OpenJourney : Midjourney, but open source!

OpenJourney is a text-to-image AI model based on prompthero/midjourney-v4-diffusion and is currently under development by Muhammadreza Haghiri. The main goal of this model is to bring a simple yet effective way of bringing Midjourney style images to the FLOSS world.

Sample Image

How to use with diffusers library

Installing necessary libraries

NOTE: This model currently works on a computer which has at least one NVIDIA GPU with CUDA support.

pip install diffusers transformers ftfy scipy accelerate

Logging in

For logging in, you have to use huggingface-cli login command.

Importing necessary libraries

import torch
from torch import autocast
from diffusers.models import AutoencoderKL
from diffusers import StableDiffusionPipeline

Creating the pipeline

pipe = StableDiffusionPipeline.from_pretrained("openjourney/openjourney", use_auth_token=True)
pipe = pipe.to("cuda")

(Optional) Disabling NSFW Filter

NOTE: Remember disabling this is not recommended, but since people had problems with some very basic prompts, we offer this. Remember AI art has a vast majority of users, so keep underage and sensitive users safe.

def dummy(images, **kwargs): 
	return images, False
			
pipe.safety_checker = dummy

Image Generation

prompt = "painting of a building in a stunning landscape"
prompt = f"mdjrny-v4 style {prompt}"

with autocast("cuda"):
  image = pipe(prompt=prompt, num_inference_steps=100, width=512, height=512, guidance_scale=15).images[0]
  
image.save("image.png")

Links

website