Adding torch.Generator to WeightedRandomSampler to ensure pseudorandom results in development phase · Issue #40274 · pytorch/pytorch · GitHub
Improving Control and Reproducibility of PyTorch DataLoader with Sampler Instead of Shuffle Argument | by hengtao tantai | Medium
![GitHub - ufoym/imbalanced-dataset-sampler: A (PyTorch) imbalanced dataset sampler for oversampling low frequent classes and undersampling high frequent ones. GitHub - ufoym/imbalanced-dataset-sampler: A (PyTorch) imbalanced dataset sampler for oversampling low frequent classes and undersampling high frequent ones.](https://user-images.githubusercontent.com/2270240/40677251-b08f504a-63af-11e8-9653-f28e973a5664.png)