Docs for onecold (#2134)

nathanielvirgo · web-flow · commit 4a2dc27034bd · 2022-12-15T17:05:03.000-05:00
* Update onehot.md

Describe the behaviour of `onecold` when given a probability distribution as input, since it's not what one would expect.

* a better description of `onecold`
diff --git a/docs/src/data/onehot.md b/docs/src/data/onehot.md
@@ -18,7 +18,7 @@ julia> onehot(:c, [:a, :b, :c])
  1
 ```
 
-The inverse is `onecold` (which can take a general probability distribution, as well as just booleans).
+There is also a `onecold` function, which is an inverse of `onehot`. It can also be given an array of numbers instead of booleans, in which case it performs an `argmax`-like operation, returning the label with the highest corresponding weight.
 
 ```jldoctest onehot
 julia> onecold(ans, [:a, :b, :c])

Original file line number	Diff line number	Diff line change
`@@ -18,7 +18,7 @@ julia> onehot(:c, [:a, :b, :c])`
`18`	`18`	`1`
`19`	`19`	```
`20`	`20`
`21`		-The inverse is `onecold` (which can take a general probability distribution, as well as just booleans).
	`21`	+There is also a `onecold` function, which is an inverse of `onehot`. It can also be given an array of numbers instead of booleans, in which case it performs an `argmax`-like operation, returning the label with the highest corresponding weight.
`22`	`22`
`23`	`23`	```jldoctest onehot
`24`	`24`	`julia> onecold(ans, [:a, :b, :c])`