Instructions to use stabilityai/stable-diffusion-3.5-medium with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use stabilityai/stable-diffusion-3.5-medium with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-3.5-medium", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps
- Draw Things
- DiffusionBee
Downloading the file textencoders/clip_g.safetensors actually gets the file model.fp16.safetensors
When trying to download the file textencoders/clip_g.safetensors, the downloaded file is actually model.fp16.safetensors. Maybe there is some error/bug with xet, pointing to the wrong file?
Thank you for this report, @SynapStorm ! I work on the Xet team and we believe we've identified the underlying issue and are working on a fix.
For anyone else encountering this, please note that this should only happen when downloading the file from the browser.
The file contents of textencoders/clip_g.safetensors are the exact same as the file contents in text_encoder_2/model.fp16.safetensors and text_encoder_2/model.safetensors (you can validate this by verifying the SHA-256 checksum of these files). Only the filename is different when downloading due to an issue with the Content-Disposition headers returned to the browser.
We'll follow up here when it's fixed!
This should be done! Thanks again @SynapStorm for reporting this (and @John6666 for the ping in https://github.com/huggingface/huggingface_hub/issues/3314#issuecomment-3235504663)
Thank you!