Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix link and typo in developer_guide.md #7219

Merged
merged 2 commits into from
Apr 6, 2024
Merged

Fix link and typo in developer_guide.md #7219

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
9f032c9
Fix link and typo in developer_guide.md
pipme Mar 26, 2024
d9efe32
Update docs/source/contributing/developer_guide.md
pipme Mar 28, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions docs/source/contributing/developer_guide.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -542,10 +542,10 @@ There are some example in the ``CompoundStep`` doc:
The base class for most MCMC sampler (except SMC) is in [ArrayStep](https://github.com/pymc-devs/pymc/blob/main/pymc/step_methods/arraystep.py).
You can see that the ``step.step()`` is mapping the ``point`` into an array, and call ``self.astep()``, which is an array in, array out function.
A PyMC model compiles a conditional logp/dlogp function that replace the input RVs with a shared 1D tensor (flatten and stack view of the original RVs).
And the transition kernel (i.e., ``.astep()``) takes an array as input and output an array.
See for example in the [MH sampler](https://github.com/pymc-devs/pymc/blob/6d07591962a6c135640a3c31903eba66b34e71d8/pymc/step_methods/metropolis.py#L139-L173).
And the transition kernel (i.e., ``.astep()``) takes an array as input and outputs an array.
For example, see the [MH sampler](https://github.com/pymc-devs/pymc/blob/89f6fcf751774fb50016561dc448a87fba7ed3aa/pymc/step_methods/metropolis.py#L235-L289).

This is of course very different compare to the transition kernel in e.g. TFP, which is a tenor in tensor out function.
This is of course very different compared to the transition kernel in e.g. TFP, which is a tenor in tensor out function.
Moreover, transition kernels in TFP do not flatten the tensors, see eg docstring of [tensorflow\_probability/python/mcmc/random\_walk\_metropolis.py](https://github.com/tensorflow/probability/blob/master/tensorflow_probability/python/mcmc/random_walk_metropolis.py):

```python
Expand Down
Loading