GoogLe Net implementation

[mdit108]

In which part of the code is the GoogLe Net as the first part of the DSN specified?

[HERIUN]

from torchvision.models import googlenet
import torch

model = googlenet(pretrained=True)

extractor = torch.nn.Sequential(*list(model.children())[:-2])

im = torch.randn(1,3,720,1280) # NCHW
feature = extractor(im).cpu().numpy().flatten() # [1,1024,1,1] -> [1024]

i try like this...

[ruanzhijian]

I wonder if he is using the pool5 layer of the googlenet network for feature extraction, so is that the code you wrote? Or is there some other additional code.

[ehdrndd]

I wonder if he is using the pool5 layer of the googlenet network for feature extraction, so is that the code you wrote? Or is there some other additional code.

I write the code myself.

[ruanzhijian]

@ehdrndd Can you share your feature extraction code? Or give a link, thanks! You can add a contact if it is convenient

[HERIUN]

@ruanzhijian It may help you.

https://github.com/HERIUN/vsumm-reinforce_re/blob/main/generate_dataset.py