Instructions to use prompthero/openjourney-v4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use prompthero/openjourney-v4 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("prompthero/openjourney-v4", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
You asked on model card for tests so i did some
#9
by sosojni - opened
I created grids of 9 images with same seed where aplicable + compared to v2.1 and 2.1+some embeddings
Prompt: cute baby groot grinch, dramatic lighting,by robert longo,realistic,high detail,hyper realistic,ultry high detail, 8k,photo realistic,dark background
DPM++ 2M
20 Steps
CFG 7
v2.1 models were rendered on their native 768x768 resolution. One is bare naked 2.1 model, another one has 3 embeddings all of which i found on SD reddit and downloaded some time ago so i do not have links. Embeddings are :CinemaHelper, PhotoHelper, midjourney
Make your own decision and pick what you like.
Thank you very much!!
Really good work. We've chosen epoch3.
prompthero changed discussion status to closed







