Aleksei-grovety opened a new pull request, #14464:
URL: https://github.com/apache/tvm/pull/14464

   Before format NHCWB16 was used for operations performed by NPU, but there 
are limitations, for example ReduceSum requires NHWC input in case input tensor 
has type int32 or accelerator is Ethos_U65_512 
https://review.mlplatform.org/plugins/gitiles/ml/ethos-u/ethos-u-vela/+/refs/tags/3.7.0/ethosu/vela/graph_optimiser_util.py#126
   
   cc @neildhickey, @lhutton1, @ekalda, @leandron, @ilyag-grovety 


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to