Change shuffle to do sampling without replacement

`numpy.random.randint` does sampling with replacement. Change to `numpy.random.choice` and set `replace=False` for sampling without replacement.

Author: arun-rangarajan

6.3-advanced-usage-of-recurrent-neural-networks.ipynb

@@ -1157,8 +1157,7 @@
     "    i = min_index + lookback\n",
     "    while 1:\n",
     "        if shuffle:\n",
-    "            rows = np.random.randint(\n",
-    "                min_index + lookback, max_index, size=batch_size)\n",
+    "            rows = np.random.choice(range(min_index + lookback, max_index), size=batch_size, replace=False)\n"
     "        else:\n",
     "            if i + batch_size >= max_index:\n",
     "                i = min_index + lookback\n",