Artur Chakhvadze

Why do you store separate prior mean and covariance for each sample within a batch? https://github.com/chaiyujin/glow-pytorch/blob/487a6b149295f4ec4b36e408f63604c593ff2031/glow/models.py#L199 It doesn't make sense to me. As I understand, you sample each element of...

Semi-transparent shapes and automatic flood fill.

2

### Add an option to make drawn shapes semi-transparent by specifying a keyword argument `opacity`, like this: ```julia ImageDraw.draw!(image, ImageDraw.Cross(ImageDraw.Point(x, y), 3), color, opacity=0.2) ``` ### Also change the polygon...

A way to assign a variable to a field with the same name

1

**Is your feature request related to a problem? Please describe.** Here is an example config: ``` let learning_rate = 1e-3 in { optimizer = { learning_rate = learning_rate } }...

eqx.filter breaks type annotations and post init checks

3

Since exq.filter replaces non-matching leaves with None, all non-static module attributes should be annotated as T | None. Also, checks in `__post_init__` should assume that attributes can be None. This...