Allow rendering of semisupervised models

[fritzo]

Addresses this forum post

This uses color gradients. It doesn't look perfect, but it's a good first pass. For example:

def model(x, y=None):
    with pyro.plate("N", 2):
        z = pyro.sample("z", dist.Normal(0, 1))
        y = pyro.sample("y", dist.Normal(0, 1), obs=y)
        pyro.sample("x", dist.Normal(y + z, 1), obs=x)

pyro.render_model(
    model,
    model_kwargs=[
        {"x": torch.zeros(2)},
        {"x": torch.zeros(2), "y": torch.zeros(2)},
    ]
)

Tested

• added a new section to the rendering tutorial

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[fehiepsi]

This looks pretty cool!! I'll take a pass soon

[fehiepsi]

The logic is a little bit complicated to follow so it would be nice to add a test for _deep_merge.

[fehiepsi]

Thanks, Fritz!