🐛 Describe the bug

If you try to set the seed = 0 as such:

dp = IterableWrapper(range(20)).shuffle()
rs = MultiProcessingReadingService(num_workers=2)

dl = DataLoader2(datapipe=dp, reading_service=rs)
dl.seed(0)

The seed will not be set because this line returns False. I think the true intention is to skip over None, but 0 also gets skipped as it is currently implemented.

https://github.com/pytorch/data/blob/7d9747588173f38e3ce0d5c58ba3746dacb4438f/torchdata/dataloader2/dataloader2.py#L203

I suppose we can say seed = 0 is invalid but it is probably better to check for self._seed is not None instead.

Versions

main