Merge #1783

1783: Doc update (training.md): fix DataLoader example in Training section r=ToucheSir a=eliascarv

Change from:
```julia
X = rand(28, 28, 60000)
Y = rand(0:9, 60000)
data = DataLoader(X, Y, batchsize=128)
```
To:
```julia
X = rand(28, 28, 60000)
Y = rand(0:9, 60000)
data = DataLoader((X, Y), batchsize=128)
```


Co-authored-by: eliascarv <73039601+eliascarv@users.noreply.github.com>

Author: bors[bot]

docs/src/training/training.md

@@ -102,7 +102,7 @@ Training data can be conveniently  partitioned for mini-batch training using the
 ```julia
 X = rand(28, 28, 60000)
 Y = rand(0:9, 60000)
-data = DataLoader(X, Y, batchsize=128) 
+data = DataLoader((X, Y), batchsize=128) 
 ```
 
 Note that, by default, `train!` only loops over the data once (a single "epoch").