Merge pull request #15 from pitmonticone/main

Update README.md

README.md

@@ -27,7 +27,7 @@ There is, however, only partial support for multidimensional layers (convolution
 
 ## How it works
 
-There is a distinction between netwoks starting with a layer with fixed-size input (Dense and Recurrent) and networks starting with a layer with variable-size input (Convolutional, Pooling, and functional).
+There is a distinction between networks starting with a layer with fixed-size input (Dense and Recurrent) and networks starting with a layer with variable-size input (Convolutional, Pooling, and functional).
 
 In the former case, just passing a network `m = Chain(...)` to plot works, e.g. `plot(m)`. In the latter case, one needs to pass along an initial input `inp`, or input size `inpsz = size(inp)`, as the second argument, like `plot(m, inp)` or `plot(m, inpsz)`, so that the plot recipe can properly figure out the size of each layer.