Code not working

[chrisdottel]

Code references a custom_dataset module that doesn't exist. Removed it. Then I get this ghost parameter

test.py: error: the following arguments are required: --dataroot

There seems to be no description to what this parameters is for. Passing it in blank we get

TypeError: Can't instantiate abstract class Pix2PixModel with abstract methods optimize_parameters

Did anyone work through all of this to get the base pretrained model working?

[valillon]

Last thrown error means the class Pix2PixModel does not implement the method optimize_parameters(), previously declared in the class BaseModel.

For testing purposes, a quick fix can be done by implementing a pass statement, since optimization is not required:

# inside Pix2PixModel class in pix2pix_model.py
def optimize_parameters(self):
    """No optimization for test model."""
    pass

After that, there is a bunch of errors in test.py regarding imports and a leftover self in load_image(). This code snippet should fix it:

# from custom_dataset import DepthDataset
from torchvision import transforms
from PIL import Image
import glob

def load_image(image_path):

    # Convert image to tensor
    to_tensor = transforms.Compose([
        transforms.ToTensor(),
        transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
    ])

    # image
    image = Image.open(image_path).convert('RGB')
    image = to_tensor(image)    # remove `self.`
    return { 'A': image.unsqueeze(0) }

And finally, there's a missing method convert_model().

Traceback (most recent call last):
  File "test.py", line 119, in <module>
    model = create_model(opt)   # create a model given opt.model and other options
  File "/home/Imatge/projects/vivim/PeeledHuman/models/__init__.py", line 65, in create_model
    instance = model(opt)
  File "/home/Imatge/projects/vivim/PeeledHuman/models/test_model.py", line 46, in __init__
    opt.norm, not opt.no_dropout, opt.init_type, opt.init_gain, self.gpu_ids)
  File "/home/Imatge/projects/vivim/PeeledHuman/models/networks.py", line 149, in define_G
    return init_net(net, init_type, init_gain, gpu_ids)
  File "/home/Imatge/projects/vivim/PeeledHuman/models/networks.py", line 115, in init_net
    net = convert_model(net)
NameError: name 'convert_model' is not defined

Not sure what it really does, it's anywhere declared. But comment that 115 line and the code apparently runs with no crash.

I also tried to use the class TestModel via --model test (instead of pix2pix), but it's apparently incompatible with the provided checkpoint.

On a side note, I tested on several video frames and the code throws a bunch cluttered files (*_mesh.obj, *.inp.png, and *_depth.png) as expected, however some frames and meshes were missing. But hey, I should probably read the paper and code more carefully.

[robinsonkwame]

@valillon the checkpoint link is no longer available, would mind uploading it to a google drive link so that others can work with this repository? It would be much appreciated!