mohwald / gandtr Goto Github PK

Hello there, I was wondering where did you get the pre-trained HED model from your config /mnt/fry2/landmarkdb/models/pytorch/weights/hed/sniklaus_github.pth, the one I get from the https://github.com/sniklaus/pytorch-hed has different model layer names so I can't load it directly on this code, did you train a model or simply changed the names of the layers from this one?

Hellod, Mohwald.
I want to conduct some comparative experiments with DSA, which is very important to me.
I tried your sample code to reproduce it, however I ran into some difficulties.
I'm going to add your method to the project for testing. (https://github.com/gmberton/VPR-methods-evaluation)
Many famous methods have been added to the models folder of this project.

import torch
import torchvision.transforms as tfm

from models import utils


class DSAModel(torch.nn.Module):
    def __init__(self, device='cuda'):
        super().__init__()        
        self.device = torch.device(device if torch.cuda.is_available() else 'cpu')
        
        self.net = torch.hub.load('mohwald/gandtr', 'gem_vgg16_hedngan').to(self.device)
        self.state_dict = torch.load("/home/ubuntu/.cache/torch/hub/checkpoints/hedngan_embed_vgg16.pth")
        self.net.model.load_state_dict(self.state_dict['model_state'])

        self.un_normalize = utils.UnNormalize(mean=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225))
        self.normalize = tfm.Normalize(mean=[0.48501960784313836, 0.4579568627450961, 0.4076039215686255],
                                       std=[0.00392156862745098, 0.00392156862745098, 0.00392156862745098])

    def forward(self, images):
        images = self.normalize(self.un_normalize(images))
        descriptors = self.net(images)
        return descriptors

1. I cannot load the model onto the GPU.
2. Are the normalization parameters I use fair for DSA?
3. Do you have any other comments about this code?

Thank you.

I can't open any of these download links.

Hello! @mohwald
I successfully tested your trained cyclegan, but could not import Hedgan's checkpoints correctly in the open source CUT project.
I use the default parameters in CUT's base_options.py