Use numpy.arange in RandomSampler (#10768)

Significant speedup for large datasets:

In [2]: %timeit current_sample(1529*8192)
12.3 s ± 721 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

In [3]: %timeit np_sample(1529*8192)
641 ms ± 6.7 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

python/mxnet/gluon/data/sampler.py

@@ -20,7 +20,7 @@
 """Dataset sampler."""
 __all__ = ['Sampler', 'SequentialSampler', 'RandomSampler', 'BatchSampler']
 
-import random
+import numpy as np
 
 class Sampler(object):
     """Base class for samplers.
@@ -65,8 +65,8 @@ def __init__(self, length):
         self._length = length
 
     def __iter__(self):
-        indices = list(range(self._length))
-        random.shuffle(indices)
+        indices = np.arange(self._length)
+        np.random.shuffle(indices)
         return iter(indices)
 
     def __len__(self):