Instructions to use prism-ml/bonsai-image-ternary-4B-gemlite-2bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use prism-ml/bonsai-image-ternary-4B-gemlite-2bit with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("prism-ml/bonsai-image-ternary-4B-gemlite-2bit", torch_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
Bonsai Image Ternary: Clarification
You are hereby greeted.
I congratulate you on your release of this model. I am impressed by the speed and relative quality I experienced on your Hugging Face space. The only point of improvement I could think of is human anatomy, though the used base model, FLUX.2 [klein] 4B, is infamous for being exceptionally horrifying in that field.
Using this model raised one main question which the model card did not answer. Does this repository "only" host a quantized variant of the base model, or did you retrain it to be used with this level of quantization? I would be happy to have the answer to this question.
You are hereby bidden farewell.
You are hereby greeted.
Since I did not receive a response in over a week, I am closing this discussion.
You are hereby bidden farewell.