Created using Colaboratory

keras_tokenizer_fix.ipynb

@@ -53,6 +53,7 @@
         "tk = Tokenizer(num_words=(n_words+1), filters=string.punctuation, lower=True, split=' ', char_level=False, oov_token='<unk>')\n",
         "texts = [\"my name is far faraway asdasd\", \"my name is\",\"your name is\"]\n",
         "tk.fit_on_texts(texts)\n",
+        "## NOTE: You can fit chunks of data, it will work same way. When having big data don't need to load all files\n",
         "\n",
         "print('Original text: %s\\n'%texts)\n",
         "print('Only use top %s words\\n\\n'%n_words)\n",