Instructions to use woodfireind/MiniMax-H3-GGUF-MiniStack with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use woodfireind/MiniMax-H3-GGUF-MiniStack with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("woodfireind/MiniMax-H3-GGUF-MiniStack", 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
H3SmallTELoader
Hello. Where I can find H3SmallTELoader?
sorry, i forgot to upload the custom node
it's there now, just make sure to put it in your comfuUI /custom_nodes directory
Thanks. Which version or fork of the ComfyUI-GGUF needed? I have error ValueError: Unexpected architecture type in GGUF file: 'minimax_h3' with ComfyUI-GGUF from main repository.
you need the one with PR #15224 : https://github.com/Comfy-Org/ComfyUI/pull/15224
should be merged into main by now so it should be good as long as you update it
you also need the ComfyUI-GGUF node if you don't have it already
you may also have to update transformers in python
if that doesn't work for whatever reason you can try whitelisting it:
- Open the file throwing the error (e.g., look at your terminal traceback for a path ending in loader.py or modeling_gguf_pytorch_utils.py).
- Locate the supported architecture list variable, which is usually named IMG_ARCH_LIST, ARCH_LIST, or a similar dictionary.
- Manually append "minimax_h3" to that array or list of strings.
- Save the file and restart your application.
Thanks!