lora_tuning2_ipynb.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "### 使用 LoRA 和 Keras 進行微調\n",
        "\n",
        "[官方文件](https://ai.google.dev/gemma/docs/lora_tuning)"
      ],
      "metadata": {
        "id": "Yz6SNIk7mWP5"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 此 Colab 使用 T4 Runtime 約需 20 分鐘完成"
      ],
      "metadata": {
        "id": "74d7b6qKpwwo"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### 事先準備\n",
        "\n",
        "- Huggingface [Token](https://huggingface.co/settings/tokens)\n",
        "- Kaggle Username / Key: [取得說明](https://christianjmills.com/posts/kaggle-obtain-api-key-tutorial/)\n",
        "- Colab Runtime 請選擇 T4\n",
        "\n",
        "![Screenshot 2024-08-02 at 11.06.46 PM.png](data:image/png;base64,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)\n",
        "\n",
        "##### 密鑰放置在 Colab `Secret`\n",
        "- HuggingFace Token: `HF_TOKEN`\n",
        "- Kaggle username: `KAGGLE_USERNAME`\n",
        "- Kaggle key: `KAGGLE_KEY`"
      ],
      "metadata": {
        "id": "u-qdwMNWnOCB"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Tce3stUlHN0L"
      },
      "source": [
        "##### Copyright 2024 Google LLC."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "tuOe1ymfHZPu"
      },
      "outputs": [],
      "source": [
        "#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n",
        "# you may not use this file except in compliance with the License.\n",
        "# You may obtain a copy of the License at\n",
        "#\n",
        "# https://www.apache.org/licenses/LICENSE-2.0\n",
        "#\n",
        "# Unless required by applicable law or agreed to in writing, software\n",
        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
        "# See the License for the specific language governing permissions and\n",
        "# limitations under the License."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SDEExiAk4fLb"
      },
      "source": [
        "# Fine-tune Gemma models in Keras using LoRA"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZFWzQEqNosrS"
      },
      "source": [
        "<table class=\"tfo-notebook-buttons\" align=\"left\">\n",
        "  <td>\n",
        "    <a target=\"_blank\" href=\"https://ai.google.dev/gemma/docs/lora_tuning\"><img src=\"https://ai.google.dev/static/site-assets/images/docs/notebook-site-button.png\" height=\"32\" width=\"32\" />View on ai.google.dev</a>\n",
        "  <td>\n",
        "    <a target=\"_blank\" href=\"https://colab.research.google.com/github/google/generative-ai-docs/blob/main/site/en/gemma/docs/lora_tuning.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />Run in Google Colab</a>\n",
        "  </td>\n",
        "  <td>\n",
        "    <a target=\"_blank\" href=\"https://console.cloud.google.com/vertex-ai/workbench/deploy-notebook?download_url=https://raw.githubusercontent.com/google/generative-ai-docs/main/site/en/gemma/docs/lora_tuning.ipynb\"><img src=\"https://ai.google.dev/images/cloud-icon.svg\" width=\"40\" />Open in Vertex AI</a>\n",
        "  </td>\n",
        "  <td>\n",
        "    <a target=\"_blank\" href=\"https://github.com/google/generative-ai-docs/blob/main/site/en/gemma/docs/lora_tuning.ipynb\"><img src=\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\" />View source on GitHub</a>\n",
        "  </td>\n",
        "</table>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lSGRSsRPgkzK"
      },
      "source": [
        "## Overview\n",
        "\n",
        "Gemma is a family of lightweight, state-of-the art open models built from the same research and technology used to create the Gemini models.\n",
        "\n",
        "Large Language Models (LLMs) like Gemma have been shown to be effective at a variety of NLP tasks. An LLM is first pre-trained on a large corpus of text in a self-supervised fashion. Pre-training helps LLMs learn general-purpose knowledge, such as statistical relationships between words. An LLM can then be fine-tuned with domain-specific data to perform downstream tasks (such as sentiment analysis).\n",
        "\n",
        "LLMs are extremely large in size (parameters in the order of billions). Full fine-tuning (which updates all the parameters in the model) is not required for most applications because typical fine-tuning datasets are relatively much smaller than the pre-training datasets.\n",
        "\n",
        "[Low Rank Adaptation (LoRA)](https://arxiv.org/abs/2106.09685) is a fine-tuning technique which greatly reduces the number of trainable parameters for downstream tasks by freezing the weights of the model and inserting a smaller number of new weights into the model. This makes training with LoRA much faster and more memory-efficient, and produces smaller model weights (a few hundred MBs), all while maintaining the quality of the model outputs.\n",
        "\n",
        "This tutorial walks you through using KerasNLP to perform LoRA fine-tuning on a Gemma 2B model using the [Databricks Dolly 15k dataset](https://huggingface.co/datasets/databricks/databricks-dolly-15k). This dataset contains 15,000 high-quality human-generated prompt / response pairs specifically designed for fine-tuning LLMs.\n",
        "Gemma 是一系列輕量級、最先進的開放式模型，採用與創建 Gemini 模型相同的研究和技術而建造。\n",
        "\n",
        "像 Gemma 這樣的大型語言模型 (LLM) 已被證明在各種 NLP 任務中都很有效。法學碩士首先以自我監督的方式對大量文本進行預訓練。預訓練幫助法學碩士學習通用知識，例如單字之間的統計關係。然後，法學碩士可以使用特定領域的數據進行微調，以執行下游任務（例如情緒分析）。\n",
        "\n",
        "LLM 的規模非常大（參數數量級為數十億）。大多數應用不需要完全微調（更新模型中的所有參數），因為典型的微調資料集相對比預訓練資料集小得多。\n",
        "\n",
        "低秩適應（LoRA）是一種微調技術，透過凍結模型的權重並向模型中插入較少數量的新權重，大大減少了下游任務的可訓練參數的數量。這使得 LoRA 的訓練速度更快、記憶體效率更高，並產生更小的模型權重（幾百 MB），同時保持模型輸出的品質。\n",
        "\n",
        "本教學將引導您使用 KerasNLP 使用 Databricks Dolly 15k 資料集對 Gemma 2B 模型執行 LoRA 微調。該資料集包含 15,000 個高品質的人工生成的提示/響應對，專為微調法學碩士而設計。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w1q6-W_mKIT-"
      },
      "source": [
        "## Setup"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lyhHCMfoRZ_v"
      },
      "source": [
        "### Get access to Gemma\n",
        "\n",
        "To complete this tutorial, you will first need to complete the setup instructions at [Gemma setup](https://ai.google.dev/gemma/docs/setup). The Gemma setup instructions show you how to do the following:\n",
        "\n",
        "* Get access to Gemma on [kaggle.com](https://kaggle.com).\n",
        "* Select a Colab runtime with sufficient resources to run\n",
        "  the Gemma 2B model.\n",
        "* Generate and configure a Kaggle username and API key.\n",
        "\n",
        "After you've completed the Gemma setup, move on to the next section, where you'll set environment variables for your Colab environment.\n",
        "拜訪傑瑪\n",
        "要完成本教學課程，您首先需要完成 Gemma 設定中的設定說明。 Gemma 設定說明向您展示如何執行以下操作：\n",
        "\n",
        "在 kaggle.com 上造訪 Gemma。\n",
        "選擇具有足夠資源的 Colab 運行時來執行 Gemma 2B 模型。\n",
        "產生並配置 Kaggle 使用者名稱和 API 金鑰。\n",
        "完成 Gemma 設定後，請繼續下一部分，您將為 Colab 環境設定環境變數。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AZ5Qo0fxRZ1V"
      },
      "source": [
        "### Select the runtime\n",
        "\n",
        "To complete this tutorial, you'll need to have a Colab runtime with sufficient resources to run the Gemma model. In this case, you can use a T4 GPU:\n",
        "\n",
        "1. In the upper-right of the Colab window, select &#9662; (**Additional connection options**).\n",
        "2. Select **Change runtime type**.\n",
        "3. Under **Hardware accelerator**, select **T4 GPU**.\n",
        "選擇運行時\n",
        "要完成本教學課程，您需要擁有一個具有足夠資源的 Colab 執行階段來執行 Gemma 模型。在這種情況下，您可以使用 T4 GPU：\n",
        "\n",
        "在 Colab 視窗的右上角，選擇 ▾（其他連線選項）。\n",
        "選擇更改運行時類型。\n",
        "在硬體加速器下，選擇 T4 GPU。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hsPC0HRkJl0K"
      },
      "source": [
        "### Configure your API key\n",
        "\n",
        "To use Gemma, you must provide your Kaggle username and a Kaggle API key.\n",
        "\n",
        "To generate a Kaggle API key, go to the **Account** tab of your Kaggle user profile and select **Create New Token**. This will trigger the download of a `kaggle.json` file containing your API credentials.\n",
        "\n",
        "In Colab, select **Secrets** (🔑) in the left pane and add your Kaggle username and Kaggle API key. Store your username under the name `KAGGLE_USERNAME` and your API key under the name `KAGGLE_KEY`.\n",
        "配置您的 API 金鑰\n",
        "要使用 Gemma，您必須提供您的 Kaggle 使用者名稱和 Kaggle API 金鑰。\n",
        "\n",
        "若要產生 Kaggle API 金鑰，請前往 Kaggle 使用者個人資料的「帳戶」選項卡，然後選擇「建立新令牌」。這將觸發包含您的 API 憑證的 kaggle.json 檔案的下載。\n",
        "\n",
        "在 Colab 中，選擇左側窗格中的 Secrets (🔑)，然後新增您的 Kaggle 使用者名稱和 Kaggle API 金鑰。將您的使用者名稱儲存在 KAGGLE_USERNAME 名稱下，並將您的 API 金鑰儲存在 KAGGLE_KEY 名稱下。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7iOF6Yo-wUEC"
      },
      "source": [
        "### Set environment variables\n",
        "\n",
        "Set environment variables for `KAGGLE_USERNAME` and `KAGGLE_KEY`."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0_EdOg9DPK6Q"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "from google.colab import userdata\n",
        "\n",
        "# Note: `userdata.get` is a Colab API. If you're not using Colab, set the env\n",
        "# vars as appropriate for your system.\n",
        "\n",
        "os.environ[\"KAGGLE_USERNAME\"] = userdata.get('KAGGLE_USERNAME')\n",
        "os.environ[\"KAGGLE_KEY\"] = userdata.get('KAGGLE_KEY')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CuEUAKJW1QkQ"
      },
      "source": [
        "### Install dependencies\n",
        "\n",
        "Install Keras, KerasNLP, and other dependencies."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "1eeBtYqJsZPG"
      },
      "outputs": [],
      "source": [
        "# Install Keras 3 last. See https://keras.io/getting_started/ for more details.\n",
        "!pip install -q -U keras-nlp\n",
        "!pip install -q -U \"keras>=3\""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rGLS-l5TxIR4"
      },
      "source": [
        "### Select a backend\n",
        "\n",
        "Keras is a high-level, multi-framework deep learning API designed for simplicity and ease of use. Using Keras 3, you can run workflows on one of three backends: TensorFlow, JAX, or PyTorch.\n",
        "\n",
        "For this tutorial, configure the backend for JAX.\n",
        "選擇後端\n",
        "Keras 是一種進階、多框架深度學習 API，旨在簡單易用。使用 Keras 3，您可以在三個後端之一執行工作流程：TensorFlow、JAX 或 PyTorch。\n",
        "\n",
        "在本教程中，配置 JAX 的後端。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "yn5uy8X8sdD0"
      },
      "outputs": [],
      "source": [
        "os.environ[\"KERAS_BACKEND\"] = \"jax\"  # Or \"torch\" or \"tensorflow\".\n",
        "# Avoid memory fragmentation on JAX backend.\n",
        "os.environ[\"XLA_PYTHON_CLIENT_MEM_FRACTION\"]=\"1.00\""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hZs8XXqUKRmi"
      },
      "source": [
        "### Import packages\n",
        "\n",
        "Import Keras and KerasNLP."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "FYHyPUA9hKTf"
      },
      "outputs": [],
      "source": [
        "import keras\n",
        "import keras_nlp"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9T7xe_jzslv4"
      },
      "source": [
        "## Load Dataset"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "import json  # 添加这一行来导入json模块\n",
        "\n",
        "# 文件路径\n",
        "file_path = '/content/databricks-dolly-33.jsonl'\n",
        "\n",
        "# 读取JSONL文件\n",
        "def load_jsonl(file_path):\n",
        "    data = []\n",
        "    with open(file_path, 'r') as file:\n",
        "        for line in file:\n",
        "            data.append(json.loads(line))\n",
        "    return pd.DataFrame(data)\n",
        "\n",
        "# 加载数据\n",
        "df = load_jsonl(file_path)\n",
        "\n",
        "# 显示数据的前几行\n",
        "print(df.head())\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vSQyun9-kT1e",
        "outputId": "3c0a7f22-98a7-4bb7-c886-7b287438e27d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "                                         instruction context  \\\n",
            "0  You are an AI drawing master who is good at ac...           \n",
            "1  Please play the role of an AI drawing master w...           \n",
            "2          Create an 'imagine prompt' of a cute girl           \n",
            "3        Create an 'imagine prompt' of a cute kitten           \n",
            "4  Create an 'imagine prompt' of a cute girl wave...           \n",
            "\n",
            "                                            response category  \n",
            "0  PROMPT: A female character wearing a bikini wi...  open_qa  \n",
            "1  prompt:Cinderella, tripping on the stairs, gla...  open_qa  \n",
            "2  prompt: a cute young girl with rosy cheeks and...  open_qa  \n",
            "3  prompt: fluffy grey kitten, curled up asleep o...  open_qa  \n",
            "4  prompt: a cheerful young woman with rosy cheek...  open_qa  \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tEM6ZvhpbtBm",
        "outputId": "cd549086-93cf-41f8-8b2e-5527f0dc9cf0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "45UpBDfBgf0I"
      },
      "source": [
        "Preprocess the data. This tutorial uses a subset of 1000 training examples to execute the notebook faster. Consider using more training data for higher quality fine-tuning.\n",
        "預處理資料。本教學使用 1000 個訓練範例的子集來更快地執行筆記本。考慮使用更多的訓練資料來進行更高品質的微調。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZiS-KU9osh_N"
      },
      "outputs": [],
      "source": [
        "import json\n",
        "data = []\n",
        "with open(\"databricks-dolly-33.jsonl\") as file:\n",
        "    for line in file:\n",
        "        features = json.loads(line)\n",
        "        # Filter out examples with context, to keep it simple.\n",
        "        if features[\"context\"]:\n",
        "            continue\n",
        "        # Format the entire example as a single string.\n",
        "        template = \"Instruction:\\n{instruction}\\n\\nResponse:\\n{response}\"\n",
        "        data.append(template.format(**features))\n",
        "\n",
        "# Only use 1000 training examples, to keep it fast.\n",
        "data = data[:24]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7RCE3fdGhDE5"
      },
      "source": [
        "## Load Model\n",
        "\n",
        "KerasNLP provides implementations of many popular [model architectures](https://keras.io/api/keras_nlp/models/). In this tutorial, you'll create a model using `GemmaCausalLM`, an end-to-end Gemma model for causal language modeling. A causal language model predicts the next token based on previous tokens.\n",
        "\n",
        "Create the model using the `from_preset` method:\n",
        "KerasNLP 提供了許多流行模型架構的實作。在本教程中，您將使用 GemmaCausalLM 建立一個模型，這是一個用於因果語言建模的端對端 Gemma 模型。因果語言模型根據先前的標記預測下一個標記。\n",
        "\n",
        "使用 from_preset 方法建立模型："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vz5zLEyLstfn",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        },
        "outputId": "c6b7d1ab-530e-4c98-cf1e-91c34ca39db6"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1mPreprocessor: \"gemma_causal_lm_preprocessor_1\"\u001b[0m\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Preprocessor: \"gemma_causal_lm_preprocessor_1\"</span>\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
              "┃\u001b[1m \u001b[0m\u001b[1mTokenizer (type)                                  \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m                                            Vocab #\u001b[0m\u001b[1m \u001b[0m┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
              "│ gemma_tokenizer (\u001b[38;5;33mGemmaTokenizer\u001b[0m)                   │                                             \u001b[38;5;34m256,000\u001b[0m │\n",
              "└────────────────────────────────────────────────────┴─────────────────────────────────────────────────────┘\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
              "┃<span style=\"font-weight: bold\"> Tokenizer (type)                                   </span>┃<span style=\"font-weight: bold\">                                             Vocab # </span>┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
              "│ gemma_tokenizer (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">GemmaTokenizer</span>)                   │                                             <span style=\"color: #00af00; text-decoration-color: #00af00\">256,000</span> │\n",
              "└────────────────────────────────────────────────────┴─────────────────────────────────────────────────────┘\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1mModel: \"gemma_causal_lm_1\"\u001b[0m\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"gemma_causal_lm_1\"</span>\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
              "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                 \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape             \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m        Param #\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConnected to              \u001b[0m\u001b[1m \u001b[0m┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
              "│ padding_mask (\u001b[38;5;33mInputLayer\u001b[0m)     │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;45mNone\u001b[0m)              │               \u001b[38;5;34m0\u001b[0m │ -                          │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ token_ids (\u001b[38;5;33mInputLayer\u001b[0m)        │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;45mNone\u001b[0m)              │               \u001b[38;5;34m0\u001b[0m │ -                          │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ gemma_backbone                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m2304\u001b[0m)        │   \u001b[38;5;34m2,614,341,888\u001b[0m │ padding_mask[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m],        │\n",
              "│ (\u001b[38;5;33mGemmaBackbone\u001b[0m)               │                           │                 │ token_ids[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]            │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ token_embedding               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256000\u001b[0m)      │     \u001b[38;5;34m589,824,000\u001b[0m │ gemma_backbone[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]       │\n",
              "│ (\u001b[38;5;33mReversibleEmbedding\u001b[0m)         │                           │                 │                            │\n",
              "└───────────────────────────────┴───────────────────────────┴─────────────────┴────────────────────────────┘\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
              "┃<span style=\"font-weight: bold\"> Layer (type)                  </span>┃<span style=\"font-weight: bold\"> Output Shape              </span>┃<span style=\"font-weight: bold\">         Param # </span>┃<span style=\"font-weight: bold\"> Connected to               </span>┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
              "│ padding_mask (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)     │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>)              │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                          │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ token_ids (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>)              │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                          │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ gemma_backbone                │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">2304</span>)        │   <span style=\"color: #00af00; text-decoration-color: #00af00\">2,614,341,888</span> │ padding_mask[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>],        │\n",
              "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">GemmaBackbone</span>)               │                           │                 │ token_ids[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]            │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ token_embedding               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256000</span>)      │     <span style=\"color: #00af00; text-decoration-color: #00af00\">589,824,000</span> │ gemma_backbone[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]       │\n",
              "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">ReversibleEmbedding</span>)         │                           │                 │                            │\n",
              "└───────────────────────────────┴───────────────────────────┴─────────────────┴────────────────────────────┘\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m2,614,341,888\u001b[0m (9.74 GB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">2,614,341,888</span> (9.74 GB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m2,614,341,888\u001b[0m (9.74 GB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">2,614,341,888</span> (9.74 GB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        }
      ],
      "source": [
        "gemma_lm = keras_nlp.models.GemmaCausalLM.from_preset(\"gemma2_2b_en\")\n",
        "gemma_lm.summary()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Nl4lvPy5zA26"
      },
      "source": [
        "The `from_preset` method instantiates the model from a preset architecture and weights. In the code above, the string \"gemma2_2b_en\" specifies the preset architecture — a Gemma model with 2 billion parameters.\n",
        "\n",
        "NOTE: A Gemma model with 7\n",
        "billion parameters is also available. To run the larger model in Colab, you need access to the premium GPUs available in paid plans. Alternatively, you can perform [distributed tuning on a Gemma 7B model](https://ai.google.dev/gemma/docs/distributed_tuning) on Kaggle or Google Cloud.\n",
        "from_preset 方法根據預設的架構和權重實例化模型。在上面的程式碼中，字串「gemma2_2b_en」指定了預設架構 - 具有 20 億個參數的 Gemma 模型。\n",
        "\n",
        "注意：具有 70 億個參數的 Gemma 模型也可用。要在 Colab 中運行更大的模型，您需要存取付費計劃中提供的高級 GPU。或者，您可以對 Kaggle 或 Google Cloud 上的 Gemma 7B 模型執行分散式調優。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "G_L6A5J-1QgC"
      },
      "source": [
        "## Inference before fine tuning\n",
        "\n",
        "In this section, you will query the model with various prompts to see how it responds.\n",
        "## 微調前的推斷\n",
        "\n",
        "在本部分中，您將使用各種提示查詢模型以查看它如何回應。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PVLXadptyo34"
      },
      "source": [
        "### Europe Trip Prompt\n",
        "\n",
        "Query the model for suggestions on what to do on a trip to Europe.\n",
        "歐洲旅行提示\n",
        "查詢模型以取得有關去歐洲旅行時應該做什麼的建議。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZwQz3xxxKciD"
      },
      "outputs": [],
      "source": [
        "template = \"instruction: {instruction}\\nresponse: {response}\"\n",
        "prompt = template.format(\n",
        "    instruction=\"What should I do on a trip to Europe?\",\n",
        "    response=\"\",\n",
        ")\n",
        "sampler = keras_nlp.samplers.TopKSampler(k=5, seed=2)\n",
        "gemma_lm.compile(sampler=sampler)\n",
        "print(gemma_lm.generate(prompt, max_length=256))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AePQUIs2h-Ks"
      },
      "source": [
        "The model responds with generic tips on how to plan a trip.\n",
        "該模型會給出有關如何計劃旅行的一般提示"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YQ74Zz_S0iVv"
      },
      "source": [
        "### ELI5 Photosynthesis Prompt\n",
        "\n",
        "Prompt the model to explain photosynthesis in terms simple enough for a 5 year old child to understand.\n",
        "ELI5 光合作用提示\n",
        "\n",
        "提示模型以 5 歲兒童能夠理解的簡單術語解釋光合作用。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lorJMbsusgoo",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 339
        },
        "outputId": "960f9819-bc4f-4555-b12e-4312dcc14711"
      },
      "outputs": [
        {
          "output_type": "error",
          "ename": "TypeError",
          "evalue": "'NoneType' object is not callable",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-31-0a9039fc45dd>\u001b[0m in \u001b[0;36m<cell line: 5>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      3\u001b[0m     \u001b[0mresponse\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m )\n\u001b[0;32m----> 5\u001b[0;31m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgemma_lm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgenerate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprompt\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmax_length\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m256\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras_nlp/src/models/causal_lm.py\u001b[0m in \u001b[0;36mgenerate\u001b[0;34m(self, inputs, max_length, stop_token_ids)\u001b[0m\n\u001b[1;32m    378\u001b[0m                 \u001b[0minputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mpreprocess\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    379\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 380\u001b[0;31m         \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mgenerate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    381\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    382\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpreprocessor\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras_nlp/src/models/causal_lm.py\u001b[0m in \u001b[0;36m<listcomp>\u001b[0;34m(.0)\u001b[0m\n\u001b[1;32m    378\u001b[0m                 \u001b[0minputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mpreprocess\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    379\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 380\u001b[0;31m         \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mgenerate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    381\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    382\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpreprocessor\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras_nlp/src/models/causal_lm.py\u001b[0m in \u001b[0;36mgenerate\u001b[0;34m(x)\u001b[0m\n\u001b[1;32m    362\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    363\u001b[0m         \u001b[0;32mdef\u001b[0m \u001b[0mgenerate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 364\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mgenerate_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstop_token_ids\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mstop_token_ids\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    365\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    366\u001b[0m         \u001b[0;32mdef\u001b[0m \u001b[0mpostprocess\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras_nlp/src/models/causal_lm.py\u001b[0m in \u001b[0;36mwrapped_generate_function\u001b[0;34m(inputs, stop_token_ids)\u001b[0m\n\u001b[1;32m    208\u001b[0m                     \u001b[0;31m# Use the explicit variable.value to preserve the\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    209\u001b[0m                     \u001b[0;31m# sharding spec of distribution.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 210\u001b[0;31m                     \u001b[0;34m[\u001b[0m\u001b[0mv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalue\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mv\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrainable_variables\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    211\u001b[0m                     \u001b[0;34m[\u001b[0m\u001b[0mv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalue\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mv\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnon_trainable_variables\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    212\u001b[0m                 )\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras_nlp/src/models/causal_lm.py\u001b[0m in \u001b[0;36m<listcomp>\u001b[0;34m(.0)\u001b[0m\n\u001b[1;32m    208\u001b[0m                     \u001b[0;31m# Use the explicit variable.value to preserve the\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    209\u001b[0m                     \u001b[0;31m# sharding spec of distribution.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 210\u001b[0;31m                     \u001b[0;34m[\u001b[0m\u001b[0mv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalue\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mv\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrainable_variables\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    211\u001b[0m                     \u001b[0;34m[\u001b[0m\u001b[0mv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalue\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mv\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnon_trainable_variables\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    212\u001b[0m                 )\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras/src/backend/common/variables.py\u001b[0m in \u001b[0;36mvalue\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    217\u001b[0m             \u001b[0;31m# (anything else would be a bug, to be fixed.)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    218\u001b[0m             return self._maybe_autocast(\n\u001b[0;32m--> 219\u001b[0;31m                 \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_initializer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_shape\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_dtype\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    220\u001b[0m             )\n\u001b[1;32m    221\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_maybe_autocast\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_value\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mTypeError\u001b[0m: 'NoneType' object is not callable"
          ]
        }
      ],
      "source": [
        "prompt = template.format(\n",
        "    instruction=\"Explain the process of photosynthesis in a way that a child could understand.\",\n",
        "    response=\"\",\n",
        ")\n",
        "print(gemma_lm.generate(prompt, max_length=256))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WBQieduRizZf"
      },
      "source": [
        "The model response contains words that might not be easy to understand for a child such as chlorophyll.\n",
        "模型回應包含兒童可能不容易理解的單字，例如葉綠素。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Pt7Nr6a7tItO"
      },
      "source": [
        "## LoRA Fine-tuning\n",
        "\n",
        "To get better responses from the model, fine-tune the model with Low Rank Adaptation (LoRA) using the Databricks Dolly 15k dataset.\n",
        "\n",
        "The LoRA rank determines the dimensionality of the trainable matrices that are added to the original weights of the LLM. It controls the expressiveness and precision of the fine-tuning adjustments.\n",
        "\n",
        "A higher rank means more detailed changes are possible, but also means more trainable parameters. A lower rank means less computational overhead, but potentially less precise adaptation.\n",
        "\n",
        "This tutorial uses a LoRA rank of 4. In practice, begin with a relatively small rank (such as 4, 8, 16). This is computationally efficient for experimentation. Train your model with this rank and evaluate the performance improvement on your task. Gradually increase the rank in subsequent trials and see if that further boosts performance.\n",
        "LoRA微調\n",
        "\n",
        "為了從模型中獲得更好的反應，請使用 Databricks Dolly 15k 資料集透過低秩適應 (LoRA) 微調模型。\n",
        "\n",
        "LoRA 等級決定了加到 LLM 原始權重的可訓練矩陣的維數。它控制微調調整的表現力和精確度。\n",
        "\n",
        "更高的排名意味著可以進行更詳細的更改，但也意味著更多可訓練的參數。較低的等級意味著較少的計算開銷，但可能不太精確的適應。\n",
        "\n",
        "本教程使用 LoRA 等級 4。這對於實驗來說計算效率很高。使用此排名訓練您的模型並評估您的任務的表現改進。在後續試驗中逐漸提高排名，看看是否會進一步提高效能。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "RCucu6oHz53G",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        },
        "outputId": "2d96c239-50cd-4b01-e048-752b79bb87c6"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1mPreprocessor: \"gemma_causal_lm_preprocessor_1\"\u001b[0m\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Preprocessor: \"gemma_causal_lm_preprocessor_1\"</span>\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
              "┃\u001b[1m \u001b[0m\u001b[1mTokenizer (type)                                  \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m                                            Vocab #\u001b[0m\u001b[1m \u001b[0m┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
              "│ gemma_tokenizer (\u001b[38;5;33mGemmaTokenizer\u001b[0m)                   │                                             \u001b[38;5;34m256,000\u001b[0m │\n",
              "└────────────────────────────────────────────────────┴─────────────────────────────────────────────────────┘\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
              "┃<span style=\"font-weight: bold\"> Tokenizer (type)                                   </span>┃<span style=\"font-weight: bold\">                                             Vocab # </span>┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
              "│ gemma_tokenizer (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">GemmaTokenizer</span>)                   │                                             <span style=\"color: #00af00; text-decoration-color: #00af00\">256,000</span> │\n",
              "└────────────────────────────────────────────────────┴─────────────────────────────────────────────────────┘\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1mModel: \"gemma_causal_lm_1\"\u001b[0m\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"gemma_causal_lm_1\"</span>\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
              "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                 \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape             \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m        Param #\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConnected to              \u001b[0m\u001b[1m \u001b[0m┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
              "│ padding_mask (\u001b[38;5;33mInputLayer\u001b[0m)     │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;45mNone\u001b[0m)              │               \u001b[38;5;34m0\u001b[0m │ -                          │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ token_ids (\u001b[38;5;33mInputLayer\u001b[0m)        │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;45mNone\u001b[0m)              │               \u001b[38;5;34m0\u001b[0m │ -                          │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ gemma_backbone                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m2304\u001b[0m)        │   \u001b[38;5;34m2,617,270,528\u001b[0m │ padding_mask[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m],        │\n",
              "│ (\u001b[38;5;33mGemmaBackbone\u001b[0m)               │                           │                 │ token_ids[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]            │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ token_embedding               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256000\u001b[0m)      │     \u001b[38;5;34m589,824,000\u001b[0m │ gemma_backbone[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]       │\n",
              "│ (\u001b[38;5;33mReversibleEmbedding\u001b[0m)         │                           │                 │                            │\n",
              "└───────────────────────────────┴───────────────────────────┴─────────────────┴────────────────────────────┘\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓\n",
              "┃<span style=\"font-weight: bold\"> Layer (type)                  </span>┃<span style=\"font-weight: bold\"> Output Shape              </span>┃<span style=\"font-weight: bold\">         Param # </span>┃<span style=\"font-weight: bold\"> Connected to               </span>┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩\n",
              "│ padding_mask (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)     │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>)              │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                          │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ token_ids (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>)              │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                          │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ gemma_backbone                │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">2304</span>)        │   <span style=\"color: #00af00; text-decoration-color: #00af00\">2,617,270,528</span> │ padding_mask[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>],        │\n",
              "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">GemmaBackbone</span>)               │                           │                 │ token_ids[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]            │\n",
              "├───────────────────────────────┼───────────────────────────┼─────────────────┼────────────────────────────┤\n",
              "│ token_embedding               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256000</span>)      │     <span style=\"color: #00af00; text-decoration-color: #00af00\">589,824,000</span> │ gemma_backbone[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]       │\n",
              "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">ReversibleEmbedding</span>)         │                           │                 │                            │\n",
              "└───────────────────────────────┴───────────────────────────┴─────────────────┴────────────────────────────┘\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m2,617,270,528\u001b[0m (9.75 GB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">2,617,270,528</span> (9.75 GB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m2,928,640\u001b[0m (11.17 MB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">2,928,640</span> (11.17 MB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m2,614,341,888\u001b[0m (9.74 GB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">2,614,341,888</span> (9.74 GB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        }
      ],
      "source": [
        "# Enable LoRA for the model and set the LoRA rank to 4.\n",
        "gemma_lm.backbone.enable_lora(rank=4)\n",
        "gemma_lm.summary()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hQQ47kcdpbZ9"
      },
      "source": [
        "Note that enabling LoRA reduces the number of trainable parameters significantly (from 2.6 billion to 2.9 million).\n",
        "請注意，啟用 LoRA 會顯著減少可訓練參數的數量（從 26 億減少到 290 萬）。"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# test,do not run it\n",
        "import tensorflow as tf\n",
        "\n",
        "def preprocess_data(data):\n",
        "    instructions = [item['instruction'] for item in data]\n",
        "    responses = [item['response'] for item in data]\n",
        "\n",
        "    # 假設 `instructions` 和 `responses` 已經是處理過的張量\n",
        "    instructions_tensor = tf.convert_to_tensor(instructions, dtype=tf.string)\n",
        "    responses_tensor = tf.convert_to_tensor(responses, dtype=tf.string)\n",
        "\n",
        "    dataset = tf.data.Dataset.from_tensor_slices((instructions_tensor, responses_tensor))\n",
        "    return dataset\n",
        "\n",
        "# 假設 `data` 是你的字典列表\n",
        "dataset = preprocess_data(data)\n"
      ],
      "metadata": {
        "id": "P85Eeur9iuLP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# test,do not run it\n",
        "for feature, label in dataset.take(1):\n",
        "    print(\"Feature:\", feature)\n",
        "    print(\"Label:\", label)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "adCuSBLNi3rA",
        "outputId": "719de086-c6eb-4463-bc92-51725099b40f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Feature: tf.Tensor(b\"You are an AI drawing master who is good at accurately describing keywords into prompts. Please describe the drawing of 'a woman wearing a bikini'. Please draw a drawing in the style of a comic character, including characters, colors, light and shadow, background, composition...etc. Prompts that make the drawing more similar to the Q version character style are a must! Please convert it into prompt in the form of English words. If it is a short sentence, please use () to cover the short sentence. You can customize the weight and give it the weight of prompt(:1~1.5). and Start with 'prompt:' and after completing the description, change to the next line and start with '--' to add an error prompt that will affect the drawing itself.\", shape=(), dtype=string)\n",
            "Label: tf.Tensor(b'PROMPT: A female character wearing a bikini with a Q-version style. (:1.2) Character: Cute, lively, and with exaggerated facial features. (:1.3) Bikini: Colorful and vibrant, with a sense of playfulness. (:1.2) Colors: Bright and eye-catching, with a preference for pastel shades. (:1.1) Lighting and Shadows: Soft, with gentle shading to add depth to the character. (:1.2) Background: A sunny beach with palm trees and a blue sky, creating a cheerful atmosphere. (:1.3) Composition: The character placed in the center, striking a cute and confident pose. (:1.5) Attention to Details: Accessorize the character with sunglasses and a beach hat.', shape=(), dtype=string)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install jsonlines"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IgdxugZswdAt",
        "outputId": "8030e3fd-0923-4517-acb6-c94a89edbd6d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting jsonlines\n",
            "  Downloading jsonlines-4.0.0-py3-none-any.whl.metadata (1.6 kB)\n",
            "Requirement already satisfied: attrs>=19.2.0 in /usr/local/lib/python3.10/dist-packages (from jsonlines) (24.2.0)\n",
            "Downloading jsonlines-4.0.0-py3-none-any.whl (8.7 kB)\n",
            "Installing collected packages: jsonlines\n",
            "Successfully installed jsonlines-4.0.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_Peq7TnLtHse",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "c588fa82-7306-40c0-a460-817d1cd6c858"
      },
      "outputs": [
        {
          "output_type": "error",
          "ename": "ValueError",
          "evalue": "Attempt to convert a value ({'instruction': \"You are an AI drawing master who is good at accurately describing keywords into prompts. Please describe the drawing of 'a woman wearing a bikini'. Please draw a drawing in the style of a comic character, including characters, colors, light and shadow, background, composition...etc. Prompts that make the drawing more similar to the Q version character style are a must! Please convert it into prompt in the form of English words. If it is a short sentence, please use () to cover the short sentence. You can customize the weight and give it the weight of prompt(:1~1.5). and Start with 'prompt:' and after completing the description, change to the next line and start with '--' to add an error prompt that will affect the drawing itself.\", 'context': '', 'response': 'PROMPT: A female character wearing a bikini with a Q-version style. (:1.2) Character: Cute, lively, and with exaggerated facial features. (:1.3) Bikini: Colorful and vibrant, with a sense of playfulness. (:1.2) Colors: Bright and eye-catching, with a preference for pastel shades. (:1.1) Lighting and Shadows: Soft, with gentle shading to add depth to the character. (:1.2) Background: A sunny beach with palm trees and a blue sky, creating a cheerful atmosphere. (:1.3) Composition: The character placed in the center, striking a cute and confident pose. (:1.5) Attention to Details: Accessorize the character with sunglasses and a beach hat.', 'category': 'open_qa'}) with an unsupported type (<class 'dict'>) to a Tensor.",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/data/util/structure.py\u001b[0m in \u001b[0;36mnormalize_element\u001b[0;34m(element, element_signature)\u001b[0m\n\u001b[1;32m    104\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mspec\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 105\u001b[0;31m           \u001b[0mspec\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtype_spec_from_value\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mt\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0muse_fallback\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    106\u001b[0m       \u001b[0;32mexcept\u001b[0m \u001b[0mTypeError\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/data/util/structure.py\u001b[0m in \u001b[0;36mtype_spec_from_value\u001b[0;34m(element, use_fallback)\u001b[0m\n\u001b[1;32m    513\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 514\u001b[0;31m   raise TypeError(\"Could not build a `TypeSpec` for {} with type {}\".format(\n\u001b[0m\u001b[1;32m    515\u001b[0m       \u001b[0melement\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mTypeError\u001b[0m: Could not build a `TypeSpec` for [{'instruction': \"You are an AI drawing master who is good at accurately describing keywords into prompts. Please describe the drawing of 'a woman wearing a bikini'. Please draw a drawing in the style of a comic character, including characters, colors, light and shadow, background, composition...etc. Prompts that make the drawing more similar to the Q version character style are a must! Please convert it into prompt in the form of English words. If it is a short sentence, please use () to cover the short sentence. You can customize the weight and give it the weight of prompt(:1~1.5). and Start with 'prompt:' and after completing the description, change to the next line and start with '--' to add an error prompt that will affect the drawing itself.\", 'context': '', 'response': 'PROMPT: A female character wearing a bikini with a Q-version style. (:1.2) Character: Cute, lively, and with exaggerated facial features. (:1.3) Bikini: Colorful and vibrant, with a sense of playfulness. (:1.2) Colors: Bright and eye-catching, with a preference for pastel shades. (:1.1) Lighting and Shadows: Soft, with gentle shading to add depth to the character. (:1.2) Background: A sunny beach with palm trees and a blue sky, creating a cheerful atmosphere. (:1.3) Composition: The character placed in the center, striking a cute and confident pose. (:1.5) Attention to Details: Accessorize the character with sunglasses and a beach hat.', 'categor...",
            "\nDuring handling of the above exception, another exception occurred:\n",
            "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-10-bc06812f22bd>\u001b[0m in \u001b[0;36m<cell line: 24>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     22\u001b[0m )\n\u001b[1;32m     23\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 24\u001b[0;31m \u001b[0mgemma_lm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mepochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras_nlp/src/utils/pipeline_model.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, sample_weight, validation_data, validation_split, **kwargs)\u001b[0m\n\u001b[1;32m    177\u001b[0m             )\n\u001b[1;32m    178\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 179\u001b[0;31m         \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_convert_inputs_to_dataset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msample_weight\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    180\u001b[0m         x = x.map(\n\u001b[1;32m    181\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpreprocess_samples\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_parallel_calls\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mAUTOTUNE\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras_nlp/src/utils/pipeline_model.py\u001b[0m in \u001b[0;36m_convert_inputs_to_dataset\u001b[0;34m(x, y, sample_weight, batch_size)\u001b[0m\n\u001b[1;32m     83\u001b[0m                 \u001b[0;34m\"input, e.g., wrap it in a list.\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     84\u001b[0m             ) from e\n\u001b[0;32m---> 85\u001b[0;31m         \u001b[0;32mraise\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     86\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     87\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mds\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbatch\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbatch_size\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0;36m32\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras_nlp/src/utils/pipeline_model.py\u001b[0m in \u001b[0;36m_convert_inputs_to_dataset\u001b[0;34m(x, y, sample_weight, batch_size)\u001b[0m\n\u001b[1;32m     71\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     72\u001b[0m         \u001b[0minputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtree\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmap_structure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconvert\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 73\u001b[0;31m         \u001b[0mds\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mDataset\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfrom_tensor_slices\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     74\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0mValueError\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     75\u001b[0m         \u001b[0;31m# If our inputs are unbatched, re-raise with a more friendly error\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/data/ops/dataset_ops.py\u001b[0m in \u001b[0;36mfrom_tensor_slices\u001b[0;34m(tensors, name)\u001b[0m\n\u001b[1;32m    824\u001b[0m     \u001b[0;31m# pylint: disable=g-import-not-at-top,protected-access\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    825\u001b[0m     \u001b[0;32mfrom\u001b[0m \u001b[0mtensorflow\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpython\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mops\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mfrom_tensor_slices_op\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 826\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mfrom_tensor_slices_op\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_from_tensor_slices\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtensors\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    827\u001b[0m     \u001b[0;31m# pylint: enable=g-import-not-at-top,protected-access\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    828\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/data/ops/from_tensor_slices_op.py\u001b[0m in \u001b[0;36m_from_tensor_slices\u001b[0;34m(tensors, name)\u001b[0m\n\u001b[1;32m     23\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     24\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m_from_tensor_slices\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtensors\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 25\u001b[0;31m   \u001b[0;32mreturn\u001b[0m \u001b[0m_TensorSliceDataset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtensors\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     26\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     27\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/data/ops/from_tensor_slices_op.py\u001b[0m in \u001b[0;36m__init__\u001b[0;34m(self, element, is_files, name)\u001b[0m\n\u001b[1;32m     31\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0melement\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mis_files\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     32\u001b[0m     \u001b[0;34m\"\"\"See `Dataset.from_tensor_slices` for details.\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 33\u001b[0;31m     \u001b[0melement\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstructure\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnormalize_element\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0melement\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     34\u001b[0m     \u001b[0mbatched_spec\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstructure\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtype_spec_from_value\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0melement\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     35\u001b[0m     \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_tensors\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstructure\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto_batched_tensor_list\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbatched_spec\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0melement\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/data/util/structure.py\u001b[0m in \u001b[0;36mnormalize_element\u001b[0;34m(element, element_signature)\u001b[0m\n\u001b[1;32m    108\u001b[0m         \u001b[0;31m# the value. As a fallback try converting the value to a tensor.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    109\u001b[0m         normalized_components.append(\n\u001b[0;32m--> 110\u001b[0;31m             ops.convert_to_tensor(t, name=\"component_%d\" % i))\n\u001b[0m\u001b[1;32m    111\u001b[0m       \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    112\u001b[0m         \u001b[0;31m# To avoid a circular dependency between dataset_ops and structure,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/profiler/trace.py\u001b[0m in \u001b[0;36mwrapped\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    181\u001b[0m         \u001b[0;32mwith\u001b[0m \u001b[0mTrace\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrace_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mtrace_kwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    182\u001b[0m           \u001b[0;32mreturn\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 183\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    184\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    185\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mwrapped\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/framework/ops.py\u001b[0m in \u001b[0;36mconvert_to_tensor\u001b[0;34m(value, dtype, name, as_ref, preferred_dtype, dtype_hint, ctx, accepted_result_types)\u001b[0m\n\u001b[1;32m    711\u001b[0m   \u001b[0;31m# TODO(b/142518781): Fix all call-sites and remove redundant arg\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    712\u001b[0m   \u001b[0mpreferred_dtype\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpreferred_dtype\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mdtype_hint\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 713\u001b[0;31m   return tensor_conversion_registry.convert(\n\u001b[0m\u001b[1;32m    714\u001b[0m       \u001b[0mvalue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mas_ref\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpreferred_dtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maccepted_result_types\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    715\u001b[0m   )\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/framework/tensor_conversion_registry.py\u001b[0m in \u001b[0;36mconvert\u001b[0;34m(value, dtype, name, as_ref, preferred_dtype, accepted_result_types)\u001b[0m\n\u001b[1;32m    232\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    233\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mret\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 234\u001b[0;31m       \u001b[0mret\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mconversion_func\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mas_ref\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mas_ref\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    235\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    236\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mret\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0mNotImplemented\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/framework/constant_tensor_conversion.py\u001b[0m in \u001b[0;36m_constant_tensor_conversion_function\u001b[0;34m(v, dtype, name, as_ref)\u001b[0m\n\u001b[1;32m     27\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     28\u001b[0m   \u001b[0m_\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mas_ref\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 29\u001b[0;31m   \u001b[0;32mreturn\u001b[0m \u001b[0mconstant_op\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconstant\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mv\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     30\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     31\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/ops/weak_tensor_ops.py\u001b[0m in \u001b[0;36mwrapper\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    140\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0mwrapper\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    141\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_auto_dtype_conversion_enabled\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 142\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mop\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    143\u001b[0m     \u001b[0mbound_arguments\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msignature\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbind\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    144\u001b[0m     \u001b[0mbound_arguments\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mapply_defaults\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/framework/constant_op.py\u001b[0m in \u001b[0;36mconstant\u001b[0;34m(value, dtype, shape, name)\u001b[0m\n\u001b[1;32m    274\u001b[0m     \u001b[0mValueError\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mcalled\u001b[0m \u001b[0mon\u001b[0m \u001b[0ma\u001b[0m \u001b[0msymbolic\u001b[0m \u001b[0mtensor\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    275\u001b[0m   \"\"\"\n\u001b[0;32m--> 276\u001b[0;31m   return _constant_impl(value, dtype, shape, name, verify_shape=False,\n\u001b[0m\u001b[1;32m    277\u001b[0m                         allow_broadcast=True)\n\u001b[1;32m    278\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/framework/constant_op.py\u001b[0m in \u001b[0;36m_constant_impl\u001b[0;34m(value, dtype, shape, name, verify_shape, allow_broadcast)\u001b[0m\n\u001b[1;32m    287\u001b[0m       \u001b[0;32mwith\u001b[0m \u001b[0mtrace\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTrace\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"tf.constant\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    288\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0m_constant_eager_impl\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mctx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mshape\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mverify_shape\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 289\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0m_constant_eager_impl\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mctx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mshape\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mverify_shape\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    290\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    291\u001b[0m   const_tensor = ops._create_graph_constant(  # pylint: disable=protected-access\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/framework/constant_op.py\u001b[0m in \u001b[0;36m_constant_eager_impl\u001b[0;34m(ctx, value, dtype, shape, verify_shape)\u001b[0m\n\u001b[1;32m    299\u001b[0m ) -> ops._EagerTensorBase:\n\u001b[1;32m    300\u001b[0m   \u001b[0;34m\"\"\"Creates a constant on the current device.\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 301\u001b[0;31m   \u001b[0mt\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mconvert_to_eager_tensor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mctx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    302\u001b[0m   \u001b[0;32mif\u001b[0m \u001b[0mshape\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    303\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/tensorflow/python/framework/constant_op.py\u001b[0m in \u001b[0;36mconvert_to_eager_tensor\u001b[0;34m(value, ctx, dtype)\u001b[0m\n\u001b[1;32m    106\u001b[0m       \u001b[0mdtype\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdtypes\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mas_dtype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdtype\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mas_datatype_enum\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    107\u001b[0m   \u001b[0mctx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mensure_initialized\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 108\u001b[0;31m   \u001b[0;32mreturn\u001b[0m \u001b[0mops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mEagerTensor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mctx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdevice_name\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    109\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    110\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mValueError\u001b[0m: Attempt to convert a value ({'instruction': \"You are an AI drawing master who is good at accurately describing keywords into prompts. Please describe the drawing of 'a woman wearing a bikini'. Please draw a drawing in the style of a comic character, including characters, colors, light and shadow, background, composition...etc. Prompts that make the drawing more similar to the Q version character style are a must! Please convert it into prompt in the form of English words. If it is a short sentence, please use () to cover the short sentence. You can customize the weight and give it the weight of prompt(:1~1.5). and Start with 'prompt:' and after completing the description, change to the next line and start with '--' to add an error prompt that will affect the drawing itself.\", 'context': '', 'response': 'PROMPT: A female character wearing a bikini with a Q-version style. (:1.2) Character: Cute, lively, and with exaggerated facial features. (:1.3) Bikini: Colorful and vibrant, with a sense of playfulness. (:1.2) Colors: Bright and eye-catching, with a preference for pastel shades. (:1.1) Lighting and Shadows: Soft, with gentle shading to add depth to the character. (:1.2) Background: A sunny beach with palm trees and a blue sky, creating a cheerful atmosphere. (:1.3) Composition: The character placed in the center, striking a cute and confident pose. (:1.5) Attention to Details: Accessorize the character with sunglasses and a beach hat.', 'category': '..."
          ]
        }
      ],
      "source": [
        "import jsonlines\n",
        "\n",
        "# 使用 jsonlines 读取 jsonl 文件\n",
        "with jsonlines.open('/content/databricks-dolly-33.jsonl') as reader:\n",
        "    data = list(reader)\n",
        "\n",
        "\n",
        "# Limit the input sequence length to 256 (to control memory usage).\n",
        "gemma_lm.preprocessor.sequence_length = 256\n",
        "# Use AdamW (a common optimizer for transformer models).\n",
        "optimizer = keras.optimizers.AdamW(\n",
        "    learning_rate=5e-5,\n",
        "    weight_decay=0.01,\n",
        ")\n",
        "# Exclude layernorm and bias terms from decay.\n",
        "optimizer.exclude_from_weight_decay(var_names=[\"bias\", \"scale\"])\n",
        "\n",
        "gemma_lm.compile(\n",
        "    loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "    optimizer=optimizer,\n",
        "    weighted_metrics=[keras.metrics.SparseCategoricalAccuracy()],\n",
        ")\n",
        "\n",
        "gemma_lm.fit(data, epochs=1, batch_size=1)\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#TEST FIRST\n",
        "# Limit the input sequence length to 256 (to control memory usage).\n",
        "gemma_lm.preprocessor.sequence_length = 256\n",
        "# Use AdamW (a common optimizer for transformer models).\n",
        "optimizer = keras.optimizers.AdamW(\n",
        "    learning_rate=5e-5,\n",
        "    weight_decay=0.01,\n",
        ")\n",
        "# Exclude layernorm and bias terms from decay.\n",
        "optimizer.exclude_from_weight_decay(var_names=[\"bias\", \"scale\"])\n",
        "\n",
        "gemma_lm.compile(\n",
        "    loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "    optimizer=optimizer,\n",
        "    weighted_metrics=[keras.metrics.SparseCategoricalAccuracy()],\n",
        ")\n",
        "gemma_lm.fit(data, epochs=1, batch_size=1)\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "tkdbWenetaBF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Limit the input sequence length to 256 (to control memory usage).\n",
        "gemma_lm.preprocessor.sequence_length = 256\n",
        "# Use AdamW (a common optimizer for transformer models).\n",
        "optimizer = keras.optimizers.AdamW(\n",
        "    learning_rate=5e-5,\n",
        "    weight_decay=0.01,\n",
        ")\n",
        "# Exclude layernorm and bias terms from decay.\n",
        "optimizer.exclude_from_weight_decay(var_names=[\"bias\", \"scale\"])\n",
        "\n",
        "gemma_lm.compile(\n",
        "    loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "    optimizer=optimizer,\n",
        "    weighted_metrics=[keras.metrics.SparseCategoricalAccuracy()],\n",
        ")\n",
        "gemma_lm.fit(data, epochs=50, batch_size=1)\n",
        "\n",
        "# 儲存為 Keras 原生格式\n",
        "gemma_lm.save('my_model4.keras')\n",
        "gemma_lm.save('/content/drive/My Drive/my_model4.keras')\n"
      ],
      "metadata": {
        "id": "xFYFEeGvhjO1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 582
        },
        "outputId": "1ace6acf-8bc5-4016-c8a4-ba51e65095f3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "RuntimeError",
          "evalue": "Unable to automatically build the model. Please build it yourself before calling fit/evaluate/predict. A model is 'built' when its variables have been created and its `self.built` attribute is True. Usually, calling the model on a batch of data is the right way to build it.\nException encountered:\n'Exception encountered when calling ReversibleEmbedding.call().\n\n\u001b[1m'NoneType' object is not callable\u001b[0m\n\nArguments received by ReversibleEmbedding.call():\n  • inputs=jnp.ndarray(shape=(1, 256), dtype=float32)\n  • reverse=False'",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-29-a04e6117b3fb>\u001b[0m in \u001b[0;36m<cell line: 16>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     14\u001b[0m     \u001b[0mweighted_metrics\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmetrics\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSparseCategoricalAccuracy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     15\u001b[0m )\n\u001b[0;32m---> 16\u001b[0;31m \u001b[0mgemma_lm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mepochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m50\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     17\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     18\u001b[0m \u001b[0;31m# 儲存為 Keras 原生格式\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras_nlp/src/utils/pipeline_model.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, sample_weight, validation_data, validation_split, **kwargs)\u001b[0m\n\u001b[1;32m    191\u001b[0m                 )\n\u001b[1;32m    192\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 193\u001b[0;31m         return super().fit(\n\u001b[0m\u001b[1;32m    194\u001b[0m             \u001b[0mx\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    195\u001b[0m             \u001b[0my\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras/src/utils/traceback_utils.py\u001b[0m in \u001b[0;36merror_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    120\u001b[0m             \u001b[0;31m# To get the full stack trace, call:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    121\u001b[0m             \u001b[0;31m# `keras.config.disable_traceback_filtering()`\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 122\u001b[0;31m             \u001b[0;32mraise\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwith_traceback\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfiltered_tb\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    123\u001b[0m         \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    124\u001b[0m             \u001b[0;32mdel\u001b[0m \u001b[0mfiltered_tb\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras/src/trainers/trainer.py\u001b[0m in \u001b[0;36m_symbolic_build\u001b[0;34m(self, iterator, data_batch)\u001b[0m\n\u001b[1;32m   1038\u001b[0m                 \u001b[0my_pred\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbackend\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcompute_output_spec\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtraining\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1039\u001b[0m             \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1040\u001b[0;31m                 raise RuntimeError(\n\u001b[0m\u001b[1;32m   1041\u001b[0m                     \u001b[0;34m\"Unable to automatically build the model. \"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1042\u001b[0m                     \u001b[0;34m\"Please build it yourself before calling \"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mRuntimeError\u001b[0m: Unable to automatically build the model. Please build it yourself before calling fit/evaluate/predict. A model is 'built' when its variables have been created and its `self.built` attribute is True. Usually, calling the model on a batch of data is the right way to build it.\nException encountered:\n'Exception encountered when calling ReversibleEmbedding.call().\n\n\u001b[1m'NoneType' object is not callable\u001b[0m\n\nArguments received by ReversibleEmbedding.call():\n  • inputs=jnp.ndarray(shape=(1, 256), dtype=float32)\n  • reverse=False'"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Limit the input sequence length to 256 (to control memory usage).\n",
        "gemma_lm.preprocessor.sequence_length = 256\n",
        "# Use AdamW (a common optimizer for transformer models).\n",
        "optimizer = keras.optimizers.AdamW(\n",
        "    learning_rate=5e-5,\n",
        "    weight_decay=0.01,\n",
        ")\n",
        "# Exclude layernorm and bias terms from decay.\n",
        "optimizer.exclude_from_weight_decay(var_names=[\"bias\", \"scale\"])\n",
        "\n",
        "gemma_lm.compile(\n",
        "    loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "    optimizer=optimizer,\n",
        "    weighted_metrics=[keras.metrics.SparseCategoricalAccuracy()],\n",
        ")\n",
        "\n",
        "# 然后再开始训练\n",
        "gemma_lm.fit(data, epochs=50, batch_size=1, callbacks=[model_checkpoint_callback])\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 228
        },
        "id": "P0AxZDSdXyKK",
        "outputId": "b39d1969-e9ab-4dd0-c562-e17d421e1e34"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "name 'input_shape' is not defined",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-27-925f2f256840>\u001b[0m in \u001b[0;36m<cell line: 1>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0m_\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mgemma_lm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrandom\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muniform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput_shape\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# 使用随机数据来初始化模型\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;31m# 然后再开始训练\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mgemma_lm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mepochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m500\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mmodel_checkpoint_callback\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'input_shape' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras.callbacks import ModelCheckpoint\n",
        "\n",
        "# ... (你的模型定义、数据加载等代码) ...\n",
        "# Limit the input sequence length to 256 (to control memory usage).\n",
        "gemma_lm.preprocessor.sequence_length = 256\n",
        "# Use AdamW (a common optimizer for transformer models).\n",
        "optimizer = keras.optimizers.AdamW(\n",
        "    learning_rate=5e-5,\n",
        "    weight_decay=0.01,\n",
        ")\n",
        "# Exclude layernorm and bias terms from decay.\n",
        "optimizer.exclude_from_weight_decay(var_names=[\"bias\", \"scale\"])\n",
        "\n",
        "gemma_lm.compile(\n",
        "    loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "    optimizer=optimizer,\n",
        "    weighted_metrics=[keras.metrics.SparseCategoricalAccuracy()],\n",
        ")\n",
        "\n",
        "# 定义 ModelCheckpoint 回调函数\n",
        "checkpoint_filepath = '/content/drive/My Drive/model/my_model_checkpoint_{epoch:04d}.keras'\n",
        "model_checkpoint_callback = ModelCheckpoint(\n",
        "    filepath=checkpoint_filepath,\n",
        "    save_weights_only=False,  # 保存整个模型（架构 + 权重）\n",
        "    monitor='loss',  # 可选：监控某个指标，例如验证集损失\n",
        "    save_best_only=False,  # 可选：只保存指标最佳的模型\n",
        "    verbose=1  # 显示保存信息\n",
        ")\n",
        "\n",
        "# 在训练时加入回调函数\n",
        "gemma_lm.fit(\n",
        "    data,\n",
        "    epochs=50,\n",
        "    batch_size=1,\n",
        "    callbacks=[model_checkpoint_callback]  # 加入回调函数列表\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QiCZPz0RW4RT",
        "outputId": "e85b503f-da6f-4d71-86e4-910f58d32257"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/50\n",
            "\u001b[1m22/22\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2s/step - loss: 2.0382 - sparse_categorical_accuracy: 0.4027\n",
            "Epoch 1: saving model to /content/drive/My Drive/model/my_model_checkpoint_0001.keras\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "gemma_lm.compile(\n",
        "    loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "    optimizer=optimizer,\n",
        "    weighted_metrics=[keras.metrics.SparseCategoricalAccuracy()],\n",
        ")\n",
        "# 儲存為 Keras 原生格式\n",
        "gemma_lm.save('/content/drive/My Drive/model/my_model2.keras')\n",
        "\n",
        "# 儲存模型權重\n",
        "gemma_lm.save_weights('/content/drive/My Drive/model/my_model_weights')\n",
        "# gemma_lm.save_weights('/content/drive/My Drive/model/my_model_weights.keras')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 339
        },
        "id": "0K7QoUL-AlrF",
        "outputId": "de0c4ea9-7da7-4be8-eddb-89f74a1c93dd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "TypeError",
          "evalue": "'NoneType' object is not callable",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-18-52a73b5e00ff>\u001b[0m in \u001b[0;36m<cell line: 7>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      5\u001b[0m )\n\u001b[1;32m      6\u001b[0m \u001b[0;31m# 儲存為 Keras 原生格式\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 7\u001b[0;31m \u001b[0mgemma_lm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msave\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'/content/drive/My Drive/model/my_model2.keras'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras/src/utils/traceback_utils.py\u001b[0m in \u001b[0;36merror_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    120\u001b[0m             \u001b[0;31m# To get the full stack trace, call:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    121\u001b[0m             \u001b[0;31m# `keras.config.disable_traceback_filtering()`\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 122\u001b[0;31m             \u001b[0;32mraise\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwith_traceback\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfiltered_tb\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    123\u001b[0m         \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    124\u001b[0m             \u001b[0;32mdel\u001b[0m \u001b[0mfiltered_tb\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras/src/backend/common/variables.py\u001b[0m in \u001b[0;36mvalue\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    217\u001b[0m             \u001b[0;31m# (anything else would be a bug, to be fixed.)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    218\u001b[0m             return self._maybe_autocast(\n\u001b[0;32m--> 219\u001b[0;31m                 \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_initializer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_shape\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_dtype\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    220\u001b[0m             )\n\u001b[1;32m    221\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_maybe_autocast\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_value\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mTypeError\u001b[0m: 'NoneType' object is not callable"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "gemma_lm.compile(\n",
        "    loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "    optimizer=optimizer,\n",
        "    weighted_metrics=[keras.metrics.SparseCategoricalAccuracy()],\n",
        ")\n",
        "\n",
        "# 保存模型权重到 Google Drive\n",
        "gemma_lm.save_weights('/content/drive/My Drive/model/my_model_weights.weights.h5')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 478
        },
        "id": "WDPXxEQfSIEg",
        "outputId": "f937649f-4982-42b8-f9f7-cd1dceec69e9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "OSError",
          "evalue": "Unable to synchronously create file (unable to truncate a file which is already open)",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mOSError\u001b[0m                                   Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-16-a48886d8fd6c>\u001b[0m in \u001b[0;36m<cell line: 8>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0;31m# 保存模型权重到 Google Drive\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m \u001b[0mgemma_lm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msave_weights\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'/content/drive/My Drive/model/my_model_weights.weights.h5'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras/src/utils/traceback_utils.py\u001b[0m in \u001b[0;36merror_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    120\u001b[0m             \u001b[0;31m# To get the full stack trace, call:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    121\u001b[0m             \u001b[0;31m# `keras.config.disable_traceback_filtering()`\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 122\u001b[0;31m             \u001b[0;32mraise\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwith_traceback\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfiltered_tb\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    123\u001b[0m         \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    124\u001b[0m             \u001b[0;32mdel\u001b[0m \u001b[0mfiltered_tb\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/h5py/_hl/files.py\u001b[0m in \u001b[0;36m__init__\u001b[0;34m(self, name, mode, driver, libver, userblock_size, swmr, rdcc_nslots, rdcc_nbytes, rdcc_w0, track_order, fs_strategy, fs_persist, fs_threshold, fs_page_size, page_buf_size, min_meta_keep, min_raw_keep, locking, alignment_threshold, alignment_interval, meta_block_size, **kwds)\u001b[0m\n\u001b[1;32m    560\u001b[0m                                  \u001b[0mfs_persist\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mfs_persist\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfs_threshold\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mfs_threshold\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    561\u001b[0m                                  fs_page_size=fs_page_size)\n\u001b[0;32m--> 562\u001b[0;31m                 \u001b[0mfid\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmake_fid\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmode\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0muserblock_size\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfapl\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfcpl\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mswmr\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mswmr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    563\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    564\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlibver\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtuple\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/h5py/_hl/files.py\u001b[0m in \u001b[0;36mmake_fid\u001b[0;34m(name, mode, userblock_size, fapl, fcpl, swmr)\u001b[0m\n\u001b[1;32m    239\u001b[0m         \u001b[0mfid\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mh5f\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcreate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mh5f\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mACC_EXCL\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfapl\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mfapl\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfcpl\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mfcpl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    240\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0mmode\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'w'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 241\u001b[0;31m         \u001b[0mfid\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mh5f\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcreate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mh5f\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mACC_TRUNC\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfapl\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mfapl\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfcpl\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mfcpl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    242\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0mmode\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'a'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    243\u001b[0m         \u001b[0;31m# Open in append mode (read/write).\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32mh5py/_objects.pyx\u001b[0m in \u001b[0;36mh5py._objects.with_phil.wrapper\u001b[0;34m()\u001b[0m\n",
            "\u001b[0;32mh5py/_objects.pyx\u001b[0m in \u001b[0;36mh5py._objects.with_phil.wrapper\u001b[0;34m()\u001b[0m\n",
            "\u001b[0;32mh5py/h5f.pyx\u001b[0m in \u001b[0;36mh5py.h5f.create\u001b[0;34m()\u001b[0m\n",
            "\u001b[0;31mOSError\u001b[0m: Unable to synchronously create file (unable to truncate a file which is already open)"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for epoch in range(476):\n",
        "    history = gemma_lm.fit(data, epochs=476, batch_size=1)\n",
        "\n",
        "    # 每 50 個 epoch 儲存一次模型\n",
        "    if (epoch + 1) % 50 == 0:\n",
        "        gemma_lm.save(f'/content/drive/My Drive/gemma_lm_epoch_{epoch+1}.keras')"
      ],
      "metadata": {
        "id": "A1x-eQgWV-qM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Assume `model` is the variable holding your trained model\n",
        "\n",
        "# 儲存整個模型\n",
        "gemma_lm.save('my_model1.h5')\n",
        "# gemma_lm.save('my_saved_model')\n",
        "\n",
        "# 儲存模型權重\n",
        "gemma_lm.save_weights('my_model_weights')\n"
      ],
      "metadata": {
        "id": "_hzBWkhnqbi7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 411
        },
        "outputId": "8efaad35-eff5-46fb-8436-b3a5f94bc4bd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:You are saving your model as an HDF5 file via `model.save()` or `keras.saving.save_model(model)`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')` or `keras.saving.save_model(model, 'my_model.keras')`. \n"
          ]
        },
        {
          "output_type": "error",
          "ename": "TypeError",
          "evalue": "'NoneType' object is not callable",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-25-d193235c79ee>\u001b[0m in \u001b[0;36m<cell line: 4>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;31m# 儲存整個模型\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0mgemma_lm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msave\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'my_model1.h5'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m \u001b[0;31m# gemma_lm.save('my_saved_model')\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras/src/utils/traceback_utils.py\u001b[0m in \u001b[0;36merror_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    120\u001b[0m             \u001b[0;31m# To get the full stack trace, call:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    121\u001b[0m             \u001b[0;31m# `keras.config.disable_traceback_filtering()`\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 122\u001b[0;31m             \u001b[0;32mraise\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwith_traceback\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfiltered_tb\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    123\u001b[0m         \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    124\u001b[0m             \u001b[0;32mdel\u001b[0m \u001b[0mfiltered_tb\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras/src/backend/common/variables.py\u001b[0m in \u001b[0;36mvalue\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    217\u001b[0m             \u001b[0;31m# (anything else would be a bug, to be fixed.)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    218\u001b[0m             return self._maybe_autocast(\n\u001b[0;32m--> 219\u001b[0;31m                 \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_initializer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_shape\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdtype\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_dtype\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    220\u001b[0m             )\n\u001b[1;32m    221\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_maybe_autocast\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_value\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mTypeError\u001b[0m: 'NoneType' object is not callable"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.callbacks import ModelCheckpoint\n",
        "\n",
        "# 建立 ModelCheckpoint 回调函数\n",
        "checkpoint_filepath = '/content/drive/My Drive/gemma_lm_checkpoint_{epoch:02d}.keras'\n",
        "model_checkpoint_callback = ModelCheckpoint(\n",
        "    filepath=checkpoint_filepath,\n",
        "    save_weights_only=False,  # 儲存整個模型 (架構+權重)\n",
        "    save_best_only=False,  # 儲存每個 epoch 的模型\n",
        "    monitor='loss',  # 可以根據驗證集上的指標來儲存最佳模型\n",
        "    verbose=1  # 顯示儲存訊息\n",
        ")\n",
        "\n",
        "# 訓練模型時，加入回调函数\n",
        "gemma_lm.fit(data, epochs=500, batch_size=1, callbacks=[model_checkpoint_callback])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 564
        },
        "id": "rL8Ddl0C_KZy",
        "outputId": "278f2335-8da8-4fdf-ce97-d92e49db3b9e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "RuntimeError",
          "evalue": "Unable to automatically build the model. Please build it yourself before calling fit/evaluate/predict. A model is 'built' when its variables have been created and its `self.built` attribute is True. Usually, calling the model on a batch of data is the right way to build it.\nException encountered:\n'Exception encountered when calling ReversibleEmbedding.call().\n\n\u001b[1m'NoneType' object is not callable\u001b[0m\n\nArguments received by ReversibleEmbedding.call():\n  • inputs=jnp.ndarray(shape=(1, 256), dtype=float32)\n  • reverse=False'",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-26-5e0e23c47c26>\u001b[0m in \u001b[0;36m<cell line: 14>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     12\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m \u001b[0;31m# 訓練模型時，加入回调函数\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 14\u001b[0;31m \u001b[0mgemma_lm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mepochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m500\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mmodel_checkpoint_callback\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras_nlp/src/utils/pipeline_model.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, sample_weight, validation_data, validation_split, **kwargs)\u001b[0m\n\u001b[1;32m    191\u001b[0m                 )\n\u001b[1;32m    192\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 193\u001b[0;31m         return super().fit(\n\u001b[0m\u001b[1;32m    194\u001b[0m             \u001b[0mx\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    195\u001b[0m             \u001b[0my\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras/src/utils/traceback_utils.py\u001b[0m in \u001b[0;36merror_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    120\u001b[0m             \u001b[0;31m# To get the full stack trace, call:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    121\u001b[0m             \u001b[0;31m# `keras.config.disable_traceback_filtering()`\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 122\u001b[0;31m             \u001b[0;32mraise\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwith_traceback\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfiltered_tb\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    123\u001b[0m         \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    124\u001b[0m             \u001b[0;32mdel\u001b[0m \u001b[0mfiltered_tb\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/keras/src/trainers/trainer.py\u001b[0m in \u001b[0;36m_symbolic_build\u001b[0;34m(self, iterator, data_batch)\u001b[0m\n\u001b[1;32m   1038\u001b[0m                 \u001b[0my_pred\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbackend\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcompute_output_spec\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtraining\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1039\u001b[0m             \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1040\u001b[0;31m                 raise RuntimeError(\n\u001b[0m\u001b[1;32m   1041\u001b[0m                     \u001b[0;34m\"Unable to automatically build the model. \"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1042\u001b[0m                     \u001b[0;34m\"Please build it yourself before calling \"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mRuntimeError\u001b[0m: Unable to automatically build the model. Please build it yourself before calling fit/evaluate/predict. A model is 'built' when its variables have been created and its `self.built` attribute is True. Usually, calling the model on a batch of data is the right way to build it.\nException encountered:\n'Exception encountered when calling ReversibleEmbedding.call().\n\n\u001b[1m'NoneType' object is not callable\u001b[0m\n\nArguments received by ReversibleEmbedding.call():\n  • inputs=jnp.ndarray(shape=(1, 256), dtype=float32)\n  • reverse=False'"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bx3m8f1dB7nk"
      },
      "source": [
        "### Note on mixed precision fine-tuning on NVIDIA GPUs or AMD GPUs\n",
        "\n",
        "Full precision is recommended for fine-tuning. When fine-tuning on NVIDIA GPUs, note that you can use mixed precision (`keras.mixed_precision.set_global_policy('mixed_bfloat16')`) to speed up training with minimal effect on training quality. Mixed precision fine-tuning does consume more memory so is useful only on larger GPUs.\n",
        "\n",
        "\n",
        "For inference, half-precision (`keras.config.set_floatx(\"bfloat16\")`) will work and save memory while mixed precision is not applicable.\n",
        "google colab do not have AMD GPUs.  Try to use AMD GPUs at home.\n",
        "建議使用全精度進行微調。在 NVIDIA GPU 上進行微調時，請注意，您可以使用混合精確度 (keras.mixed_ precision.set_global_policy('mixed_bfloat16')) 來加速訓練，同時對訓練品質的影響最小。混合精度微調確實會消耗更多內存，因此僅在較大的 GPU 上有用。\n",
        "\n",
        "對於推理，半精度 (keras.config.set_floatx(\"bfloat16\")) 將起作用並節省內存，而混合精度不適用。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "T0lHxEDX03gp"
      },
      "outputs": [],
      "source": [
        "# Uncomment the line below if you want to enable mixed precision training on GPUs\n",
        "# keras.mixed_precision.set_global_policy('mixed_bfloat16')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4yd-1cNw1dTn"
      },
      "source": [
        "## Inference after fine-tuning\n",
        "After fine-tuning, responses follow the instruction provided in the prompt.\n",
        "微調後，將按照提示中提供的說明進行回應。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H55JYJ1a1Kos"
      },
      "source": [
        "### Europe Trip Prompt"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Y7cDJHy8WfCB",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "15435701-bf16-4d11-b0e0-bc955d7d6c39"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Instruction:\n",
            "What should I do on a trip to Europe?\n",
            "\n",
            "Response:\n",
            "It's always fun to travel.  If you are going to Europe, here are some things you might want to do.  If you are traveling by air, you should check your bag at the gate.  You can bring a small personal item on the plane, such as a purse or backpack.  When you land, you should pick up your bags from the carousel.  If you have an airline that has lost your bags, it can be frustrating to wait for them.  Once you have your bags, you can go through customs and immigration.  You might have to fill out a customs declaration form and answer a few questions.  You can then get your luggage through the luggage claim area and pick up your car.  You can then drive to your hotel or Airbnb.  You can then get settled in and enjoy your trip to Europe.\n"
          ]
        }
      ],
      "source": [
        "prompt = template.format(\n",
        "    instruction=\"What should I do on a trip to Europe?\",\n",
        "    response=\"\",\n",
        ")\n",
        "sampler = keras_nlp.samplers.TopKSampler(k=5, seed=2)\n",
        "gemma_lm.compile(sampler=sampler)\n",
        "print(gemma_lm.generate(prompt, max_length=256))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OXP6gg2mjs6u"
      },
      "source": [
        "The model now recommends places to visit in Europe."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H7nVd8Mi1Yta"
      },
      "source": [
        "### ELI5 Photosynthesis Prompt"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "yb3pXS48ooGa"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "X-2sYl2jqwl7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2a1fb73e-c61c-45e0-b041-f129f9c59b58"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Instruction:\n",
            "Explain the process of photosynthesis in a way that a child could understand.\n",
            "\n",
            "Response:\n",
            "The process of photosynthesis is the process where plants convert sunlight into food, which they use for their growth. Plants have chlorophyll, a pigment that absorbs light energy to power the process of photosynthesis. The chlorophyll is present in the leaves. The process of photosynthesis happens in the chloroplast of leaves, where the light energy is absorbed, and water and carbon dioxide is used to create sugar, oxygen and energy for growth. This happens in two stages: The first stage is called the light dependent stage and the second is called the light independent stage. In the light-dependent stage, water is split to produce hydrogen ions, oxygen, and electrons. The electrons are used to create NADPH. The hydrogen ions and electrons are used to produce ATP, the energy currency of the cell. In the light-independent stage, the sugar created in the light dependent stage is used for energy and growth, and the carbon dioxide is used to create sugar and oxygen.\n"
          ]
        }
      ],
      "source": [
        "prompt = template.format(\n",
        "    instruction=\"Explain the process of photosynthesis in a way that a child could understand.\",\n",
        "    response=\"\",\n",
        ")\n",
        "print(gemma_lm.generate(prompt, max_length=256))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PCmAmqrvkEhc"
      },
      "source": [
        "The model now explains photosynthesis in simpler terms.\n",
        "該模型現在用更簡單的術語解釋了光合作用。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I8kFG12l0mVe"
      },
      "source": [
        "Note that for demonstration purposes, this tutorial fine-tunes the model on a small subset of the dataset for just one epoch and with a low LoRA rank value. To get better responses from the fine-tuned model, you can experiment with:\n",
        "\n",
        "1. Increasing the size of the fine-tuning dataset\n",
        "2. Training for more steps (epochs)\n",
        "3. Setting a higher LoRA rank\n",
        "4. Modifying the hyperparameter values such as `learning_rate` and `weight_decay`.\n",
        "請注意，出於演示目的，本教程僅在一個時期內對資料集的一小部分子集上的模型進行微調，並使用較低的 LoRA 等級值。為了從微調模型中獲得更好的反應，您可以嘗試：\n",
        "\n",
        "增加微調資料集的大小\n",
        "訓練更多步驟（epochs）\n",
        "設定更高的 LoRA 等級\n",
        "修改超參數值，例如learning_rate和weight_decay。"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "save model"
      ],
      "metadata": {
        "id": "rZRMWpPkmSnH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Assume `model` is the variable holding your trained model\n",
        "\n",
        "# Save the entire model\n",
        "model.save('path_to_save_model/my_model')\n",
        "\n",
        "# Save the model weights\n",
        "model.save_weights('path_to_save_weights/my_model_weights')\n"
      ],
      "metadata": {
        "id": "HZVJG08ca52q"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gSsRdeiof_rJ"
      },
      "source": [
        "## Summary and next steps\n",
        "\n",
        "This tutorial covered LoRA fine-tuning on a Gemma model using KerasNLP. Check out the following docs next:\n",
        "\n",
        "* Learn how to [generate text with a Gemma model](https://ai.google.dev/gemma/docs/get_started).\n",
        "* Learn how to perform [distributed fine-tuning and inference on a Gemma model](https://ai.google.dev/gemma/docs/distributed_tuning).\n",
        "* Learn how to [use Gemma open models with Vertex AI](https://cloud.google.com/vertex-ai/docs/generative-ai/open-models/use-gemma).\n",
        "* Learn how to [fine-tune Gemma using KerasNLP and deploy to Vertex AI](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_gemma_kerasnlp_to_vertexai.ipynb).\n",
        "摘要與後續步驟\n",
        "本教程介紹了使用 KerasNLP 對 Gemma 模型進行 LoRA 微調。接下來查看以下文檔：\n",
        "\n",
        "了解如何使用 Gemma 模型產生文字。\n",
        "了解如何對 Gemma 模型執行分散式微調和推理。\n",
        "了解如何將 Gemma 開放式模型與 Vertex AI 結合使用。\n",
        "了解如何使用 KerasNLP 微調 Gemma 並部署到 Vertex AI。"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}