Training UCF-101 dataset

[Fazlik995]

Can someone train UCF-101 dataset

I can not train it.

If you could train please, leave comment how(details)

Thank you

[HuangZuShu]

@Fazlik995 I think there has some code in dataset.py we need to change, because there is some difference between "trainlist*.txt" and "testlist*.txt" of ucf101.

You can see the difference between "trainlist01.txt" and "testlist01.txt" of ucf101 and you will find that "trainlist01.txt" is end with label id.

trainlist01.txt: ApplyEyeMakeup/v_ApplyEyeMakeup_g08_c01.avi 1 testlist01.txt: ApplyEyeMakeup/v_ApplyEyeMakeup_g01_c01.avi

so you need to change the code in dataset.py

        for line in f:
            class_id = self.class_idx.get(line.split('/')[0]) - 1
            if os.path.exists(os.path.join(self.opt.frame_dir, line.strip('\n')[:-4])) :
                self.data.append((os.path.join(self.opt.frame_dir, line.strip('\n')[:-4]), class_id))