Colin Carroll
Colin Carroll
Hi! This is a useful project, but appears to fail on generators. Namely, this function: ``` import time from timeout_decorator import timeout @timeout(seconds=0.5) def sleeper(s): for j in range(10): time.sleep(s)...
Right now, running ```bash $ python -m pydocstyle --convention=numpy pymc/ | wc -l 3090 ``` Shows 3090. It would be a bit much to try to fix all this at...
Apparently this can give us very close control over drawing, binning, and aggregating. This does not look like the right tutorial, but might be close. https://matplotlib.org/users/artists.html
@ahartikainen suggested adding unconstrained variables to the schema somewhere. It seems like they could go 1. Make new `unconstrained_posterior` group Good: - Explicit is better than implicit: it is clear...
In both `azure-pipelines.yml` and `scripts/lint.sh`, the `pydocstyle` check is commented out. It would be great for documentation to start to catch up to development. I started trying, but need more...
As we grow the example gallery, we should probably also check out performance compared to a direct implementation in TFP, and make sure PyMC4 is not adding too much (any?)...
Related to #171, exposing more TFP step methods through the API, and automatically assigning a reasonable sampler to each variable. PyMC3 uses a `competence` attribute to assign this.
No mentions of divergences so far, which are maybe the most common problem with modern samplers (or benefit, since they give obvious clues to deficiencies). I also have a few...
This also makes `from_dict` work for nested dictionaries. I joined the keys with double underscores because using a `.` would break attribute access (like, it would still work, but you'd...
In certain cases, if an element of `var_names` is not in the dataset, should we should throw an exception? Right not, it will plot any var_names it can find, and...