Sahil Chopra
Sahil Chopra
**What would you like to be added**: Integrate MS-AMP with PyTorch Lightning **Why is this needed**: MS-AMP shows huge gains in throughput when training in FP8. That's very exciting. Adoption...
## 🚀 Feature `CombinedStreamingDataset` allows you to combine multiple `StreamingDataset`s with a sampling ratio -- but it assumes that that the `batch_size` is the same for each dataset. ### Motivation...
## 📚 Documentation Right now, the README shows an example of pausing and resuming the StreamingDataLoader state within a simple for loop over the dataloader. It's not clear to me...
## 🐛 Bug My training run crashed so I tried to resume it from the previous PyTorch Lightning checkpoint. When I do so, I get the following error -- ```...