tlopex commented on PR #20109: URL: https://github.com/apache/tvm/pull/20109#issuecomment-5335703647
Sorry for the late reply since the CI was broken. The float64 promotion happens too early here. For float32 inputs, we should compute `limit - start` in the original dtype first, then cast the difference and `delta` to float64. For example, `Range(-7.8f, -1.4f, 0.2f)` currently produces 33 elements in TVM but 32 in ONNX Runtime. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
