diff --git a/PrepareDataset.ipynb b/PrepareDataset.ipynb
index 2cee886..6a3edd0 100644
--- a/PrepareDataset.ipynb
+++ b/PrepareDataset.ipynb
@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 9,
    "id": "b0992ed3-6fb2-4f7c-b5b2-6c1aaf632b18",
    "metadata": {},
    "outputs": [],
@@ -11,21 +11,24 @@
     "import os\n",
     "import shutil\n",
     "from tqdm import tqdm\n",
+    "import torchaudio\n",
+    "import torchaudio.transforms as T\n",
+    "import glob\n",
     "\n",
     "SRC_PATH = r'F:\\Hackathon_Dataset'\n",
-    "DST_PATH = r'C:\\Users\\amitli\\Repo\\SpeakerVerification2024\\Dataset_SR32000'\n",
+    "DST_PATH = r'C:\\Users\\amitli\\Repo\\SpeakerVerification2024\\Dataset_SR16000'\n",
     "df       = pd.read_csv(r'F:\\Hackathon_Dataset\\Hackathon_final_without_outliers.csv')\n",
-    "df       = df[df.language == 'arabic']\n"
+    "df       = df[df.language       == 'arabic']\n",
+    "df       = df[df.hackathon_type == 'tactic']"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": 10,
    "id": "23bedaf6-6695-42b4-99f4-e632d0a77ef7",
    "metadata": {},
    "outputs": [],
    "source": [
-    "\n",
     "total_collect = 0\n",
     "l_speaker     = []\n",
     "l_fname       = []\n",
@@ -45,44 +48,19 @@
     "\n",
     "        l_speaker.append(speaker)\n",
     "        l_fname  .append(file_name)        \n",
-    "        shutil.copy(rf'{SRC_PATH}/{file_name}', rf'{DST_PATH}/{file_name}')\n",
+    "        #shutil.copy(rf'{SRC_PATH}/{file_name}', rf'{DST_PATH}/{file_name}')\n",
+    "\n",
+    "        # Load the audio file    \n",
+    "        waveform, sample_rate = torchaudio.load(rf'{SRC_PATH}/{file_name}')\n",
+    "        if sample_rate != 16000:        \n",
+    "            resampler = T.Resample(orig_freq=sample_rate, new_freq=16000)\n",
+    "            waveform  = resampler(waveform)\n",
+    "        torchaudio.save(rf'{DST_PATH}/{file_name}', waveform, 16000) \n",
+    "        \n",
     "\n",
     "df_results = pd.DataFrame({'speaker': l_speaker, 'file_name': l_fname})\n",
     "df_results.to_csv(rf\"{DST_PATH}/speakers.csv\")"
    ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 5,
-   "id": "270f4f53-fd46-4eb9-a05c-a636d34f0873",
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "100%|███████████████████████████████████████████████████████████████████████████████| 100/100 [00:00<00:00, 166.34it/s]\n"
-     ]
-    }
-   ],
-   "source": [
-    "import torchaudio\n",
-    "import torchaudio.transforms as T\n",
-    "import glob\n",
-    "\n",
-    "\n",
-    "DST_PATH_16000 = r'C:\\Users\\amitli\\Repo\\SpeakerVerification2024\\Dataset_SR16000'\n",
-    "l_files = glob.glob(rf'{DST_PATH}\\*.wav')\n",
-    "for file in tqdm(l_files):\n",
-    "\n",
-    "    # Load the audio file    \n",
-    "    waveform, sample_rate = torchaudio.load(file)\n",
-    "    if sample_rate != 16000:        \n",
-    "        resampler = T.Resample(orig_freq=sample_rate, new_freq=16000)\n",
-    "        waveform = resampler(waveform)\n",
-    "\n",
-    "        torchaudio.save(rf'{DST_PATH_16000}\\{os.path.basename(file)}', waveform, 16000) \n"
-   ]
   }
  ],
  "metadata": {