jankolf / ser-fiq-pytorch Goto Github PK

Hello! Is it possible to provide something like thresholds for the quality SER-FIQ Score? Eg does the 0.75 is a high score?
Cause even I test the method on total blurry face images the lowest value I get is around 0.64.

My input image size is 512x512

- [ ] Traceback (most recent call last):
- [ ]   File "/home/dell/workspace/pycharmProject/eva_exp/evaluations/ser-fiq-pytorch-main/main.py", line 34, in <module>
- [ ]     scores = resnet.calculate_serfiq(image, T=10, scaling=5.0)
- [ ]   File "/home/dell/anaconda3/envs/eva/lib/python3.9/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
- [ ]     return func(*args, **kwargs)
- [ ]   File "/home/dell/workspace/pycharmProject/eva_exp/evaluations/ser-fiq-pytorch-main/backbones/iresnet.py", line 205, in calculate_serfiq
- [ ]     x = self.fc(x)
- [ ]   File "/home/dell/anaconda3/envs/eva/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
- [ ]     return forward_call(*args, **kwargs)
- [ ]   File "/home/dell/anaconda3/envs/eva/lib/python3.9/site-packages/torch/nn/modules/linear.py", line 114, in forward
- [ ]     return F.linear(input, self.weight, self.bias)
- [ ] RuntimeError: mat1 and mat2 shapes cannot be multiplied (10x524288 and 25088x512)

And is the results same as the https://github.com/pterhoer/FaceImageQuality?

@jankolf
how to make inference on the Gpu ?

changing device to cuda gives error..

also i changed to resnet50 it gives error. do I need extra code for resnet50 ? which one is better and %xdifference?

device=cuda :error:

.py
Traceback (most recent call last):
File "/data/OpenSource/ser-fiq-pytorch/main.py", line 37, in
scores = resnet.calculate_serfiq(image, T=10, scaling=5.0)
File "/data/OpenSource/ser-fiq-pytorch/venvSERFIQ/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
File "/data/OpenSource/ser-fiq-pytorch/backbones/iresnet.py", line 194, in calculate_serfiq
x = self.conv1(x)
File "/data/OpenSource/ser-fiq-pytorch/venvSERFIQ/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/data/OpenSource/ser-fiq-pytorch/venvSERFIQ/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/data/OpenSource/ser-fiq-pytorch/venvSERFIQ/lib/python3.10/site-packages/torch/nn/modules/conv.py", line 460, in forward
return self._conv_forward(input, self.weight, self.bias)
File "/data/OpenSource/ser-fiq-pytorch/venvSERFIQ/lib/python3.10/site-packages/torch/nn/modules/conv.py", line 456, in _conv_forward
return F.conv2d(input, weight, bias, self.stride,
RuntimeError: Input type (torch.FloatTensor) and weight type (torch.cuda.FloatTensor) should be the same or input should be a MKLDNN tensor and weight is a dense tensor