Gemma3n Audio Transcription Ollama

AudioTranscriptionGemma3n.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "ce742a3a",
   "metadata": {},
   "source": [
    "# Audio transcription with Gemma3n\n",
    "\n",
    "https://ai.google.dev/gemma/docs/core/huggingface_inference#audio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "ea2d3e06",
   "metadata": {
    "language_info": {
     "name": "polyglot-notebook"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div><div></div><div></div><div><strong>Installed Packages</strong><ul><li><span>Microsoft.Extensions.AI, 9.6.0</span></li><li><span>OllamaSharp, 5.2.3</span></li></ul></div></div>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#r \"nuget: OllamaSharp, 5.2.3\"\n",
    "#r \"nuget: Microsoft.Extensions.AI, 9.6.0\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "37d83a68",
   "metadata": {
    "language_info": {
     "name": "polyglot-notebook"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "using OllamaSharp;\n",
    "using Microsoft.Extensions.AI;"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8efc0d32",
   "metadata": {
    "language_info": {
     "name": "polyglot-notebook"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "List<AIContent> messages = \n",
    "    [\n",
    "        new TextContent(\"Transcribe this audio and complete the statement\"),\n",
    "        new UriContent(\"https://ai.google.dev/gemma/docs/audio/roses-are.wav\", \"audio/wav\")\n",
    "    ];"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8ebd5c3f",
   "metadata": {},
   "source": [
    "## Ollama\n",
    "\n",
    "NOTE: As of 06-2025, only text generation works with Gemma 3n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "118d96a8",
   "metadata": {
    "language_info": {
     "name": "polyglot-notebook"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "IChatClient ollamaClient = new OllamaApiClient(\"http://localhost:11434\", \"gemma3n\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "054fa1c5",
   "metadata": {
    "language_info": {
     "name": "polyglot-notebook"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [],
   "source": [
    "var ollamaRes = await ollamaClient.GetResponseAsync(new ChatMessage(ChatRole.User,messages));"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "0764cce6",
   "metadata": {
    "language_info": {
     "name": "polyglot-notebook"
    },
    "polyglot_notebook": {
     "kernelName": "csharp"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Please **provide me with the audio!** I need the audio file or a link to the audio to be able to transcribe it and complete the statement. \n",
       "\n",
       "Once you provide the audio, I will:\n",
       "\n",
       "1. **Transcribe the audio:** I will convert the spoken words into text.\n",
       "2. **Identify the statement:** I will look for a statement that is being made or implied in the audio.\n",
       "3. **Complete the statement:**  I will fill in any missing words or phrases to make the statement grammatically correct and meaningful.\n",
       "\n",
       "\n",
       "\n",
       "**Just paste the audio file here, or give me a link to it, and I'll get started!**\n",
       "\n",
       "\n",
       "\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ollamaRes.Text"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".NET (C#)",
   "language": "C#",
   "name": ".net-csharp"
  },
  "language_info": {
   "name": "polyglot-notebook"
  },
  "polyglot_notebook": {
   "kernelInfo": {
    "defaultKernelName": "csharp",
    "items": [
     {
      "aliases": [],
      "languageName": "csharp",
      "name": "csharp"
     }
    ]
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}