Redirect download directory

[Jeffery-Ni]

when running the example:
from stable_diffusion_videos import StableDiffusionWalkPipeline
import torch

pipeline = StableDiffusionWalkPipeline.from_pretrained(
"CompVis/stable-diffusion-v1-4",
torch_dtype=torch.float16,
).to("cuda")

video_path = pipeline.walk(
prompts=['a cat', 'a dog'],
seeds=[42, 1337],
num_interpolation_steps=3,
height=512, # use multiples of 64 if > 512. Multiples of 8 if < 512.
width=512, # use multiples of 64 if > 512. Multiples of 8 if < 512.
output_dir='dreams', # Where images/videos will be saved
name='animals_test', # Subdirectory of output_dir where images/videos will be saved
guidance_scale=8.5, # Higher adheres to prompt more, lower lets model take the wheel
num_inference_steps=50, # Number of diffusion steps per image generated. 50 is good default
)
I want to redirect the downloading of the pre-trained data to a directory of designated directory of my wish, and possibly use that directory for the later generation, how can i accomplish that?

[yshashanky]

Which directory is it using now for downloading?

[AlanZeng423]

Like this

from pathlib import Path

# Define the directory where you want to store the model checkpoints and data
download_location = Path("/path/to/your/download/directory")

# Use the `cache_dir` parameter when loading the pre-trained models or datasets
pipe = StableDiffusionWalkPipeline.from_pretrained(
    'runwayml/stable-diffusion-v1-5',
    torch_dtype=torch.float16,
    safety_checker=None,
    vae=AutoencoderKL.from_pretrained("stabilityai/sd-vae-ft-mse", torch_dtype=torch.float16).to("cuda"),
    scheduler=LMSDiscreteScheduler(
        beta_start=0.00085, beta_end=0.012, beta_schedule="scaled_linear"
    ),
    cache_dir=download_location  # Specify the download location here
).to("cuda")

# ...

if __name__ == '__main__':
    interface.launch(debug=True)