jimmychame/lsfnet

Pytorch code for "Low-light Image Restoration with Short- and Long-exposure Raw Pairs"

lsfnet's Issues

How to get access to your datasets？

model and dataset

Hi, thanks for sharing the code.
Can you share the model and dataset?

Why gain is applied two times?

Hi,
First of all thank you for sharing the code.
My question is regarding the function gen_blur in the file preprocess.py:
In line 65, the gain is applied before applying the blur kernel. Does this correspond to the scaling operation mentioned in the paper? How is that different from the same operation which is applied in line 58 of file dataloader.py?
Also why quantization operation is not applied on the blurry image? (as it is applied on noisy image in line 90 of dataloader.py)
I feel there are some inconsistencies between the data generation pipeline illustrated in Figure 3 of the paper with the provided implementation. Can you please clarify more on this?

About DCN

Hi, thanks for your excellent work.
My question is why I can't compile dcn, it will report the following error.I really want to use dcn in my environment(pytorch1.4)
/home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/c10/util/Exception.h:339:40: warning: ‘void c10::detail::deprecated_AT_CHECK()’ is deprecated [-Wdeprecated-declarations]
::c10::detail::deprecated_AT_CHECK();
^
src/deform_conv_cuda.cpp:142:5: note: in expansion of macro ‘AT_CHECK’
AT_CHECK((gradOutput->size(dimh) == outputHeight &&
^~~~~~~~
In file included from /home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/c10/core/Device.h:5:0,
from /home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/c10/core/Allocator.h:6,
from /home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/ATen/ATen.h:3,
from /home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/torch/csrc/api/include/torch/types.h:3,
from /home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/torch/csrc/api/include/torch/data/dataloader_options.h:4,
from /home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/torch/csrc/api/include/torch/data/dataloader/base.h:3,
from /home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/torch/csrc/api/include/torch/data/dataloader/stateful.h:3,
from /home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/torch/csrc/api/include/torch/data/dataloader.h:3,
from /home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/torch/csrc/api/include/torch/data.h:3,
from /home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/torch/csrc/api/include/torch/all.h:4,
from /home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/torch/extension.h:4,
from src/deform_conv_cuda.cpp:4:
/home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/c10/util/Exception.h:314:13: note: declared here
inline void deprecated_AT_CHECK() {}
^~~~~~~~~~~~~~~~~~~
/usr/include/c++/6/type_traits:1558:8: note: provided for ‘template<class _From, class _To> struct std::is_convertible’
struct is_convertible
^~~~~~~~~~~~~~
/usr/include/c++/6/tuple:502:1: error: body of constexpr function ‘static constexpr bool std::_TC<, _Elements>::_NonNestedTuple() [with _SrcTuple = const std::tuple<at::Tensor&, at::Tensor&, at::Tensor&>&; bool = true; _Elements = {at::Tensor&, at::Tensor&, at::Tensor&}]’ not a return-statement
}
^
/usr/include/c++/6/tuple: In instantiation of ‘static constexpr bool std::_TC<, _Elements>::_NonNestedTuple() [with _SrcTuple = std::tuple<at::Tensor&, at::Tensor&, at::Tensor&>&&; bool = true; _Elements = {at::Tensor&, at::Tensor&, at::Tensor&}]’:
/usr/include/c++/6/tuple:686:422: required by substitution of ‘template<class ... _UElements, class _Dummy, typename std::enable_if<((std::_TC<(1ul == sizeof... (_UElements)), at::Tensor&, at::Tensor&, at::Tensor&>::_MoveConstructibleTuple<_UElements ...>() && std::_TC<(1ul == sizeof... (_UElements)), at::Tensor&, at::Tensor&, at::Tensor&>::_ImplicitlyMoveConvertibleTuple<_UElements ...>()) && std::_TC<(std::is_same<_Dummy, void>::value && (1ul == 1)), at::Tensor&, at::Tensor&, at::Tensor&>::_NonNestedTuple<tuple<_Elements ...>&&>()), bool>::type > constexpr std::tuple< >::tuple(std::tuple<_Args1 ...>&&) [with _UElements = {at::Tensor&, at::Tensor&, at::Tensor&}; _Dummy = void; typename std::enable_if<((std::_TC<(1ul == sizeof... (_UElements)), at::Tensor&, at::Tensor&, at::Tensor&>::_MoveConstructibleTuple<_UElements ...>() && std::_TC<(1ul == sizeof... (_UElements)), at::Tensor&, at::Tensor&, at::Tensor&>::_ImplicitlyMoveConvertibleTuple<_UElements ...>()) && std::_TC<(std::is_same<_Dummy, void>::value && (1ul == 1)), at::Tensor&, at::Tensor&, at::Tensor&>::_NonNestedTuple<tuple<_Elements ...>&&>()), bool>::type = ]’
/home/amax/anaconda3/envs/pytorch1/lib/python3.7/site-packages/torch/include/ATen/Functions.h:4128:229: required from here
/usr/include/c++/6/tuple:495:244: error: wrong number of template arguments (4, should be 2)
return _and<_not<is_same<tuple<_Elements...>,
^
/usr/include/c++/6/type_traits:1558:8: note: provided for ‘template<class _From, class _To> struct std::is_convertible’
struct is_convertible
^~~~~~~~~~~~~~
/usr/include/c++/6/tuple:502:1: error: body of constexpr function ‘static constexpr bool std::_TC<, _Elements>::_NonNestedTuple() [with _SrcTuple = std::tuple<at::Tensor&, at::Tensor&, at::Tensor&>&&; bool = true; _Elements = {at::Tensor&, at::Tensor&, at::Tensor&}]’ not a return-statement
}
^

About test dataset

How to choose 70 images from SID dataset to form the test set?

DCNv2 model

Hi，after downloading the dcn file from EDVR ,How to get the DCNv2 model file which has to be saved in .\ckpt\

.MAT files

Hi, the .mat file is processed into a grayscale map after getting a .mat file or a three-channel .mat file directly from a color image?
When getting the h5py file of the train dataset, the following error appears when running the gen_deatset.py:
meta['WhiteLevel'] = [metadata['meta']['saturation'][0][0][0][0]]
KeyError: 'meta'
So consider whether the .mat file and the data in the dictionary do not correspond?
thanks！