acmmmsys / 2021-grandchallenge-cheapfakes-cosmosonsteroids Goto Github PK

Hi guys,
It has been so long since your work, I have tried

python trainer_script.py -m train

but I got some errors :(
First, if I keep the code

bbox_feats = torch.stack(
            [self.maskrcnn_predictor(img.cpu().numpy(), target) for img, target in zip(imgs, targets)])

I got (see the below images):

TypeError: __call__() takes 2 positional arguments but 3 were given

Second, I tried removing target, now the code became

bbox_feats = torch.stack(
            [self.maskrcnn_predictor(img.cpu().numpy()) for img, target in zip(imgs, targets)])

I did not get the above error anymore, but I got:

TypeError: expected Tensor as element 0 in argument 0, but got dict

Now, I guessed the problem with .numpy() function call, so I tried removing it, the code became:

bbox_feats = torch.stack(
            [self.maskrcnn_predictor(img.cpu()) for img, target in zip(imgs, targets)])

At here I got

AttributeError: 'Tensor' object has no attribute 'strides'

Could you please take a look? I don't know where the cause is from :(

I am a graduate student from Chengdu University of Information Technolgy in China. I am currently learning NLP and have finished reading your paper "COSMOS: Catching Out-of-Context Misinformation with Self-Supervised Learning", however, I am unable to download the dataset. Could you be so kind as to share it with me? It would greatly appreciate it. My google account is [email protected]. Thank you for youor consideration!

Hi, if it possible, can you send me your paper through email [email protected].
Thank you very much.