Sasha Rush

Yeah, there will be a writeup with clear description, but it is still being worked out.

This should be 2 but is failing `circle(1).get_envelope()(2 * unit_x)` . All the logic is in envelope `apply_transform` so my guess that is not handling magnitude properly. Yes, I think...

Can you send the full error with stack trace? Looks like it is something wrong with our code. `expand(torch.LongTensor{[1, 5]}, size=[5])` is wrong, but I don't know where that occurs.

multitrack

It is something we are thinking about ...

I think you installed an old version of `chalk`. Can you make sure you install the one from my branch (command is in the notebook).

Ah thanks. This is a bug with one of the chalk dependencies. I will fix it there. Sorry about that.

Okay, this issue should be fixed now. Could you try it again?

@GdMacmillan can you confirm that you installed Chalk from GitHub, as in the first notebook cell?

Hi @jheek Thanks for the response. Yes, I setup seqinternal to have a carry and a null scan outputs. When I say batching, I mean I want my `setup` method...