03_Image_Generation_App.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "2c635960",
   "metadata": {},
   "source": [
    "# **Gradio:** Image Generation App 🎨"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a2c186d0",
   "metadata": {},
   "source": [
    "Load your HF API key and relevant Python libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e0447fb8-b684-4835-9a31-1d0a824d2607",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import io\n",
    "import IPython.display\n",
    "from PIL import Image\n",
    "import base64 \n",
    "from dotenv import load_dotenv, find_dotenv\n",
    "_ = load_dotenv(find_dotenv()) # read local .env file\n",
    "hf_api_key = os.environ['HF_API_KEY']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2e1b5677-271a-46ac-a02e-656f2e315e15",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Helper function\n",
    "import requests, json\n",
    "\n",
    "#Text-to-image endpoint\n",
    "def get_completion(inputs, parameters=None, ENDPOINT_URL=os.environ['HF_API_TTI_BASE']):\n",
    "    headers = {\n",
    "      \"Authorization\": f\"Bearer {hf_api_key}\",\n",
    "      \"Content-Type\": \"application/json\"\n",
    "    }   \n",
    "    data = { \"inputs\": inputs }\n",
    "    if parameters is not None:\n",
    "        data.update({\"parameters\": parameters})\n",
    "    response = requests.request(\"POST\",\n",
    "                                ENDPOINT_URL,\n",
    "                                headers=headers,\n",
    "                                data=json.dumps(data))\n",
    "    return json.loads(response.content.decode(\"utf-8\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "764a85ed",
   "metadata": {},
   "source": [
    "## Building an image generation app "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c05fab52",
   "metadata": {},
   "source": [
    "Here we are going to run `runwayml/stable-diffusion-v1-5` using the `🧨 diffusers` library."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "47bde488",
   "metadata": {},
   "source": [
    "### How about running it locally?\n",
    "The code would look very similar if you were running it locally instead of from an API.\n",
    "```py\n",
    "from diffusers import DiffusionPipeline\n",
    "\n",
    "pipeline = DiffusionPipeline.from_pretrained(\"runwayml/stable-diffusion-v1-5\")\n",
    "\n",
    "def get_completion(prompt):\n",
    "    return pipeline(prompt).images[0]    \n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a28748bf-2618-4b8c-aa6f-7aad9c573af8",
   "metadata": {},
   "outputs": [],
   "source": [
    "prompt = \"a dog in a park\"\n",
    "\n",
    "result = get_completion(prompt)\n",
    "IPython.display.HTML(f'<img src=\"data:image/png;base64,{result}\" />')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1ee90440",
   "metadata": {},
   "source": [
    "## Generating with `gr.Interface()`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "024026ce-7832-4749-8992-50d31d61fcda",
   "metadata": {},
   "outputs": [],
   "source": [
    "import gradio as gr \n",
    "\n",
    "#A helper function to convert the PIL image to base64\n",
    "#so you can send it to the API\n",
    "def base64_to_pil(img_base64):\n",
    "    base64_decoded = base64.b64decode(img_base64)\n",
    "    byte_stream = io.BytesIO(base64_decoded)\n",
    "    pil_image = Image.open(byte_stream)\n",
    "    return pil_image\n",
    "\n",
    "def generate(prompt):\n",
    "    output = get_completion(prompt)\n",
    "    result_image = base64_to_pil(output)\n",
    "    return result_image\n",
    "\n",
    "gr.close_all()\n",
    "demo = gr.Interface(fn=generate,\n",
    "                    inputs=[gr.Textbox(label=\"Your prompt\")],\n",
    "                    outputs=[gr.Image(label=\"Result\")],\n",
    "                    title=\"Image Generation with Stable Diffusion\",\n",
    "                    description=\"Generate any image with Stable Diffusion\",\n",
    "                    allow_flagging=\"never\",\n",
    "                    examples=[\"the spirit of a tamagotchi wandering in the city of Vienna\",\"a mecha robot in a favela\"])\n",
    "\n",
    "demo.launch(share=True, server_port=int(os.environ['PORT1']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "37d16450",
   "metadata": {},
   "outputs": [],
   "source": [
    "demo.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3e7a7330",
   "metadata": {},
   "source": [
    "## Building a more advanced interface"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0c844370-4c46-4115-858e-f5cd5fe337ab",
   "metadata": {},
   "outputs": [],
   "source": [
    "import gradio as gr \n",
    "\n",
    "#A helper function to convert the PIL image to base64 \n",
    "# so you can send it to the API\n",
    "def base64_to_pil(img_base64):\n",
    "    base64_decoded = base64.b64decode(img_base64)\n",
    "    byte_stream = io.BytesIO(base64_decoded)\n",
    "    pil_image = Image.open(byte_stream)\n",
    "    return pil_image\n",
    "\n",
    "def generate(prompt, negative_prompt, steps, guidance, width, height):\n",
    "    params = {\n",
    "        \"negative_prompt\": negative_prompt,\n",
    "        \"num_inference_steps\": steps,\n",
    "        \"guidance_scale\": guidance,\n",
    "        \"width\": width,\n",
    "        \"height\": height\n",
    "    }\n",
    "    \n",
    "    output = get_completion(prompt, params)\n",
    "    pil_image = base64_to_pil(output)\n",
    "    return pil_image\n",
    "\n",
    "gr.close_all()\n",
    "demo = gr.Interface(fn=generate,\n",
    "                    inputs=[\n",
    "                        gr.Textbox(label=\"Your prompt\"),\n",
    "                        gr.Textbox(label=\"Negative prompt\"),\n",
    "                        gr.Slider(label=\"Inference Steps\", minimum=1, maximum=100, value=25,\n",
    "                                 info=\"In how many steps will the denoiser denoise the image?\"),\n",
    "                        gr.Slider(label=\"Guidance Scale\", minimum=1, maximum=20, value=7, \n",
    "                                  info=\"Controls how much the text prompt influences the result\"),\n",
    "                        gr.Slider(label=\"Width\", minimum=64, maximum=512, step=64, value=512),\n",
    "                        gr.Slider(label=\"Height\", minimum=64, maximum=512, step=64, value=512),\n",
    "                    ],\n",
    "                    outputs=[gr.Image(label=\"Result\")],\n",
    "                    title=\"Image Generation with Stable Diffusion\",\n",
    "                    description=\"Generate any image with Stable Diffusion\",\n",
    "                    allow_flagging=\"never\"\n",
    "                    )\n",
    "\n",
    "demo.launch(share=True, server_port=int(os.environ['PORT2']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "71999364",
   "metadata": {},
   "outputs": [],
   "source": [
    "demo.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3a95b525",
   "metadata": {},
   "source": [
    "## `gr.Blocks()` to the rescue!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e6818b5e-1342-439b-b129-a8904719d09c",
   "metadata": {},
   "outputs": [],
   "source": [
    "with gr.Blocks() as demo:\n",
    "    gr.Markdown(\"# Image Generation with Stable Diffusion\")\n",
    "    prompt = gr.Textbox(label=\"Your prompt\")\n",
    "    with gr.Row():\n",
    "        with gr.Column():\n",
    "            negative_prompt = gr.Textbox(label=\"Negative prompt\")\n",
    "            steps = gr.Slider(label=\"Inference Steps\", minimum=1, maximum=100, value=25,\n",
    "                      info=\"In many steps will the denoiser denoise the image?\")\n",
    "            guidance = gr.Slider(label=\"Guidance Scale\", minimum=1, maximum=20, value=7,\n",
    "                      info=\"Controls how much the text prompt influences the result\")\n",
    "            width = gr.Slider(label=\"Width\", minimum=64, maximum=512, step=64, value=512)\n",
    "            height = gr.Slider(label=\"Height\", minimum=64, maximum=512, step=64, value=512)\n",
    "            btn = gr.Button(\"Submit\")\n",
    "        with gr.Column():\n",
    "            output = gr.Image(label=\"Result\")\n",
    "\n",
    "    btn.click(fn=generate, inputs=[prompt,negative_prompt,steps,guidance,width,height], outputs=[output])\n",
    "gr.close_all()\n",
    "demo.launch(share=True, server_port=int(os.environ['PORT3']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "655908ab-fa08-4b86-b218-f685f62d3638",
   "metadata": {},
   "outputs": [],
   "source": [
    "with gr.Blocks() as demo:\n",
    "    gr.Markdown(\"# Image Generation with Stable Diffusion\")\n",
    "    with gr.Row():\n",
    "        with gr.Column(scale=4):\n",
    "            prompt = gr.Textbox(label=\"Your prompt\") #Give prompt some real estate\n",
    "        with gr.Column(scale=1, min_width=50):\n",
    "            btn = gr.Button(\"Submit\") #Submit button side by side!\n",
    "    with gr.Accordion(\"Advanced options\", open=False): #Let's hide the advanced options!\n",
    "            negative_prompt = gr.Textbox(label=\"Negative prompt\")\n",
    "            with gr.Row():\n",
    "                with gr.Column():\n",
    "                    steps = gr.Slider(label=\"Inference Steps\", minimum=1, maximum=100, value=25,\n",
    "                      info=\"In many steps will the denoiser denoise the image?\")\n",
    "                    guidance = gr.Slider(label=\"Guidance Scale\", minimum=1, maximum=20, value=7,\n",
    "                      info=\"Controls how much the text prompt influences the result\")\n",
    "                with gr.Column():\n",
    "                    width = gr.Slider(label=\"Width\", minimum=64, maximum=512, step=64, value=512)\n",
    "                    height = gr.Slider(label=\"Height\", minimum=64, maximum=512, step=64, value=512)\n",
    "    output = gr.Image(label=\"Result\") #Move the output up too\n",
    "            \n",
    "    btn.click(fn=generate, inputs=[prompt,negative_prompt,steps,guidance,width,height], outputs=[output])\n",
    "\n",
    "gr.close_all()\n",
    "demo.launch(share=True, server_port=int(os.environ['PORT4']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d91ece7c-f5a5-4025-8558-23bd2e007ed2",
   "metadata": {},
   "outputs": [],
   "source": [
    "gr.close_all()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f30d2cbd-a799-40dc-9c50-a3e26080a16b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3d29fd6c-ab10-44e1-93b9-46404dba15ac",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d73a8109-6904-4d08-9df9-8bacc300ac93",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4a03f8af-25d4-4f50-a2a9-203348440550",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "53ea6093-7353-4e40-8b61-2c1b712f6e4f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3a4483b1-850c-4491-9e2b-b2ddc4d155b8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a28194a7-1d6a-4bfa-a3c8-f1acee01c74c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "361a6a03-227a-4095-9ba1-2d5cd29d3cab",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f0e20904-8f5f-4cc9-a0e6-19a46c38e3ce",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6a24cc12-247a-451a-b274-5100c89b1411",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9444cad0-10e4-4fee-a7a3-61ec2d663a09",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6184fbc4-0c18-499e-b409-ece2ef223736",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8adf5565-fb5d-47f7-919c-0c6790090883",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "45274e82-2919-4f2e-9abc-c2d6b9781b69",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "59e908aa-cab9-4c9f-876f-4eeb3d746f21",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5eec29fb-a992-47c6-9148-dee01802a6b1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c4bd6045-da58-474d-8db4-e7eeb6de36f3",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d4212dc4-f33d-4b55-8ed6-5f4103b2bfa0",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}