The MLIR module may be written in any MLIR dialect with some limitations. Can you explain the details of limitaions? #101
Description
“MLIR_to_BEF walks a MLIR module and emits a BEF binary. The MLIR module may be written in any MLIR dialect with some limitations (one block per region, for example)” I read this from https://github.com/tensorflow/runtime/blob/master/documents/tfrt_host_runtime_design.md.
And I wondor if I want to use MLIR_to_BEF tool to convert any MLIR module(input) to BEF binary(output), are there a tool to convert this any MLIR module to a rule-compliant MLIR module?
Thank you very much!