weitingchen83 / jstasr-desnownet-eccv-2020 Goto Github PK

Hello, I have some questions.
I tested your pre-trained model (my environment -> windows 10, Cuda 9.0, Python 3.6, tensorflow 1.6.0, keras 2.2.0) with [SRRS_15004 dataset's Syn] 1000 random samples.
but I have some problems..😥
In my case, i got the very low PSNR results.. (roughly 19.5)
Image was loaded normally. <- I already checked!
Please check if there is anything I need to do to get the right PSNR values.
Thank u.
Have a good day! 😊

Command - python predict.py -dataroot testImg -datatype tif -predictpath output -batch_size 3

import end
Read img from: testImg
Len of the file: 3
[||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||] 100.0% ...Loading data...
(3, 480, 640, 3) data shape
2022-01-25 10:51:29.790963: W tensorflow/core/common_runtime/gpu/gpu_bfc_allocator.cc:39] Overriding allow_growth setting because the TF_FORCE_GPU_ALLOW_GROWTH environment variable is set. Original config value was 0.
Build TR
Traceback (most recent call last):
File "predict.py", line 95, in
modelRecoverCombine = model.model.build_combine_model()
File "/content/JSTASR-DesnowNet-ECCV-2020/model/model.py", line 181, in build_combine_model
TRinp=Concatenate(axis=3)([pmrc,inp])
File "/usr/local/lib/python3.7/dist-packages/keras/utils/traceback_utils.py", line 67, in error_handler
raise e.with_traceback(filtered_tb) from None
File "/usr/local/lib/python3.7/dist-packages/keras/layers/merge.py", line 499, in build
'A Concatenate layer should be called on a list of '
ValueError: A Concatenate layer should be called on a list of at least 1 input. Received: input_shape=[[(None, 480, 640, 1)], (None, 480, 640, 3)]

Facing this issue when trying to use the pretrained model and running predict.py

Hello, can you provide the pre-training model of DesnowNet? I want to use it for the comparative experiments.

Hello!
I have recently taken an interest in your work. Could you please provide the relevant training code?

Hi! I tried to download the A.h5 weight for modelparam folder but the 2 google-drive web links you provided cannot be found (i.e., open). Please can you provide a new link containing the trained weight. Thanks

File "./predict.py", line 95, in
modelRecoverCombine = model.model.build_combine_model()
File "/home/zhila/Project/Research_Journal/JSTASR-DesnowNet-ECCV-2020/model/model.py", line 192, in build_combine_model
modelSnow.load_weights('./modelParam/snowmodel.h5',by_name=False)
File "/home/zhila/.local/lib/python3.6/site-packages/keras/engine/saving.py", line 492, in load_wrapper
return load_function(*args, **kwargs)
File "/home/zhila/.local/lib/python3.6/site-packages/keras/engine/network.py", line 1230, in load_weights
f, self.layers, reshape=reshape)
File "/home/zhila/.local/lib/python3.6/site-packages/keras/engine/saving.py", line 1224, in load_weights_from_hdf5_group
reshape=reshape)
File "/home/zhila/.local/lib/python3.6/site-packages/keras/engine/saving.py", line 861, in preprocess_weights_for_loading
weights = convert_nested_model(weights)
File "/home/zhila/.local/lib/python3.6/site-packages/keras/engine/saving.py", line 836, in convert_nested_model
original_backend=original_backend))
File "/home/zhila/.local/lib/python3.6/site-packages/keras/engine/saving.py", line 980, in preprocess_weights_for_loading
weights[0] = np.transpose(weights[0], (3, 2, 0, 1))
File "<array_function internals>", line 6, in transpose
File "/home/zhila/.local/lib/python3.6/site-packages/numpy/core/fromnumeric.py", line 653, in transpose
return _wrapfunc(a, 'transpose', axes)
File "/home/zhila/.local/lib/python3.6/site-packages/numpy/core/fromnumeric.py", line 58, in _wrapfunc
return bound(*args, **kwds)
ValueError: axes don't match array

Thank you for your sharing!
As you mentioned in your paper JSTASR that "for each snow image, various types of snow are synthesized by Photoshop and the corresponding snow information is labeled", I want to know how did you use Photoshop to achieve such a large number of non-repetitive mask generation work.
Did you rely on manual operation, or writing scripts to achieve automatic batch processing?
Could you please tell me some related operations or scripts tutorials? Thans a lot!