Jacques Pienaar
Jacques Pienaar
SG, so if one wants any performance here, you should be using the attributed form?
Ah, and while we have this question: what is the tool/config for the different pipelines here? (e.g., is it xla-opt with pipeline foo that captures all the passes)
Issues referenced here have been addressed, pending one is a request to rename which could best be done by way of a PR & discussion there, so closing this issue.
@doru1004 is there perhaps more information you could add here? Perhaps a branch in your fork that you've been making changes on to demonstrate the failure?
Hey, have you tried tensor_load/tensor_store ops here? That seems to do what you want.
Any additional work needed here?
It is/was an integration point for IREE & torch-mlir where one could get a more user friendly checkout experience. We paused/deprioritized it a couple of months ago as we switched...
I think there'll be a few iterations. My gist is "wrong" in that it marks all constants as mutable memory. Embedding the weights in the MLIR file generated is to...
@matthiaskramm for visibility (can't seem to assign you)
@NatashaKnk for visibility given you are looking close to here.