{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Improve image quality with deterministic generation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A common way to improve the quality of generated images is with *deterministic batch generation*, generate a batch of images and select one image to improve with a more detailed prompt in a second round of inference. The key is to pass a list of [`torch.Generator`](https://pytorch.org/docs/stable/generated/torch.Generator.html#generator)'s to the pipeline for batched image generation, and tie each `Generator` to a seed so you can reuse it for an image.\n",
    "\n",
    "Let's use [`runwayml/stable-diffusion-v1-5`](https://huggingface.co/docs/diffusers/main/en/using-diffusers/runwayml/stable-diffusion-v1-5) for example, and generate several versions of the following prompt:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "prompt = \"Labrador in the style of Vermeer\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Instantiate a pipeline with [DiffusionPipeline.from_pretrained()](https://huggingface.co/docs/diffusers/main/en/api/pipelines/overview#diffusers.DiffusionPipeline.from_pretrained) and place it on a GPU (if available):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from diffusers import DiffusionPipeline\n",
    "\n",
    "pipe = DiffusionPipeline.from_pretrained(\"runwayml/stable-diffusion-v1-5\", torch_dtype=torch.float16)\n",
    "pipe = pipe.to(\"cuda\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, define four different `Generator`'s and assign each `Generator` a seed (`0` to `3`) so you can reuse a `Generator` later for a specific image:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "\n",
    "generator = [torch.Generator(device=\"cuda\").manual_seed(i) for i in range(4)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Generate the images and have a look:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "images = pipe(prompt, generator=generator, num_images_per_prompt=4).images\n",
    "images"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![img](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/reusabe_seeds.jpg)\n",
    "\n",
    "In this example, you'll improve upon the first image - but in reality, you can use any image you want (even the image with double sets of eyes!). The first image used the `Generator` with seed `0`, so you'll reuse that `Generator` for the second round of inference. To improve the quality of the image, add some additional text to the prompt:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "prompt = [prompt + t for t in [\", highly realistic\", \", artsy\", \", trending\", \", colorful\"]]\n",
    "generator = [torch.Generator(device=\"cuda\").manual_seed(0) for i in range(4)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create four generators with seed `0`, and generate another batch of images, all of which should look like the first image from the previous round!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "images = pipe(prompt, generator=generator).images\n",
    "images"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![img](https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/reusabe_seeds_2.jpg)"
   ]
  }
 ],
 "metadata": {},
 "nbformat": 4,
 "nbformat_minor": 4
}