Typo fix in quantum_data.ipynb (tensorflow#433)

docs/tutorials/quantum_data.ipynb

@@ -850,7 +850,7 @@
     "\n",
     "1. It's very unlikely that the quantum models of today will beat classical model performance on classical data. Especially on today's classical datasets that can have upwards of a million datapoints.\n",
     "\n",
-    "2. Just because the data might come from a hard to classically simulate quantum circuit, doesn't nescessarily make the data hard to learn for a classical model.\n",
+    "2. Just because the data might come from a hard to classically simulate quantum circuit, doesn't necessarily make the data hard to learn for a classical model.\n",
     "\n",
     "3. Datasets (ultimately quantum in nature) that are easy for quantum models to learn and hard for classical models to learn do exist, regardless of model architecture or training algorithms used."
    ]