Added required doc for categorical cross entropy #870
Conversation
Sentient07
force-pushed
the
crossentropy-doc
branch
from
426d7a3
to
40f58f1
Compare
lasagne/objectives.py
Outdated
| """Computes the categorical cross-entropy between predictions and targets. | ||
| """Computes the categorical cross-entropy between predictions $p_{i,j}$ | ||
| and targets $t_{i,j}$, where $i$ denotes the data point and $j$ | ||
| denotes the class. |
There was a problem hiding this comment.
Sounds good, but you completely messed up the formatting :)
a) the first sentence must be short since it's used for the overview table
b) the text must not be indented otherwise it's assumed to be a code block
c) the math syntax in Sphinx does not use $
Please always compile and look at the documentation when changing docstrings in a nontrivial way: http:https://lasagne.readthedocs.io/en/latest/user/development.html#documentation
So the explanation of p, t, i, j should come after the formula, or in a second sentence before the formula, not in the first sentence.
lasagne/objectives.py
Outdated
| @@ -157,6 +159,8 @@ def categorical_crossentropy(predictions, targets): | |||
| targets : Theano 2D tensor or 1D tensor | |||
| Either targets in [0, 1] matching the layout of `predictions`, or | |||
| a vector of int giving the correct class index per data point. | |||
| In the case of an integer vector argument, each element | |||
| represents the position of the '1' in a 1-of-N encoding. | |||
|
I've pushed a commit with the suggested changes. Here's how the doc looks, |
|
Sorry about that, there was a PEP8 error. Fixed that and committed. Once this PR is LGTM(to @f0k ), I will squash all the commits into one |
f0k
force-pushed
the
crossentropy-doc
branch
from
4ac8555
to
f3dc0bb
Compare
|
Modified the formulation, corrected the punctuation and squashed it. Hope you don't mind, it was easier this way. Thanks again, merging! |
Fixes #591