Fix the pre-process branch (#147)

Author: alan-cooney

docs/content/pre-process-datasets.ipynb

@@ -214,7 +214,8 @@
     "        # Get just the first few (each file is 11GB so this should be enough for a large dataset)\n",
     "        data_files=[\n",
     "            \"00.jsonl.zst\",\n",
-    "            \"01.jsonl.zst\" \"02.jsonl.zst\",\n",
+    "            \"01.jsonl.zst\",\n",
+    "            \"02.jsonl.zst\",\n",
     "            \"03.jsonl.zst\",\n",
     "            \"04.jsonl.zst\",\n",
     "            \"05.jsonl.zst\",\n",
@@ -226,7 +227,8 @@
     "        tokenizer_name=\"EleutherAI/gpt-neox-20b\",\n",
     "        data_files=[\n",
     "            \"00.jsonl.zst\",\n",
-    "            \"01.jsonl.zst\" \"02.jsonl.zst\",\n",
+    "            \"01.jsonl.zst\",\n",
+    "            \"02.jsonl.zst\",\n",
     "            \"03.jsonl.zst\",\n",
     "            \"04.jsonl.zst\",\n",
     "            \"05.jsonl.zst\",\n",