Instructions to use a-r-r-o-w/AnyText with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use a-r-r-o-w/AnyText with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("a-r-r-o-w/AnyText", 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
Fix `eps`
As you also know, I am working on this pipeline at the moment. I would like to continue where you left off. Can I take your advice, when I need it?
Probably this is not something important but I just realized this.
@a-r-r-o-w
Thanks for working on this @tolgacangoz ! I just haven't found the time to re-implement the pipeline in diffusers format after I commented on the issue. I probably don't even remember the details anymore but based on their code here, there's a good amount of pre-processing required on the text, which might need to be looked at carefully.
Also, since you're working on this now, it might be good to transfer the repository to your account once you've completed implementation. I will speak with the HF team regarding what can be done, and if download stats can be maintained after transfer. But, you could also run the conversion script yourself if needed. Once again, thank you very much for working on this.