Instructions to use AlperKTS/Krea2_FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use AlperKTS/Krea2_FP8 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("AlperKTS/Krea2_FP8", 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
Error indicating something wrong with model
#6
by FuzzyMeeper - opened
I'm pretty sure the model I got was from this repo specifically, but yeah I have the "type" set to krea2 and when I use this model I often get the following error
ValueError: Krea2 expects conditioning with 12x2560=30720 features (a 12-layer Qwen3-VL stack) but got 4096. Load the text encoder with CLIPLoader type 'krea2'.
Is this a "prompt enhancement only" model?
I tried it with the krea edit workflow provided by the krea edit custom node pack using krea2_turbo_fp8_scaled.safetensors. Pretty sure it does it with other krea2 turbo models also.