Resampling for IterableDatasets

Hello everyone. I noticed that the sampler parameter of DataLoader doesn’t work with an IterableDataset, as explained in this issue. I have an interest in online learning and have thus developed some methods for balancing a streaming dataset through resampling. I’ve recently implemented these as part of a small package dedicated to PyTorch.

I’m not very accustomed to PyTorch and its ecosystem, and would thus appreciate some feedback on the package. Do you find it useful? Is it idiomatic? Would you like to see any other features?

I’m sorry in advance if this post is off-topic. I’m happy to remove my post if this isn’t suitable.