kevinthesun edited a comment on pull request #6449: URL: https://github.com/apache/incubator-tvm/pull/6449#issuecomment-692285615
> But we should still know when it is appropriate to run the inference, should we not? > I must admit I'm still troubled by the idea that we don't know beforehand which inputs to process with which method. We do this for some ops which have dynamic attributes. When those dynamic attributes are relay Expr, we need to try to infer value to make the generated relay program as static as possible. There is no good way to further tell which Relay Expr needs to be inferred(And not necessary since _infer_value does general evaluation for Relay Expr). ---------------------------------------------------------------- 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: us...@infra.apache.org