Erjia Guan

Thank you for asking about it. I am currently working on a PR to enable proper shuffling for `MapDataPipe`. The above behavior is the `map.shuffle` is not shuffled per epoch....

Thinking a little bit on those options. I think the only possible solution might be always calling `finalize_iteration` before `initialize_iteration` in `DL2`'s `__iter__`. In the following case with the current...

I generally have concern on the snapshotting with `itertools` as it heavily uses generator function that won't be serializable. What's your opinion since you might have better knowledge on snapshotting?...

On MacOS, I can't reproduce it either - nightly releases ``` torch=1.13.0.dev20220711 torchtext=0.14.0.dev20220707 torchdata=0.5.0.dev20220711 ``` - official releases ``` torch=1.12.0 torchtext=0.13.0 torchdata=0.4.0 ``` @rkingery Could you please try to use...

@ivaquero You need torchdata==0.4.0 to work with pytorch 1.12.0

@ivaquero I see. I will see if torchdata can provide one for m1 mac. It should be minimum to install `torchdata` from source by ``` git clone https://github.com/pytorch/data cd data...

I am closing this Issue because TorchData currently provides pre-assembled binaries for Mac M1 on [PyPI](https://pypi.org/project/torchdata/#files) and [Conda](https://anaconda.org/pytorch/torchdata/files) after 0.4.1 You should be able to install `torchdata == 0.4.1` and...

This is super weird and can't reproduce on torchdata side. @yushikmr Could you try to run ```py import torchdata from torchdata.datapipes.iter import IterableWrapper ``` in your environment to see if...

Yeah, it would be great if we TorchData are aware of the requirements/criteria to make this happen. Then, we can investigate if we have any step missing.

After #1979 is fixed, could we start to make torchdata as a hard dependency in the main branch? Then, we will be able to spend time on figuring out what's...