Requirements to Get Started

To embark on this journey, you’ll only need a fundamental understanding of Python. The instructions provided here are tailored for Windows, but with minor tweaks, they can be adapted for Linux or Mac systems.

Installation Process

Creating Your Image Generation Script

Let’s put together a simple script that will generate images based on prompts. Start by creating a new project in PyCharm. Open the application and initiate a new project:

You can name your project 'OpenJourney'. Ensure the option to create a ‘main.py’ welcome script is selected, then replace its content. Once your project is set up, you should see the following files:

Now, let’s update the ‘main.py’ file with this code:

from diffusers import StableDiffusionPipeline

# Load the model

model_id = "prompthero/openjourney"

pipe = StableDiffusionPipeline.from_pretrained(model_id)

pipe = pipe.to("cpu")

# Prompt

prompt = input("Enter a prompt: ")

prompt += ' mdjrny-v4 style'

# Generate image

image = pipe(prompt).images[0]

image.save("image.png")

This code allows us to generate an image from the provided prompt. Here’s a brief breakdown: we load the model using the StableDiffusion pipeline from Hugging Face, specifying to use the CPU. We then ask for the prompt and add ‘mdjrny-v4 style’ to it, which is essential for producing images in the Midjourney style. Finally, the image is generated and saved to a file.

Before running the script, we’ll need to install a couple of external libraries via PIP:

pip install diffusers transformers accelerate

You can do this in the PyCharm command line:

Additionally, we need to install PyTorch, a crucial library for AI applications in Python. This can also be done with PIP:

pip install torch torchvision torchaudio

This will set up PyTorch with CPU support, which is what we’ll be using in our script, though CUDA support is available for those with NVIDIA GPUs.

Running Your Script

Now it’s time to run your script and create your first image! Click the play button in PyCharm:

Your code will start downloading the necessary model files, and then it will prompt you for input:

After entering your prompt, like ‘muscle car’, the image generation will begin:

Within 10 to 15 minutes, depending on your CPU, you'll receive a beautifully generated image:

Congratulations! You've successfully created a Midjourney-style image locally on your computer. While it may not have been instantaneous, you now have the ability to generate any AI images you desire.