УрФУ_Stable_Diffusion_Разное.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/dayekb/urfu_master_class/blob/main/%D0%A3%D1%80%D0%A4%D0%A3_Stable_Diffusion_%D0%A0%D0%B0%D0%B7%D0%BD%D0%BE%D0%B5.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "В этом Блокноте мы пользуемся GPU \"любезно\" предоставленной Google"
      ],
      "metadata": {
        "id": "7v_3yKTubRo6"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "![image.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "hjfF0va5bLE-"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zA9lDn4etoIr"
      },
      "source": [
        "# Импорт и установка хорошего)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "XErcan23E0fo"
      },
      "outputs": [],
      "source": [
        "!pip install diffusers transformers scipy ftfy \"ipywidgets>=7,<8\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "O-PeuZBDFFSF"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import google.colab.output\n",
        "\n",
        "import torch\n",
        "from torch import autocast\n",
        "\n",
        "from diffusers import StableDiffusionPipeline\n",
        "\n",
        "from huggingface_hub import notebook_login\n",
        "from PIL import Image, ImageDraw\n",
        "\n",
        "device = 'cuda'"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZQjBgCK8Jubb"
      },
      "source": [
        "# Pipeline по Умолчанию"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vDFnNxCBvw-H"
      },
      "source": [
        "Нам нужно знать название модели (полное)\n",
        "\n",
        "  [Модель](https://huggingface.co/stabilityai/stable-diffusion-2-base)\n",
        "\n",
        " Пользуемся готовым решением - загружаем уже обученную модель по имени (всё остальное ребята из Hugging Face сделали за нас)\n",
        "\n",
        " Функция с которой мы будем работать - `pipe` (пайплайн)\n",
        "\n",
        "  Нажимаем и ждем пару минут)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ye1YYP_1FKI4"
      },
      "outputs": [],
      "source": [
        "model_id = \"stabilityai/stable-diffusion-2-base\"\n",
        "\n",
        "pipe = StableDiffusionPipeline.from_pretrained(model_id, \n",
        "                                               torch_dtype=torch.float16,\n",
        "                                               low_cpu_mem_usage=False)\n",
        "\n",
        "pipe = pipe.to(device)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Запуск"
      ],
      "metadata": {
        "id": "OjmtSejTVimT"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "V0Kx8UQcMdt_"
      },
      "source": [
        "*Интерфейс* взаимодействия простой\n",
        "\n",
        "Есть `prompt` - текст \"чего мы хотим\"\n",
        "\n",
        "Он подается на вход нашего `pipe`\n",
        "\n",
        "И берем только изображение"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fl6yZ7wOMdHr"
      },
      "outputs": [],
      "source": [
        "prompt = 'Pikachu in field, hd, realistic '\n",
        "with autocast(device):\n",
        "    image = pipe(prompt)['images'][0]\n",
        "image"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rlKn786UNjjM"
      },
      "source": [
        "Как-бы проблема\n",
        "\n",
        "Если запустить еще раз - будет новое"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "A8_f6vlSNsoz"
      },
      "outputs": [],
      "source": [
        "prompt = 'Pikachu in field, hd, realistic '\n",
        "with autocast(device):\n",
        "    image = pipe(prompt)['images'][0]\n",
        "image"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Повторяемая генерация"
      ],
      "metadata": {
        "id": "w28UzM7tVeF_"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "65baeb64"
      },
      "source": [
        "Чтобы повторно использовать изображение (например чуточку передалать), нам нужно самим генерировать Скрытое пространство. В противном случае конвейер сделает это внутри, и у нас не будет возможности их воспроизвести.\n",
        "\n",
        "Скрытое Пространство — это начальный случайный гауссовский шум, который преобразуется в реальные изображения в процессе диффузии.\n",
        "\n",
        "Чтобы сгенерировать их, мы будем использовать другое случайное начальное число для каждого Скрытое Пространство и мы сохраним их, чтобы мы могли повторно использовать их позже."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zHcW1YNAN_EZ"
      },
      "outputs": [],
      "source": [
        "generator = torch.Generator(device=device)\n",
        "seed = generator.seed()\n",
        "generator = generator.manual_seed(seed)\n",
        "\n",
        "width = 512\n",
        "height = 512\n",
        "\n",
        "latents = torch.randn(\n",
        "    (1, pipe.unet.config.in_channels, height // 8, width // 8),\n",
        "    generator = generator,\n",
        "    device = device\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5hpUZhpRPWtL"
      },
      "outputs": [],
      "source": [
        "seed"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c1YQaTjPPgvc"
      },
      "source": [
        "Теперь подаем на вход не только текстовое описание ( `prompt` ), но и шум из которого мы рисуем картинку (`latents`)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7dXMy3DGONMt"
      },
      "outputs": [],
      "source": [
        "prompt = 'Pikachu in field, hd, realistic '\n",
        "with autocast(device):\n",
        "    image = pipe(prompt,\n",
        "                 latents = latents,)['images'][0]\n",
        "image"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5CkdaciTOWrm"
      },
      "source": [
        "Проверяем на повторяемость"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7t4zGI0aOZud"
      },
      "outputs": [],
      "source": [
        "prompt = 'Pikachu in field, hd, realistic '\n",
        "with autocast(device):\n",
        "    image = pipe(prompt,\n",
        "                 latents = latents,)['images'][0]\n",
        "image"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Про то как сделать что-то из ничего"
      ],
      "metadata": {
        "id": "cDCow6knVop8"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kCNVRMfKRBj4"
      },
      "source": [
        "Зафиксировав сид мы можем \"проследить\" историю создания"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7PgdYDFjI0tG"
      },
      "outputs": [],
      "source": [
        "def image_grid(imgs, rows, cols):\n",
        "    assert len(imgs) == rows*cols\n",
        "\n",
        "    w, h = imgs[0].size\n",
        "    grid = Image.new('RGB', size=(cols*w, rows*h))\n",
        "    grid_w, grid_h = grid.size\n",
        "    \n",
        "    for i, img in enumerate(imgs):\n",
        "        grid.paste(img, box=(i%cols*w, i//cols*h))\n",
        "    return grid"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "6cqSRW41RPkW"
      },
      "outputs": [],
      "source": [
        "num_inference_steps=[1, 2, 3, 4, 5,\n",
        "                     6, 7, 8, 9, 10,\n",
        "                     15,20,25,30,40,50,]\n",
        "images = []\n",
        "with autocast(device):\n",
        "    for num_i in num_inference_steps:\n",
        "\n",
        "        image = pipe(\n",
        "            prompt,\n",
        "            latents = latents,\n",
        "            num_inference_steps=num_i\n",
        "                )[\"images\"]\n",
        "        images.append(image[0])"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Немного страшно, но что поделать"
      ],
      "metadata": {
        "id": "Vnt8TPxySVhW"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "background_save": true
        },
        "id": "HN4MYD7KRPkX"
      },
      "outputs": [],
      "source": [
        "image_grid(images, 4, 4)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Жаловаться бывает хорошо"
      ],
      "metadata": {
        "id": "7km4IMrtV3aw"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HDrF3nH3Nr7r"
      },
      "source": [
        "Можно сказать модели не только \"как рисовать\", но и \"как точно не надо рисовать\"  `negative_prompt`"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zoEBpZ4TQMWe"
      },
      "outputs": [],
      "source": [
        "prompt = 'Pikachu in field, hd, realistic'\n",
        "negative_prompt = 'ugly, monochrome, lowres, bad anatomy, worst quality, low quality, crop, watermark, (out of frame)'\n",
        "\n",
        "with autocast(device):\n",
        "    image = pipe(prompt,\n",
        "                 latents = latents,\n",
        "                 negative_prompt = negative_prompt)['images'][0]\n",
        "image"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## С чем еще можно играться"
      ],
      "metadata": {
        "id": "H4Qy3UV9V683"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Можно еще варьировать параметр `guidance_scale`\n",
        "\n",
        "Своеобразный \"вес\" текстового описания\n",
        "\n",
        "Чем меньше тем более \"случайным\" будет картинка"
      ],
      "metadata": {
        "id": "celUvOYsUNlC"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "guidance_scale=[1, 2, 3, 4, 5, \n",
        "                6, 7, 8, 9, 10, \n",
        "                10.5, 11, 11.5, 12, 12.5,]\n",
        "images = []\n",
        "with autocast(device):\n",
        "    for gs_i in guidance_scale:\n",
        "\n",
        "        image = pipe(\n",
        "            prompt,\n",
        "            latents = latents,\n",
        "            guidance_scale=gs_i\n",
        "                )[\"images\"]\n",
        "        images.append(image[0])"
      ],
      "metadata": {
        "id": "DtUkWBICUIgT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "image_grid(images, 5, 3)"
      ],
      "metadata": {
        "id": "Yy_BTNKyVU-H"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Массовое рисование"
      ],
      "metadata": {
        "id": "y6YRmpMlW4Rz"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Можно сразу \"пачкой\" рисовать, а потом отбирать \"что получше\""
      ],
      "metadata": {
        "id": "ZW8VbOa3SvAH"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "faZob5viMM0f"
      },
      "outputs": [],
      "source": [
        "num_images = 4\n",
        "\n",
        "generator = torch.Generator(device=device)\n",
        "\n",
        "latents = None\n",
        "seeds = []\n",
        "for _ in range(num_images):\n",
        "    seed = generator.seed()\n",
        "    seeds.append(seed)\n",
        "    generator = generator.manual_seed(seed)\n",
        "    \n",
        "    image_latents = torch.randn(\n",
        "        (1, pipe.unet.config.in_channels, height // 8, width // 8),\n",
        "        generator = generator,\n",
        "        device = device\n",
        "    )\n",
        "    latents = image_latents if latents is None else torch.cat((latents, image_latents))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kzI35BmSFKRm"
      },
      "outputs": [],
      "source": [
        "prompt = 'Pikachu in field, hd, realistic'\n",
        "negative_prompt = 'ugly, monochrome, lowres, bad anatomy, worst quality, low quality, crop, watermark, (out of frame)'\n",
        "\n",
        "with autocast(device):\n",
        "    images = pipe(\n",
        "        [prompt] * num_images,\n",
        "        latents = latents,\n",
        "        negative_prompt = [negative_prompt]* num_images\n",
        "    )[\"images\"]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HC7AbiZwND2g"
      },
      "outputs": [],
      "source": [
        "image_grid(images, 2, 2)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZlxY0YeQPA2m"
      },
      "outputs": [],
      "source": [
        "seed = seeds[3]   # Например выбирем 4-ый\n",
        "seed"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "h60RHCKNPE9Y"
      },
      "outputs": [],
      "source": [
        "generator.manual_seed(seed)\n",
        "\n",
        "latents = torch.randn(\n",
        "    (1, pipe.unet.config.in_channels, height // 8, width // 8),\n",
        "    generator = generator,\n",
        "    device = device\n",
        ")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "3833bB-8PH_y"
      },
      "outputs": [],
      "source": [
        "with autocast(device):\n",
        "    image = pipe(\n",
        "        prompt,\n",
        "        latents = latents,\n",
        "        negative_prompt  = negative_prompt\n",
        "    )[\"images\"]\n",
        "    \n",
        "image[0]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zzqyoNLGscbl"
      },
      "source": [
        "# Дополнительно можно ознакомиться "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5fSi0v41sO6A"
      },
      "source": [
        "**Генерация изображений Img2Img**<br />\n",
        "https://colab.research.google.com/drive/1enyAeGJ5IP8nZSn9hdpMi52TPH9IsEDz?usp=sharing<br />\n",
        "Рисуем на базе наброска, или заменяем что-то по маске (Inpaint)<br />\n",
        "Так же небольшая демонстрация Gradio для \"интерфейса из ничего\"<br /> \n",
        "\n",
        "\n",
        "**Рисуем изображения по контурам другого**<br />\n",
        "https://colab.research.google.com/drive/1B_xAsdqzwkfZMNrs_r3jbbwiNaxv4BPO?usp=sharing<br />\n",
        "Знакомство с ControlNet<br />\n",
        "Рисуем по границам (контурам)\n",
        "\n",
        "\n",
        "**Генерируем изображения используя позы другого**<br />\n",
        "https://colab.research.google.com/drive/1_sh1_Vi9CHfhEIK2kOx0Z0OfN5hHxJAa?usp=sharing<br />\n",
        "Опять ControlNet<br />\n",
        "Позы (и для этого есть нейронка)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gZZbpKrBsqit"
      },
      "source": [
        "Разные специализированные рисовалки \n",
        "\n",
        "В основном, конечно, аниме\n",
        "\n",
        "https://huggingface.co/nitrosocke/Ghibli-Diffusion\n",
        "\n",
        "https://huggingface.co/nitrosocke/Arcane-Diffusion\n",
        "\n",
        "https://huggingface.co/hakurei/waifu-diffusion\n",
        "\n",
        "https://huggingface.co/dreamlike-art/dreamlike-photoreal-2.0 (наконец не аниме - типа фотореализм)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VMp_63OSWAgQ"
      },
      "outputs": [],
      "source": []
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "machine_shape": "hm",
      "provenance": [],
      "include_colab_link": true
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}