Pierre Stock comments

Results 5 comments of


                                            Pierre Stock

To allow batchnorm layers with frozen running stats

Hey EmmyFang, Thanks for your interest. I think that's a good idea of enhancement to include in a future release. However, one question remains: how do we exactly initialize the...

To allow batchnorm layers with frozen running stats

Hey Emmy, Apologies for the late reply. Yes, you understood my point correctly. Will add this in our roadmap. Thanks, Pierre

Remove `_module` prefix from GradSampleModule parameter and buffer names

Also loosely related to [this discussion](https://discuss.pytorch.org/t/modulevalidator-fix-does-not-immediately-change-the-model-how-to-deal-with-it/152783).

Length of BatchSplittingSampler with Poisson sampling

Hey s-zanella, Thanks for your interest and for the well-documented issue. Based on my understanding and your notebook, I deduce that this does not influence the accounting (which is correct)...

Missing argument in MNIST example

Hey danperazzo, Thanks for flagging this! Labelling is at bug, will investigate.