Gemma/Keras_Gemma_2_Quickstart.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Tce3stUlHN0L"
      },
      "source": [
        "##### Copyright 2024 Google LLC."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "cellView": "form",
        "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": "PXNm5_p_oxMF"
      },
      "source": [
        "This is a quick demo of Gemma running on KerasNLP.\n",
        "\n",
        "Note that you will need a large GPU (e.g. A100) to run this as well.\n",
        "\n",
        "General Keras reading:\n",
        "- [Getting started with Keras](https://keras.io/getting_started/)\n",
        "- [Getting started with KerasNLP](https://keras.io/guides/keras_nlp/getting_started/)\n",
        "\n",
        "<table align=\"left\">\n",
        "  <td>\n",
        "    <a target=\"_blank\" href=\"https://colab.research.google.com/github/google-gemini/gemma-cookbook/blob/main/Gemma/Keras_Gemma_2_Quickstart.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />Run in Google Colab</a>\n",
        "  </td>\n",
        "</table>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VzFhJDayXroT"
      },
      "source": [
        "## Access"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "DrBoa_Urw9Vx"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "from google.colab import userdata\n",
        "\n",
        "os.environ[\"KAGGLE_USERNAME\"] = userdata.get('KAGGLE_USERNAME')\n",
        "os.environ[\"KAGGLE_KEY\"] = userdata.get('KAGGLE_KEY')\n",
        "\n",
        "os.environ[\"KERAS_BACKEND\"] = \"jax\"  # Or \"tensorflow\" or \"torch\"."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "z9oy3QUmXtSd"
      },
      "source": [
        "## Installation"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bMboT70Xop8G"
      },
      "outputs": [],
      "source": [
        "# Install all deps\n",
        "!pip install -U keras-nlp\n",
        "!pip install -U keras==3.3.3"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FX47AUYrXwLK"
      },
      "source": [
        "## Quickstart"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "ww83zI9ToPso"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/model.safetensors...\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/model.safetensors.index.json...\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/metadata.json...\n",
            "100%|██████████| 143/143 [00:00<00:00, 153kB/s]\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/task.json...\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/config.json...\n",
            "100%|██████████| 780/780 [00:00<00:00, 884kB/s]\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/model.safetensors...\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/model.safetensors.index.json...\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/model.weights.h5...\n",
            "100%|██████████| 17.2G/17.2G [04:22<00:00, 70.5MB/s]\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/model.safetensors...\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/model.safetensors.index.json...\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/preprocessor.json...\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/tokenizer.json...\n",
            "100%|██████████| 315/315 [00:00<00:00, 434kB/s]\n",
            "Downloading from https://www.kaggle.com/api/v1/models/keras/gemma2/keras/gemma2_9b_en/1/download/assets/tokenizer/vocabulary.spm...\n",
            "100%|██████████| 4.04M/4.04M [00:00<00:00, 14.6MB/s]\n"
          ]
        },
        {
          "data": {
            "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\"</span>\n",
              "</pre>\n"
            ],
            "text/plain": [
              "\u001b[1mPreprocessor: \"gemma_causal_lm_preprocessor\"\u001b[0m\n"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "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"
            ],
            "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"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "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\"</span>\n",
              "</pre>\n"
            ],
            "text/plain": [
              "\u001b[1mModel: \"gemma_causal_lm\"\u001b[0m\n"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "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\">3584</span>)        │   <span style=\"color: #00af00; text-decoration-color: #00af00\">9,241,705,984</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\">917,504,000</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"
            ],
            "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;34m3584\u001b[0m)        │   \u001b[38;5;34m9,241,705,984\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;34m917,504,000\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"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "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\">9,241,705,984</span> (17.21 GB)\n",
              "</pre>\n"
            ],
            "text/plain": [
              "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m9,241,705,984\u001b[0m (17.21 GB)\n"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "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\">9,241,705,984</span> (17.21 GB)\n",
              "</pre>\n"
            ],
            "text/plain": [
              "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m9,241,705,984\u001b[0m (17.21 GB)\n"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "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"
            ],
            "text/plain": [
              "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import keras_nlp\n",
        "import keras\n",
        "\n",
        "# Run at half precision.\n",
        "keras.config.set_floatx(\"bfloat16\")\n",
        "\n",
        "# using 9B base model\n",
        "gemma_lm = keras_nlp.models.GemmaCausalLM.from_preset(\"gemma2_9b_en\")\n",
        "gemma_lm.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "aae5GHrdpj2_"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "It was a dark and stormy night.\n",
            "\n",
            "The wind was howling, the rain was pouring, and the thunder was rumbling.\n",
            "\n",
            "I was sitting in my living room, watching the storm rage outside.\n",
            "\n",
            "Suddenly, I heard a knock at the door.\n",
            "\n",
            "I got up and opened it, and there stood a man in a black cloak.\n",
            "\n",
            "He had a strange look in his eyes, and he was holding a lantern.\n",
            "\n",
            "\"Who are you?\" I asked.\n",
            "\n",
            "\"I am the storm,\" he replied.\n",
            "\n",
            "\"And I have come to take you away.\"\n",
            "\n",
            "I was terrified, but I couldn't move.\n",
            "\n",
            "The man in the black cloak grabbed my arm and pulled me out into the storm.\n",
            "\n",
            "We walked for what seemed like hours, until we came to a clearing in the woods.\n",
            "\n",
            "There, the man in the black cloak stopped and turned to me.\n",
            "\n",
            "\"You are mine now,\" he said.\n",
            "\n",
            "\"And I will take you to my castle.\"\n",
            "\n",
            "I tried to fight him off, but he was too strong.\n",
            "\n",
            "He dragged me into the castle, and I was never seen again.\n",
            "\n",
            "The end.\n"
          ]
        }
      ],
      "source": [
        "result = gemma_lm.generate(\"It was a dark and stormy night.\", max_length=256)\n",
        "print(result)"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "Keras_Gemma_2_Quickstart.ipynb",
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}