Skip to content

V1.1.3

Choose a tag to compare

View all tags
@umittadelen umittadelen released this 16 Feb 17:42
· 56 commits to main since this release
v1.1.3
0009a2c

EasyUI

V1.1.3

TextToImage is a free, open-source text-to-image generation tool designed for ease of use, allowing anyone to run advanced models on their computer with customizable parameters and progress tracking.

NOTE: The version available via git clone is more up-to-date than the releases because the repository contains the latest changes, which may not yet be included in the official release.

Features

  • Progress Tracking
  • Seed Control
  • Adjustable CFG (Classifier-Free Guidance)
  • Support for multiple schedulers
  • ControlNet support (currently canny/depth is fully supported | normal map is only SD)
  • Image-to-Image and Text-to-Image generation

Installation Guide

Prerequisites

  • A good GPU (it will run slowly if you don't have one)
  • CUDA (to use GPU)
  • VS Code (recommended editor)

Steps

  1. Clone the Repository:

    git clone https://github.com/umittadelen/stableDiffusionEasyUI.git
cd stableDiffusionEasyUI

  2. Change the IP:

    This step is not that important but the starting ip:port is localhost:8080

    It is possible to change this from settings page

  3. Install Dependencies:

    pip install -r requirements.txt

Usage

Running the Application

  1. Start the Flask Server: 
    python app.py
  2. Access the Application:

    Open your web browser and navigate to http://localhost:8080.

Generating Images

  1. Download a Model:

    • Click to Model Editor button and download model using the civitai model id and version id (you can get the ids from AIR: 123456 @ 654321)

  2. Choose a Model:

    • Select a model from the dropdown list or add a new model URL.

  3. Set Parameters:

    • Model Type:
      Choose the type of model (this step is automatically done if the model downloaded using the model editor page) (e.g., SDXL, SD1.5, FLUX).

    • Scheduler:
      Select the scheduler model. (default is [Euler A])

    • Prompt:
      Enter the main prompt that describes the image you want to generate.
      prevent using short prompts (it can make NSFW images)
      (Multiple images can be generated using the multi prompt feature like "prompt1§prompt2" > "prompt1","prompt2")

      [Click for a quick guide on writing better prompts]

    • Negative Prompt:
      Enter negative prompts to specify what you want to avoid in the generated image.

    • Width and Height:
      Set the dimensions of the generated image.

    • CFG Scale:
      Adjust the scale value to control how closely the AI follows your prompts.

    • Sampling Steps:
      Set the number of sampling steps. (higher number means more steps | 30 is recommended)

    • Number of Images:
      Specify the number of images to generate.

    • Custom Seed:
      If set to -1 generates variated images.

  1. Generate Images:
    Click the "Generate Images" button to start the image generation process.

Additional Features

  • Stop Generation: Click the "Stop Generation" button to stop the image generation process.
  • Clear Images: Click the "Clear Images" button to clear all generated images.
  • Preview ControlNet: Click the "Preview ControlNet" button to go to the page for previewing the ControlNet image.
  • Get Metadata: Click the "Get Metadata" button to extract metadata from an uploaded image.
  • Model Editor: Click the "Model Editor" button to view and edit the JSON configuration of models.

License

This project is licensed under the MIT License. See the LICENSE.txt file for details.

More About Me

www.umittadelen.net

Assets 3
Loading