Where are models installed by default ?

[Infinitusvoid]

When you run this code where is the model located how to find out the path to the model itself ?
``
from diffusers import StableDiffusionPipeline
import torch

model_id = "stabilityai/stable-diffusion-2-1"
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
``

[a-r-r-o-w]

Hi @Infinitusvoid.

diffusers uses huggingface/huggingface_hub to download and store models that you use. This provides you with a list of environment variables that hub uses by default.

HF_HUB_CACHE is what you're looking for, and by default it is ~/.cache/huggingface/hub, where ~ means your home directory for your profile on windows/linux/mac.

If you'd like to change the location where your models are downloaded by default, you could do this:

# If you want the variable to persist everytime you run a shell
echo "export HF_HOME=/home/profile/path/to/model/directory" >> ~/.bashrc # or config file for whatever terminal you use

# You could also create a symbolic link from your directory to HF_HUB_CACHE
ln -s /path/to/your/model/directory $HF_HUB_CACHE

You could also achieve the same by using the cache_dir parameter when downloading models:

model_id = "stabilityai/stable-diffusion-2-1"
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16, cache_dir="/path/to/model/directory")

[Infinitusvoid]

Thanks, found 70 Gb of models saved there xD

[runner22k]

does diffuser has a config file? i am on windows 11 and i want to change the default models location to a AI dedicated drive. Please help.