Hello.
First of all, thank you for releasing this cool work. I read the recent works and found this release.
I have successfully built the tiny-cuda-nn and ran mlp_learning_an_image and checked that it work.
However, I'm facing the following error when building bindings/torch.
tinycudann/bindings.cpp: In function ‘void* void_data_ptr(at::Tensor&)’:
tinycudann/bindings.cpp:60:48: error: expected primary-expression before ‘float’
case torch::kFloat32: return tensor.data_ptr();
^~~~~
tinycudann/bindings.cpp:60:48: error: expected ‘;’ before ‘float’
tinycudann/bindings.cpp:60:53: error: expected unqualified-id before ‘>’ token
case torch::kFloat32: return tensor.data_ptr();
^
tinycudann/bindings.cpp:61:56: error: expected primary-expression before ‘>’ token
case torch::kHalf: return tensor.data_ptrtorch::Half();
^
tinycudann/bindings.cpp:61:58: error: expected primary-expression before ‘)’ token
case torch::kHalf: return tensor.data_ptrtorch::Half();
^
tinycudann/bindings.cpp: In member function ‘std::tuple<tcnn::cpp::Context, at::Tensor> Module::fwd(at::Tensor, at::Tensor)’:
tinycudann/bindings.cpp:96:59: error: expected primary-expression before ‘float’
m_module->inference(stream, batch_size, input.data_ptr(), void_data_ptr(output), void_data_ptr(params));
^~~~~
tinycudann/bindings.cpp:98:63: error: expected primary-expression before ‘float’
ctx = m_module->forward(stream, batch_size, input.data_ptr(), void_data_ptr(output), void_data_ptr(params), input.requires_grad());
^~~~~
tinycudann/bindings.cpp: In member function ‘std::tuple<at::Tensor, at::Tensor> Module::bwd(const tcnn::cpp::Context&, at::Tensor, at::Tensor, at::Tenr)’:
tinycudann/bindings.cpp:161:48: error: expected primary-expression before ‘float’
input.requires_grad() ? dL_dinput.data_ptr() : nullptr,
^~~~~
tinycudann/bindings.cpp:161:48: error: expected ‘:’ before ‘float’
tinycudann/bindings.cpp:161:48: error: expected primary-expression before ‘float’
tinycudann/bindings.cpp:164:20: error: expected primary-expression before ‘float’
input.data_ptr(),
^~~~~
tinycudann/bindings.cpp: In member function ‘at::Tensor Module::initial_params(size_t)’:
tinycudann/bindings.cpp:175:53: error: expected primary-expression before ‘float’
m_module->initialize_params(seed, output.data_ptr());
^~~~~
error: command 'gcc' failed with exit status 1
I googled about this kind of error, and it says its about the wrong use of types.
However, I feel like there's nothing wrong with the code.
I've read #48, but it seems the problem is not the same.
I'm having a strong suspicion
My configuration is as below.
ubuntu 18.04
RTX8000
Driver version
cuda 11.0
torch 1.2.1
Thank you in advance!
NVlabs
