From 9d50728bb59b25f31a533803a289bb289cbef2c6 Mon Sep 17 00:00:00 2001
From: raja <69231336+Rajasimhareddybolla@users.noreply.github.com>
Date: Sat, 31 Aug 2024 17:30:41 +0530
Subject: [PATCH] Update video_classification.ipynb

Align Colab notebook with official Hugging Face documentation for video classification

- Identified missing code sections in `transformers_doc/en/tensorflow/video_classification.ipynb` that are correctly present in the official Hugging Face documentation.
- Updated the Colab notebook to include these missing code snippets, ensuring consistency between the notebook and the [Hugging Face Video Classification Docs](https://huggingface.co/docs/transformers/en/tasks/video_classification).
- This update fixes discrepancies that could lead to confusion or errors for users following the notebook in Colab, enhancing the reliability of the provided resources.
- Verified the changes to ensure the notebook runs as expected with the corrected code.
---
 .../en/tensorflow/video_classification.ipynb  | 51 ++++++++++++++++++-
 1 file changed, 50 insertions(+), 1 deletion(-)

diff --git a/transformers_doc/en/tensorflow/video_classification.ipynb b/transformers_doc/en/tensorflow/video_classification.ipynb
index 87697b86..97638100 100644
--- a/transformers_doc/en/tensorflow/video_classification.ipynb
+++ b/transformers_doc/en/tensorflow/video_classification.ipynb
@@ -149,6 +149,51 @@
     "        ...\n",
     "```\n",
     "\n",
+    "You can then count the number of total videos."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pathlib\n",
+    "dataset_root_path = \"UCF101_subset\"\n",
+    "dataset_root_path = pathlib.Path(dataset_root_path)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "video_count_train = len(list(dataset_root_path.glob(\"train/*/*.avi\")))\n",
+    "video_count_val = len(list(dataset_root_path.glob(\"val/*/*.avi\")))\n",
+    "video_count_test = len(list(dataset_root_path.glob(\"test/*/*.avi\")))\n",
+    "video_total = video_count_train + video_count_val + video_count_test\n",
+    "print(f\"Total videos: {video_total}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "all_video_file_paths = (\n",
+    "    list(dataset_root_path.glob(\"train/*/*.avi\"))\n",
+    "    + list(dataset_root_path.glob(\"val/*/*.avi\"))\n",
+    "    + list(dataset_root_path.glob(\"test/*/*.avi\"))\n",
+    " )\n",
+    "all_video_file_paths[:5]"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
     "The (`sorted`) video paths appear like so:\n",
     "\n",
     "```bash\n",
@@ -784,7 +829,11 @@
    ]
   }
  ],
- "metadata": {},
+ "metadata": {
+  "language_info": {
+   "name": "python"
+  }
+ },
  "nbformat": 4,
  "nbformat_minor": 4
 }