I’m implementing a correlation operator which is based on the following equation from FlowNet: Learning Optical Flow with Convolutional Networks:
I was able to succesfully tune the operator manually but when I create a tunable template to use autoTVM I get the following error which results in no tuning at all.
RuntimeError('Traceback (most recent call last):\n
[bt] (3) /tvm/build/libtvm.so(TVMFuncCall+0x48) [0x7f436146d4a8]\n
[bt] (2) /tvm/build/libtvm.so(std::_Function_handler<void (tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*),
tvm::runtime::RPCModuleNode::WrapRemote(void*)::{lambda(tvm::runtime::TVMArgs,
tvm::runtime::TVMRetValue*)#1}>::_M_invoke(std::_Any_data const&, tvm::runtime::TVMArgs&&,
tvm::runtime::TVMRetValue*&&)+0x42) [0x7f43614cc412]\n
[bt] (1) /tvm/build/libtvm.so(tvm::runtime::RPCSession::CallFunc(void*, tvm::runtime::TVMArgs,
tvm::runtime::TVMRetValue*, void* (*)(int, tvm::runtime::TVMArgValue const&), tvm::runtime::PackedFunc const*)+0x1e6) [0x7f43614d5b26]\n
[bt] (0) /tvm/build/libtvm.so(dmlc::LogMessageFatal::~LogMessageFatal()+0x35) [0x7f4360cb2885]\n
File "/tvm/src/runtime/rpc/rpc_session.cc", line 995\nTVMError: Check failed: code == RPCCode: :kReturn: code=4',)
Some clue of how I can further debug this error? I have already tried solutions from other similar publications without success.
Thanks.