utils.py

import os
import torch
from PIL import Image
from diffusers import AutoencoderKL

torch_device = (
    "cuda"
    if torch.cuda.is_available()
    else "mps"
    if torch.backends.mps.is_available()
    else "cpu"
)
if "mps" == torch_device:
    os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"

vae = AutoencoderKL.from_pretrained("CompVis/stable-diffusion-v1-4", subfolder="vae").to(torch_device)

def pil_to_latent(input_im):
    # Single image -> single latent in a batch (so size 1, 4, 64, 64)
    with torch.no_grad():
        latent = vae.encode(tfms.ToTensor()(input_im).unsqueeze(0).to(torch_device)*2-1) # Note scaling
    return 0.18215 * latent.latent_dist.sample()

def latents_to_pil(latents, torch_device=torch_device):
    # bath of latents -> list of images
    latents = (1 / 0.18215) * latents
    with torch.no_grad():
        image = vae.decode(latents.to(torch_device)).sample
    image = (image / 2 + 0.5).clamp(0, 1)
    image = image.detach().cpu().permute(0, 2, 3, 1).numpy()
    images = (image * 255).round().astype("uint8")
    pil_images = [Image.fromarray(image) for image in images]
    return pil_images

def load_embedding_bin(path):
    return torch.load(path)

#  Prep Scheduler
def set_timesteps(scheduler, num_inference_steps):
    scheduler.set_timesteps(num_inference_steps)
    scheduler.timesteps = scheduler.timesteps.to(torch.float32) # minor fix to ensure MPS compatibility, fixed in diffusers PR 3925