Instructions to use Comfy-Org/Lumina_Image_2.0_Repackaged with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusion Single File
How to use Comfy-Org/Lumina_Image_2.0_Repackaged with Diffusion Single File:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Conversion script?
#1
by Nap - opened
I tried to run the original weights (FP32) but there seems to be a mismatch between the tensors on the original files and these repackaged ones, could you point to the script you used to repackage them? The files in this repo are in BF16 and I wanted to test the models with FP32 in ComfyUI.
Thanks
Try loading this file with the "Load Diffusion Model" node: https://huggingface.co/Alpha-VLLM/Lumina-Image-2.0/blob/main/consolidated.00-of-01.pth
All I did is convert that file to .safetensors and the tensors in it to bf16.
Great, thanks!
Nap changed discussion status to closed