Instructions to use unsloth/ERNIE-Image-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Local Apps Settings
- Unsloth Studio
How to use unsloth/ERNIE-Image-GGUF with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for unsloth/ERNIE-Image-GGUF to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for unsloth/ERNIE-Image-GGUF to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for unsloth/ERNIE-Image-GGUF to start chatting
Load model with FastModel
pip install unsloth from unsloth import FastModel model, tokenizer = FastModel.from_pretrained( model_name="unsloth/ERNIE-Image-GGUF", max_seq_length=2048, )
base model issue
EDIT: the issue is literally just the model. its not the quant, or settings.... the model just outputs noise, even on full weights. sorry for the confusion.
so i ran the turbo model with GREAT results buit the base model is giving me these results no matter the cfg i input. tried 1.0 - 5.0
something just feels off and theres a lot of noise going on
You're saying there's an issue with the bf16 gguf? or the safetensor?
You're saying there's an issue with the bf16 gguf? or the safetensor?
it was with the any version of the model, not just these
I think it may be the small vae for flux 2, im running a few gens to see if using the full flux 2 vae will fix it. I tried to save decoding process from stalling as i only have 8gb of vram and 16gb of ram so i have to work around that alot. Cant used tiled decode because it ruins images on alot of models so using condensed vaes like the small encoder is a great workaround. However i feel it may be affecting gens with this model. Ill update with images shortly to compare






