related code:
#... some other code
target = tvm.target.Target(target="cuda", host="llvm")
device = tvm.device(target.kind.name, 0)
with tvm.transform.PassContext(opt_level=1):
func = relay.vm.compile(mod, target=target, params=params)
ctx = tvm.cuda(0)
vm = tvm.runtime.vm.VirtualMachine(func, ctx)
results = []
rays = np.load("mydatapath")
for i in range(1065):
print(i)
start = i*1420
end = start + 20
aa = rays[start: end]
print(aa.shape)
vm.set_input("main", tvm.nd.array(aa, device))
vm.invoke_stateful("main") # get error here
tvm_out = vm.get_outputs()[0]
print("tvm_out: ", len(tvm_out.asnumpy()))
results.append(tvm_out.asnumpy())
print("results: ", len(results))
result = np.stack(results)
# ... some other code to get pictures.
error:
Traceback (most recent call last):
File "test_torch.py", line 205, in <module>
vm.invoke_stateful("main")
File "/cpfs01/user/zhangyuchang/tvm_rgst_cuda/python/tvm/runtime/vm.py", line 561, in invoke_stateful
self._invoke_stateful(func_name)
File "/cpfs01/user/zhangyuchang/tvm_rgst_cuda/python/tvm/_ffi/_ctypes/packed_func.py", line 239, in __call__
raise_last_ffi_error()
File "/cpfs01/user/zhangyuchang/tvm_rgst_cuda/python/tvm/_ffi/base.py", line 476, in raise_last_ffi_error
raise py_err
tvm.error.InternalError: Traceback (most recent call last):
5: tvm::runtime::PackedFuncObj::Extractor<tvm::runtime::PackedFuncSubObj<tvm::runtime::vm::VirtualMachine::GetFunction(tvm::runtime::String const&, tvm::runtime::ObjectPtr<tvm::runtime::Object> const&)::{lambda(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)#2}> >::Call(tvm::runtime::PackedFuncObj const*, tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)
4: tvm::runtime::vm::VirtualMachine::GetFunction(tvm::runtime::String const&, tvm::runtime::ObjectPtr<tvm::runtime::Object> const&)::{lambda(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)#1}::operator()(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*) const [clone .isra.0]
3: tvm::runtime::vm::VirtualMachine::Invoke(tvm::runtime::vm::VMFunction const&, std::vector<tvm::runtime::ObjectRef, std::allocator<tvm::runtime::ObjectRef> > const&)
2: tvm::runtime::vm::VirtualMachine::RunLoop(std::vector<long, std::allocator<long> > const&)
1: tvm::runtime::vm::VirtualMachine::InvokePacked(long, tvm::runtime::PackedFunc const&, long, long, std::vector<tvm::runtime::ObjectRef, std::allocator<tvm::runtime::ObjectRef> > const&)
0: tvm::runtime::PackedFuncObj::Extractor<tvm::runtime::PackedFuncSubObj<tvm::runtime::WrapPackedFunc(int (*)(TVMValue*, int*, int, TVMValue*, int*, void*), tvm::runtime::ObjectPtr<tvm::runtime::Object> const&)::{lambda(tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)#1}> >::Call(tvm::runtime::PackedFuncObj const*, tvm::runtime::TVMArgs, tvm::runtime::TVMRetValue*)
File "/cpfs01/user/zhangyuchang/tvm_rgst_cuda/src/runtime/library_module.cc", line 76
InternalError: Check failed: ret == 0 (-1 vs. 0)
What could be the possible reasons? Is my CUDA operator not registered correctly in TVM? I can already see my custom operator in the Relay IR.
