When integrating with other framework such as PyTorch and etc, it's more desirable to avoid unnecessary copies of activations and weights when hooking up TVM runtime and run.
cc: @ajtulloch @hlu1 You can view, comment on, or merge this pull request online at: https://github.com/dmlc/tvm/pull/3416 -- Commit Summary -- * Enable set_input_zero_copy in GraphRuntime -- File Changes -- M src/runtime/graph/graph_runtime.cc (151) M src/runtime/graph/graph_runtime.h (29) -- Patch Links -- https://github.com/dmlc/tvm/pull/3416.patch https://github.com/dmlc/tvm/pull/3416.diff -- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: https://github.com/dmlc/tvm/pull/3416
