It is workable. The VM can gather the workloads and pack them as a single dynamic-shape workload and forward it to DietCode. DietCode does the auto-scheduling and send the generated code back to the VM.
-- Reply to this email directly or view it on GitHub: https://github.com/apache/tvm-rfcs/pull/72#issuecomment-1205995569 You are receiving this because you are subscribed to this thread. Message ID: <apache/tvm-rfcs/pull/72/c1205995...@github.com>
