masahi edited a comment on pull request #8084: URL: https://github.com/apache/tvm/pull/8084#issuecomment-845460104
As the onnx doc says, https://github.com/onnx/onnx/blob/master/docs/Operators.md#GatherND, `the leading B number of dimensions of data tensor and indices are representing the batches`, so there is a constraint `X_0, ... X_{B-1} == Y_0, ... Y_{B-1}`. When I wrote the output shape as `(Y_0, ..., Y_{K-1}, X_{M+B}, ..., X_{N-1})`, more precisely it means `(Y_0, ..., Y_{B-1}, ... Y_{K-1}, X_{M+B}, ..., X_{N-1})`, which is equivalent to `(X_0, ..., X_{B-1}, ... Y_{K-1}, X_{M+B}, ..., X_{N-1})`. -- 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]
