音声の理解

[ai.google.dev で表示] [Run in Google Colab] [Kaggle で実行] [Vertex AI で開く] [GitHub でソースを表示]

Gemma 3n 以降では、プロンプトとワークフローで音声を直接使用できます。音声と話し言葉は、ユーザーの意図を把握し、周囲の世界に関する情報を記録し、解決すべき特定の問題を理解するための豊富なデータソースです。

このガイドでは、Gemma 4 の音声処理機能の概要について説明します。これには、自動音声認識（ASR）、翻訳、一般的な音声理解が含まれます。

このノートブックは T4 GPU で実行されます。

Python パッケージをインストールする

Gemma モデルを実行してリクエストを行うために必要な Hugging Face ライブラリをインストールします。

# Install PyTorch & other libraries
pip install torch accelerate

# Install the transformers library
pip install transformers

モデルを読み込む

transformers ライブラリを使用して、次のコード例に示すように、AutoProcessor クラスと AutoModelForImageTextToText クラスを使用して processor と model のインスタンスを作成します。

MODEL_ID = "google/gemma-4-E2B-it" # @param ["google/gemma-4-E2B-it","google/gemma-4-E4B-it", "google/gemma-4-31B-it", "google/gemma-4-26B-A4B-it"]

from transformers import AutoProcessor, AutoModelForMultimodalLM

model = AutoModelForMultimodalLM.from_pretrained(MODEL_ID, dtype="auto", device_map="auto")
processor = AutoProcessor.from_pretrained(MODEL_ID)

Loading weights:   0%|          | 0/2011 [00:00<?, ?it/s]

音声データ

デジタル音声データには、さまざまな形式と解像度があります。Gemma で使用できる実際の音声形式（MP3 や WAV 形式など）は、音声データをテンソルに変換するために選択するフレームワークによって決まります。Gemma で処理する音声データを準備する際の具体的な考慮事項は次のとおりです。

  • トークン費用: 音声 1 秒あたり、Gemma 4 では 25 トークン（Gemma 3n では 6.25 トークン）。
  • クリップの長さ: 音声は最大 30 秒までサポートされます。
  • 音声チャンネル: 音声データは単一の音声チャンネルとして処理されます。 左右のチャンネルなど、マルチチャンネル音声を使用している場合は、チャンネルを削除するか、音声データを 1 つのチャンネルに結合して、データを 1 つのチャンネルに減らすことを検討してください。
  • 技術的なエンコード:
    • サンプルレート: 32 ミリ秒フレームを使用する 16 kHz。
    • ビット深度: 32 ビット浮動小数点形式。サンプルは [-1, 1] の範囲内で正規化されます。

処理する音声データが入力処理と大きく異なる場合（特にチャネル、サンプルレート、ビット深度に関して）、モデルで処理されるデータ解像度に合わせて音声データをリサンプリングまたはトリミングすることを検討してください。

音声エンコード

高レベル ライブラリ（Hugging Face AutoProcessor など）では音声の前処理が自動的に行われることが多いですが、カスタム エンコードを実装する必要がある場合もあります。

Gemma で使用するために独自のコード実装で音声データをエンコードする場合は、推奨される変換プロセスに従う必要があります。MP3 や WAV エンコード データなど、特定の形式でエンコードされた音声ファイルを操作する場合は、まず ffmpeg などのライブラリを使用してサンプルにデコードする必要があります。データのデコードが完了したら、音声をモノラルチャンネルの 16 kHz float32 波形（範囲 [-1, 1]）に変換します。たとえば、44.1 kHz のステレオ符号付き 16 ビット PCM 整数 WAV ファイルを操作する場合は、次の手順を行います。

  • 音声データを 16 kHz にリサンプリングする
  • 2 つのチャンネルを平均してステレオからモノラルにダウンミックスする
  • int16 から float32 に変換し、32768.0 で割って [-1, 1] の範囲にスケーリングする

音声入力

Gemma 4 E2B と E4B は多言語音声認識用にトレーニングされているため、さまざまな言語の音声入力をテキストに書き起こすことができます。次のコード例は、Hugging Face Transformers を使用して音声ファイルからテキストを文字起こしするようにモデルにプロンプトを表示する方法を示しています。

RESOURCE_URL_PREFIX = "https://raw.githubusercontent.com/google-gemma/cookbook/refs/heads/main/Demos/sample-data/"

messages = [
    {
        "role": "user",
        "content": [
            {"type": "text", "text": "Transcribe the following speech segment in its original language. Follow these specific instructions for formatting the answer:\n* Only output the transcription, with no newlines.\n* When transcribing numbers, write the digits, i.e. write 1.7 and not one point seven, and write 3 instead of three."},
            #{"type": "text", "text": "Transcribe the following speech segment in English into English text. Follow these specific instructions for formatting the answer:\n* Only output the transcription, with no newlines.\n* When transcribing numbers, write the digits, i.e. write 1.7 and not one point seven, and write 3 instead of three."},
            {"type": "audio", "audio": f"{RESOURCE_URL_PREFIX}journal1.wav"},
        ]
    }
]

input_ids = processor.apply_chat_template(
        messages,
        add_generation_prompt=True,
        tokenize=True, return_dict=True,
        return_tensors="pt",
)
input_ids = input_ids.to(model.device, dtype=model.dtype)

outputs = model.generate(**input_ids, max_new_tokens=64)

text = processor.batch_decode(
    outputs,
    skip_special_tokens=False,
    clean_up_tokenization_spaces=False
)
print(text[0])

<bos><|turn>user
Transcribe the following speech segment in its original language. Follow these specific instructions for formatting the answer:

* Only output the transcription, with no newlines.
* When transcribing numbers, write the digits, i.e. write 1.7 and not one point seven, and write 3 instead of three.<|audio><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><audio|><turn|>
<|turn>model
I woke up early today feeling really fresh the morning light was beautiful and I enjoyed a nice cup of coffee<turn|>

messages = [
    {
        "role": "user",
        "content": [
            {"type": "text", "text": "Give me a concise overview of these audio files."},
            {"type": "text", "text": "journal1:"},
            {"type": "audio", "audio": f"{RESOURCE_URL_PREFIX}journal1.wav"},
            {"type": "text", "text": "journal2:"},
            {"type": "audio", "audio": f"{RESOURCE_URL_PREFIX}journal2.wav"},
            {"type": "text", "text": "journal3:"},
            {"type": "audio", "audio": f"{RESOURCE_URL_PREFIX}journal3.wav"},
            {"type": "text", "text": "journal4:"},
            {"type": "audio", "audio": f"{RESOURCE_URL_PREFIX}journal4.wav"},
            {"type": "text", "text": "journal5:"},
            {"type": "audio", "audio": f"{RESOURCE_URL_PREFIX}journal5.wav"},
        ]
    }
]

input_ids = processor.apply_chat_template(
        messages,
        add_generation_prompt=True,
        tokenize=True, return_dict=True,
        return_tensors="pt",
)
input_ids = input_ids.to(model.device, dtype=model.dtype)

outputs = model.generate(**input_ids, max_new_tokens=1024)

text = processor.batch_decode(
    outputs,
    skip_special_tokens=False,
    clean_up_tokenization_spaces=False
)
print(text[0])

<bos><|turn>user
Give me a concise overview of these audio files.journal1:<|audio><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><audio|>journal2:<|audio><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><audio|>journal3:<|audio><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><audio|>journal4:<|audio><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><audio|>journal5:<|audio><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><audio|><turn|>
<|turn>model
Here is a concise overview of the audio files:

**Journal 1:** The speaker felt refreshed, enjoyed a morning ride, a cup of coffee, and was generally happy.

**Journal 2:** The speaker spent the afternoon at the park, which was a perfect day for a walk, and enjoyed watching the cherry blossoms.

**Journal 3:** The speaker finished the day with a good book, feeling grateful for simple moments and ready for more.

**Journal 4:** The speaker returned from work, admiring the sunset, and enjoyed a clear view from the train.

**Journal 5:** The speaker had a great lunch with an old friend, enjoyed catching up, and felt happy about the day.<turn|>

自動音声翻訳

Gemma 4 E2B と E4B は多言語音声翻訳タスク用にトレーニングされているため、話し言葉を別の言語に直接翻訳できます。次のコード例は、Hugging Face Transformers を使用して音声をテキストに翻訳するようにモデルにプロンプトを表示する方法を示しています。

messages = [
    {
        "role": "user",
        "content": [
            {"type": "text", "text": "Transcribe the following speech segment in English, then translate it into Korean. When formatting the answer, first output the transcription in English, then one newline, then output the string 'Korean: ', then the translation in Korean."},
            {"type": "audio", "audio": "https://ai.google.dev/gemma/docs/audio/roses-are.wav"},
        ]
    }
]

input_ids = processor.apply_chat_template(
        messages,
        add_generation_prompt=True,
        tokenize=True, return_dict=True,
        return_tensors="pt",
)
input_ids = input_ids.to(model.device, dtype=model.dtype)

outputs = model.generate(**input_ids, max_new_tokens=64)

text = processor.batch_decode(
    outputs,
    skip_special_tokens=False,
    clean_up_tokenization_spaces=False
)
print(text[0])

<bos><|turn>user
Transcribe the following speech segment in English, then translate it into Korean. When formatting the answer, first output the transcription in English, then one newline, then output the string 'Korean: ', then the translation in Korean.<|audio><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><audio|><turn|>
<|turn>model
Roses are red, violets are blue.
Korean: 장미는 빨갛고, 제비꽃은 파랗다.<turn|>

自動音声翻訳 / 自動音声認識

実際に試してみる

pip install ipywebrtc

丸いボタンを押して話し始めます。終了したら、もう一度丸いボタンをクリックします。ウィジェットは、キャプチャした内容の再生をすぐに開始します。

from google.colab import output
output.enable_custom_widget_manager()

from ipywebrtc import AudioRecorder, CameraStream

camera = CameraStream(constraints={'audio': True,'video':False})
recorder = AudioRecorder(stream=camera)
recorder

AudioRecorder(audio=Audio(value=b'', format='webm'), stream=CameraStream(constraints={'audio': True, 'video': …

PyTorch が理解できる wav 形式に webm ファイルを変換します。

with open('/content/recording.webm', 'wb') as f:
    f.write(recorder.audio.value)
!ffmpeg -i /content/recording.webm /content/recording.wav -y

ffmpeg version 4.4.2-0ubuntu0.22.04.1 Copyright (c) 2000-2021 the FFmpeg developers
  built with gcc 11 (Ubuntu 11.2.0-19ubuntu1)
  configuration: --prefix=/usr --extra-version=0ubuntu0.22.04.1 --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libdav1d --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librabbitmq --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzimg --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --enable-pocketsphinx --enable-librsvg --enable-libmfx --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared
  libavutil      56. 70.100 / 56. 70.100
  libavcodec     58.134.100 / 58.134.100
  libavformat    58. 76.100 / 58. 76.100
  libavdevice    58. 13.100 / 58. 13.100
  libavfilter     7.110.100 /  7.110.100
  libswscale      5.  9.100 /  5.  9.100
  libswresample   3.  9.100 /  3.  9.100
  libpostproc    55.  9.100 / 55.  9.100
Input #0, matroska,webm, from '/content/recording.webm':
  Metadata:
    encoder         : Chrome
  Duration: 00:00:04.02, start: 0.000000, bitrate: 131 kb/s
  Stream #0:0(eng): Audio: opus, 48000 Hz, mono, fltp (default)
Stream mapping:
  Stream #0:0 -> #0:0 (opus (native) -> pcm_s16le (native))
Press [q] to stop, [?] for help
Output #0, wav, to '/content/recording.wav':
  Metadata:
    ISFT            : Lavf58.76.100
  Stream #0:0(eng): Audio: pcm_s16le ([1][0][0][0] / 0x0001), 48000 Hz, mono, s16, 768 kb/s (default)
    Metadata:
      encoder         : Lavc58.134.100 pcm_s16le
size=     383kB time=00:00:04.01 bitrate= 779.7kbits/s speed=60.6x    
video:0kB audio:382kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.019914%

ASR

messages = [{
  "role": "user",
  "content": [
    {"type": "text", "text": "Transcribe the following speech segment in its original language. Follow these specific instructions for formatting the answer:\n* Only output the transcription, with no newlines.\n* When transcribing numbers, write the digits, i.e. write 1.7 and not one point seven, and write 3 instead of three."},
    {"type": "audio", "audio": "/content/recording.wav"},
  ]
}]

input_ids = processor.apply_chat_template(
        messages,
        add_generation_prompt=True,
        tokenize=True, return_dict=True,
        return_tensors="pt",
)
input_ids = input_ids.to(model.device, dtype=model.dtype)

outputs = model.generate(**input_ids, max_new_tokens=64)

text = processor.batch_decode(
    outputs,
    skip_special_tokens=False,
    clean_up_tokenization_spaces=False
)
print(text[0])

<bos><|turn>user
Transcribe the following speech segment in its original language. Follow these specific instructions for formatting the answer:

* Only output the transcription, with no newlines.
* When transcribing numbers, write the digits, i.e. write 1.7 and not one point seven, and write 3 instead of three.<|audio><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><audio|><turn|>
<|turn>model
How can I get to the station?<turn|>

AST

messages = [{
  "role": "user",
  "content": [
    {"type": "text", "text": "Transcribe the following speech segment in English, then translate it into Korean. When formatting the answer, first output the transcription in English, then one newline, then output the string 'Korean: ', then the translation in Korean."},
    {"type": "audio", "audio": "/content/recording.wav"},
  ]
}]

input_ids = processor.apply_chat_template(
        messages,
        add_generation_prompt=True,
        tokenize=True, return_dict=True,
        return_tensors="pt",
)
input_ids = input_ids.to(model.device, dtype=model.dtype)

outputs = model.generate(**input_ids, max_new_tokens=64)

text = processor.batch_decode(
    outputs,
    skip_special_tokens=False,
    clean_up_tokenization_spaces=False
)
print(text[0])

<bos><|turn>user
Transcribe the following speech segment in English, then translate it into Korean. When formatting the answer, first output the transcription in English, then one newline, then output the string 'Korean: ', then the translation in Korean.<|audio><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><|audio|><audio|><turn|>
<|turn>model
How can I get to the station?
Korean: 역에 어떻게 가나요?<turn|>

まとめと次のステップ

このガイドでは、Gemma 4 モデルを使用して音声を処理する方法について説明しました。この例では、話し言葉を書き起こす音声入力（ASR）と、話し言葉を別の言語に直接翻訳する自動音声翻訳（AST）を行う方法を示しました。また、ノートブック環境でマイクから音声をキャプチャして処理する方法についても説明しました。

詳細については、次のドキュメントをご覧ください。