opened an issue on this in the Executorch repo (pytorch/executorch#1350) and communicating with @mcr229 about it.
Thanks
The text was updated successfully, but these errors were encountered:
XNNPack now supports dynamic shapes and this is wired up to TFLite and tested and working. This seems like an ExecuTorch issue so please follow up with them.
@alankelly I am using XNNPACK with Executorch to lower my model to an Android device and run it for inference.
Running the lowered model for inference fails because of a lack of dynamic shape support in XNNPACK.
I saw that back in October dynamic shape support was being worked on (#5656).
What is the current status?
By the way, I opened an issue on this in the Executorch repo (pytorch/executorch#1350) and communicating with @mcr229 about it.
Thanks
The text was updated successfully, but these errors were encountered: