cassiepython / cddfm3d Goto Github PK

Hi! Thanks for your great work. It seems that there is no code for the algorithm "Attribute-Adaptive Layer Selection Algorithm" mentioned in your paper.

I am trying to edit the facial pose of a face using python evaluate.py --name pose --model RIGModelP --load_apnet_epoch 80 but I fail.

Could you provide an example of editing the head pose?

Thanks in advance.

Hi, thanks for the excellent work.
I am trying to test the effect of transferring attributes from one source image to another target image, (e.g. makes the target person show the same expression with the source person), as shown in the Figure from the paper.

After reading the code, I am confused that whether this function is implemented in the code? The test_shape in evaluate.py seems a bit related but I cannot figure out what the pca_id and pca_val mean.

Does the test_shape function do the thing I want? If not, what should I do to implement the expression transfer process?

Following the default procedure and parameter settings, I cannot train the APNet successfully. For different latents input, the network always produces the same result, e.g. mean face.

I tried to print the output of each layer (as shown below). It seems the network has already collapsed.

Hi, thank you for your work!

I like your approach with fixing the same style space for different domains and would have liked to reproduce it. So I want to be sure about what exactly was your approach.

To keep the latent consistency across all generators, we freeze the network layers relevant to latent embedding. To be specific, since we use the style space 𝒮, all style convolution layers and tRGB layers, Menglei: i.e., the affine transformation layers, are fixed during the finetuning.

Hence, all the affine layers are freezed together with all tRGB layers?

 Besides that, we use exactly the same loss functions and hyper-parameters as in StyleGAN2

Would it affect the result if stylegan-ada is used for finetuning a new model?

Looking forward to hearing from you!

Interesting work, could you update the remianed toDo?

Hi, firstly thank you for sharing this work.

When I run the python evaluate.py --name apnet_wpdc --model APModel command in Attribute Prediction part, I got the error as follows:

Model APModel was created
  0%|                                                                                                                                                                                     | 0/199 [00:00<?, ?it/s]
Traceback (most recent call last):
  File "evaluate.py", line 192, in <module>
    model.test_shape()
  File "evaluate.py", line 159, in test_shape
    latent_pred, params_source, params_pred = self._model.forward_test(latent, pca_id, pca_val)
TypeError: forward_test() takes 2 positional arguments but 4 were given

And when I check the code, I found the function of forward_test in APModel class exactly accept only one latent argument, however, the above function need three arguments.

I don't know how to fix this error.