charliebudd

I have tracked down the issue (for me at least). I was making a copy of the model using the python built-in `copy.deepcopy` method. The copy throws the error in...

Seeing the same thing, I had assumed it was a windows issue from the path displayed in the error message. I think windows support would be a nice addition. Even...

I may be able to find some time to help get it working on windows. In fact it seams the main issue here is that the `_substitute` method assumes the...

Got the producer init command to run successfully but it isn't pretty... changes can be seen here https://github.com/charliebudd/klutter/tree/windows-support

I can confirm this behaviour for a 31 dimension output, and that `torch.sum()` works (and I assume `torch.mean()` too). Perhaps a sanity check after the `torch.prod()` and fallback to the...

Yeah, I eventually arrived at the same solution... ```python def gaussian_probability(pi, sigma, mu, target): """Returns the probability of `target` given MoG parameters `sigma` and `mu`. Arguments: sigma (BxGxO): The standard...

What makes this tricky to add in, and what's missing from the implementation in the PR, is the length scale being a function of x. This adds a lot of...

The data used in this work is made from an existing dataset (SCARED) with some additional data (LATTE) which I don't think is released