ben-da6

calling iter twice messes up dataloaders with queues

8

### Bug description This bug has reappeared https://github.com/Lightning-AI/pytorch-lightning/issues/18414 We now call iter() twice in different places: - https://github.com/Lightning-AI/pytorch-lightning/blob/master/src/lightning/pytorch/loops/fit_loop.py#L263 - https://github.com/Lightning-AI/pytorch-lightning/blob/master/src/lightning/pytorch/loops/training_epoch_loop.py#L171C1-L172C1 ### What version are you seeing the problem on? v2.1...

How do I perform Int8 activation and int8 weight QAT and export to onnx?

3

From the tutorials and recipes it looks like you can only do dynamic Int8 Int4? Also I cannot export the trained model to onnx? ``` import torch from torchao.quantization.prototype.qat import...