anijain2305 opened a new pull request #4351: [QNN] Lowering for Depthwise Convolution. URL: https://github.com/apache/incubator-tvm/pull/4351 Adding the QNN lowering sequence for Depthwise conv2d. This creates depthwise conv2d with (u)int8 inputs, opening up path to write schedule using Intel VNNI and ARM DOT instructions. For older HW that do not have fast in8 support, this lowering will not be called (already merged upstream in a different PR). @FrozenGene @jackwish @tmoreau89 @yzhliu @zhiics Note - Will wait for #4292 to get merged first as @u99127 already had to resolve conflicts multiple times over past couple of times :) This PR will cause conflict too. But we can still get feedback.
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
