chi0tzp / pyvggface Goto Github PK

Hi, thank you for your work. I meet some mistakes when convert weights:

Traceback (most recent call last):
File "E:/BaoXiu/EX8/PyVGGFace-master/convert_weights.py", line 67, in
model=vggface_model)
File "E:/BaoXiu/EX8/PyVGGFace-master/convert_weights.py", line 32, in convert
torch_model = torchfile.load(torch_weights_file)
File "C:\ProgramData\Anaconda3\envs\temp\lib\site-packages\torchfile.py", line 424, in load
return reader.read_obj()
File "C:\ProgramData\Anaconda3\envs\temp\lib\site-packages\torchfile.py", line 370, in read_obj
obj._obj = self.read_obj()
File "C:\ProgramData\Anaconda3\envs\temp\lib\site-packages\torchfile.py", line 387, in read_obj
obj[k] = v
TypeError: unhashable type: 'numpy.ndarray'

And my envs is:

numpy (1.14.3)
Pillow (5.2.0)
pip (9.0.1)
torch (1.1.0)
torchfile (0.1.0)
torchvision (0.3.0)

Could you give me some help?

Hello, I download the pretrained file vgg_face_torch.tar.gz (Torch), and follow your step1 but I got the error "unhashable type: 'numpy.ndarray'", could you help me with this error, or may you share your converted .pth model? Thanks a lot

As describe in the title, I get confused about the process here

Is this process exactly the

‘with the average face image (computed from the training set) subtracted’

in Deep Face Recognition