Batch transcribe audio/video files using Groq Whisper

whisper_batch_groq.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "8c39022e-4551-4167-ad56-9a39da484e38",
   "metadata": {},
   "outputs": [],
   "source": [
    "from groq import Groq\n",
    "client = Groq()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "3d548229-26b0-464f-98da-77ddc3b8d5f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "filename = \"ElevenLabs_Eastend_Steve_2.mp3\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "35d8827f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " When suddenly, there's this strange sucking noise. It's Armstrong! He's only brought his own food along in a liquid form in a belt pouch and started sucking it noisily from a straw. I should say, Oi Armstrong, no! You're not on your lunar module now. Pick up your knife and fork and eat your finders chicken Italians in a terrestrial fashion.\n"
     ]
    }
   ],
   "source": [
    "with open(filename, \"rb\") as file:\n",
    "    transcription = client.audio.transcriptions.create(\n",
    "      file=(filename, file.read()),\n",
    "      model=\"whisper-large-v3-turbo\", # distil-whisper-large-v3-en or whisper-large-v3\n",
    "      response_format=\"verbose_json\",\n",
    "    )\n",
    "    print(transcription.text)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "c8b407d7-35f8-42b3-bc2d-96c5b0e8400f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Transcription(text=\" When suddenly, there's this strange sucking noise. It's Armstrong! He's only brought his own food along in a liquid form in a belt pouch and started sucking it noisily from a straw. I should say, Oi Armstrong, no! You're not on your lunar module now. Pick up your knife and fork and eat your finders chicken Italians in a terrestrial fashion.\", task='transcribe', language='English', duration=23.3, segments=[{'id': 0, 'seek': 0, 'start': 0, 'end': 4.38, 'text': \" When suddenly, there's this strange sucking noise. It's Armstrong!\", 'tokens': [50365, 1133, 5800, 11, 456, 311, 341, 5861, 38669, 5658, 13, 467, 311, 36100, 0, 50584], 'temperature': 0, 'avg_logprob': -0.2622168, 'compression_ratio': 1.5336323, 'no_speech_prob': 1.9729613e-12}, {'id': 1, 'seek': 0, 'start': 4.38, 'end': 9.08, 'text': \" He's only brought his own food along in a liquid form in a belt pouch\", 'tokens': [50584, 634, 311, 787, 3038, 702, 1065, 1755, 2051, 294, 257, 6553, 1254, 294, 257, 10750, 27781, 50819], 'temperature': 0, 'avg_logprob': -0.2622168, 'compression_ratio': 1.5336323, 'no_speech_prob': 1.9729613e-12}, {'id': 2, 'seek': 0, 'start': 9.08, 'end': 11.96, 'text': ' and started sucking it noisily from a straw.', 'tokens': [50819, 293, 1409, 38669, 309, 572, 271, 953, 490, 257, 10099, 13, 50963], 'temperature': 0, 'avg_logprob': -0.2622168, 'compression_ratio': 1.5336323, 'no_speech_prob': 1.9729613e-12}, {'id': 3, 'seek': 0, 'start': 11.96, 'end': 14.38, 'text': ' I should say, Oi Armstrong, no!', 'tokens': [50963, 286, 820, 584, 11, 31610, 36100, 11, 572, 0, 51084], 'temperature': 0, 'avg_logprob': -0.2622168, 'compression_ratio': 1.5336323, 'no_speech_prob': 1.9729613e-12}, {'id': 4, 'seek': 0, 'start': 14.38, 'end': 16.7, 'text': \" You're not on your lunar module now.\", 'tokens': [51084, 509, 434, 406, 322, 428, 32581, 10088, 586, 13, 51200], 'temperature': 0, 'avg_logprob': -0.2622168, 'compression_ratio': 1.5336323, 'no_speech_prob': 1.9729613e-12}, {'id': 5, 'seek': 0, 'start': 16.7, 'end': 23.12, 'text': ' Pick up your knife and fork and eat your finders chicken Italians in a terrestrial fashion.', 'tokens': [51200, 14129, 493, 428, 7976, 293, 17716, 293, 1862, 428, 915, 433, 4662, 43620, 294, 257, 1796, 34539, 6700, 13, 51521], 'temperature': 0, 'avg_logprob': -0.2622168, 'compression_ratio': 1.5336323, 'no_speech_prob': 1.9729613e-12}], x_groq={'id': 'req_01jg2f5xf6e8esvszdn6qrxkm3'})\n"
     ]
    }
   ],
   "source": [
    "with open(filename, \"rb\") as file:\n",
    "    transcription = client.audio.transcriptions.create(\n",
    "      file=(filename, file.read()),\n",
    "      model=\"whisper-large-v3-turbo\", # distil-whisper-large-v3-en or whisper-large-v3\n",
    "      response_format=\"verbose_json\",\n",
    "      #prompt=\"Specify context or spelling\",  # Optional\n",
    "      #language=\"es\",  # Optional\n",
    "      #temperature=0.0  # Optional\n",
    "    )\n",
    "    #print(transcription.text)\n",
    "    print(transcription)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "d5f07a6a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'id': 0,\n",
       "  'seek': 0,\n",
       "  'start': 0,\n",
       "  'end': 4.38,\n",
       "  'text': \" When suddenly, there's this strange sucking noise. It's Armstrong!\",\n",
       "  'tokens': [50365,\n",
       "   1133,\n",
       "   5800,\n",
       "   11,\n",
       "   456,\n",
       "   311,\n",
       "   341,\n",
       "   5861,\n",
       "   38669,\n",
       "   5658,\n",
       "   13,\n",
       "   467,\n",
       "   311,\n",
       "   36100,\n",
       "   0,\n",
       "   50584],\n",
       "  'temperature': 0,\n",
       "  'avg_logprob': -0.2622168,\n",
       "  'compression_ratio': 1.5336323,\n",
       "  'no_speech_prob': 1.9729613e-12},\n",
       " {'id': 1,\n",
       "  'seek': 0,\n",
       "  'start': 4.38,\n",
       "  'end': 9.08,\n",
       "  'text': \" He's only brought his own food along in a liquid form in a belt pouch\",\n",
       "  'tokens': [50584,\n",
       "   634,\n",
       "   311,\n",
       "   787,\n",
       "   3038,\n",
       "   702,\n",
       "   1065,\n",
       "   1755,\n",
       "   2051,\n",
       "   294,\n",
       "   257,\n",
       "   6553,\n",
       "   1254,\n",
       "   294,\n",
       "   257,\n",
       "   10750,\n",
       "   27781,\n",
       "   50819],\n",
       "  'temperature': 0,\n",
       "  'avg_logprob': -0.2622168,\n",
       "  'compression_ratio': 1.5336323,\n",
       "  'no_speech_prob': 1.9729613e-12},\n",
       " {'id': 2,\n",
       "  'seek': 0,\n",
       "  'start': 9.08,\n",
       "  'end': 11.96,\n",
       "  'text': ' and started sucking it noisily from a straw.',\n",
       "  'tokens': [50819,\n",
       "   293,\n",
       "   1409,\n",
       "   38669,\n",
       "   309,\n",
       "   572,\n",
       "   271,\n",
       "   953,\n",
       "   490,\n",
       "   257,\n",
       "   10099,\n",
       "   13,\n",
       "   50963],\n",
       "  'temperature': 0,\n",
       "  'avg_logprob': -0.2622168,\n",
       "  'compression_ratio': 1.5336323,\n",
       "  'no_speech_prob': 1.9729613e-12},\n",
       " {'id': 3,\n",
       "  'seek': 0,\n",
       "  'start': 11.96,\n",
       "  'end': 14.38,\n",
       "  'text': ' I should say, Oi Armstrong, no!',\n",
       "  'tokens': [50963, 286, 820, 584, 11, 31610, 36100, 11, 572, 0, 51084],\n",
       "  'temperature': 0,\n",
       "  'avg_logprob': -0.2622168,\n",
       "  'compression_ratio': 1.5336323,\n",
       "  'no_speech_prob': 1.9729613e-12},\n",
       " {'id': 4,\n",
       "  'seek': 0,\n",
       "  'start': 14.38,\n",
       "  'end': 16.7,\n",
       "  'text': \" You're not on your lunar module now.\",\n",
       "  'tokens': [51084, 509, 434, 406, 322, 428, 32581, 10088, 586, 13, 51200],\n",
       "  'temperature': 0,\n",
       "  'avg_logprob': -0.2622168,\n",
       "  'compression_ratio': 1.5336323,\n",
       "  'no_speech_prob': 1.9729613e-12},\n",
       " {'id': 5,\n",
       "  'seek': 0,\n",
       "  'start': 16.7,\n",
       "  'end': 23.12,\n",
       "  'text': ' Pick up your knife and fork and eat your finders chicken Italians in a terrestrial fashion.',\n",
       "  'tokens': [51200,\n",
       "   14129,\n",
       "   493,\n",
       "   428,\n",
       "   7976,\n",
       "   293,\n",
       "   17716,\n",
       "   293,\n",
       "   1862,\n",
       "   428,\n",
       "   915,\n",
       "   433,\n",
       "   4662,\n",
       "   43620,\n",
       "   294,\n",
       "   257,\n",
       "   1796,\n",
       "   34539,\n",
       "   6700,\n",
       "   13,\n",
       "   51521],\n",
       "  'temperature': 0,\n",
       "  'avg_logprob': -0.2622168,\n",
       "  'compression_ratio': 1.5336323,\n",
       "  'no_speech_prob': 1.9729613e-12}]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "segments = (transcription.segments)\n",
    "segments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "b381c88c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "\n",
    "def convert_seconds_to_hms(seconds):\n",
    "    hours, remainder = divmod(seconds, 3600)\n",
    "    minutes, seconds = divmod(remainder, 60)\n",
    "    milliseconds = math.floor((seconds % 1) * 1000)\n",
    "    output = f\"{int(hours):02}:{int(minutes):02}:{int(seconds):02},{milliseconds:03}\"\n",
    "    return output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "c8ff6a57",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[00:00:00,000 -> 00:00:04,379]  When suddenly, there's this strange sucking noise. It's Armstrong!\n",
      "[00:00:04,379 -> 00:00:09,080]  He's only brought his own food along in a liquid form in a belt pouch\n",
      "[00:00:09,080 -> 00:00:11,960]  and started sucking it noisily from a straw.\n",
      "[00:00:11,960 -> 00:00:14,380]  I should say, Oi Armstrong, no!\n",
      "[00:00:14,380 -> 00:00:16,699]  You're not on your lunar module now.\n",
      "[00:00:16,699 -> 00:00:23,120]  Pick up your knife and fork and eat your finders chicken Italians in a terrestrial fashion.\n"
     ]
    }
   ],
   "source": [
    "def format_segments_with_timecode(segments):\n",
    "    formatted_segments = []\n",
    "    for segment in segments:\n",
    "        start_tc = convert_seconds_to_hms(segment['start'])\n",
    "        end_tc = convert_seconds_to_hms(segment['end'])\n",
    "        formatted_line = f\"[{start_tc} -> {end_tc}] {segment['text']}\"\n",
    "        formatted_segments.append(formatted_line)\n",
    "    return formatted_segments\n",
    "\n",
    "# Print all segments\n",
    "for formatted_line in format_segments_with_timecode(segments):\n",
    "    print(formatted_line)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "77a20d69",
   "metadata": {},
   "outputs": [],
   "source": [
    "# save transcription as .srt file\n",
    "count = 0\n",
    "with open(\"subs.srt\", 'w') as f: # define subtitle file name here\n",
    "      for segment in segments:\n",
    "        count +=1\n",
    "        duration = f\"{convert_seconds_to_hms(segment['start'])} --> {convert_seconds_to_hms(segment['end'])}\\n\"\n",
    "        text = f\"{segment['text'].lstrip()}\\n\\n\"\n",
    "        f.write(f\"{count}\\n{duration}{text}\")  # Write formatted string to the file"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.13.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}