GAN_Metrics-Tensorflow
GAN_Metrics-Tensorflow copied to clipboard
Question about test images
├── real_source
├── aaa.png
├── bbb.jpg
├── real_target
├── ccc.png
├── ddd.jpg
├── fake
├── ccc_fake.png
├── ddd_fake.jpg
├── main.py
├── inception_score.py
└── frechet_kernel_Inception_distance.py
As you say, when we use the metrics, the test images should has the above format.
If I want to test the performance of selfie to anime model, the real source
is any selfie image, the fake
is my generated image, how could I select proper real target
images? Since you the xxx_fake.jpg
is paired with real target
?
Did the GAN metrics are statistics value with a lot samples, and can't evaluate on an single image?
the same question.
The reference listed should be helpful.
Have you solved the problem
Same problem 🤦♂️🤦♂️