Gradio is not auto downloading Hugging Face model weights - auto downloading on Google Colab but not on Windows

Question

The below code automatically downloading necessary model files and weights on Google Colab but not on Windows

Any ideas? How to fix?

import gradio as gr

from diffusers import DiffusionPipeline
import torch

import base64
from io import BytesIO
import os
import gc

from share_btn import community_icon_html, loading_icon_html, share_js

# SDXL code: https://github.com/huggingface/diffusers/pull/3859

model_dir = os.getenv("SDXL_MODEL_DIR")
access_token = os.getenv("my_token")

if model_dir:
    # Use local model
    model_key_base = os.path.join(model_dir, "stable-diffusion-xl-base-0.9")
    model_key_refiner = os.path.join(model_dir, "stable-diffusion-xl-refiner-0.9")
else:
    model_key_base = "nichijoufan777/stable-diffusion-xl-base-0.9"
    model_key_refiner = "nichijoufan777/stable-diffusion-xl-refiner-0.9"

# Use refiner (enabled by default)
enable_refiner = os.getenv("ENABLE_REFINER", "true").lower() == "true"
# Output images before the refiner and after the refiner
output_images_before_refiner = os.getenv("OUTPUT_IMAGES_BEFORE_REFINER", "false").lower() == "true"

# Create public link
share = os.getenv("SHARE", "false").lower() == "true"

print("Loading model", model_key_base)
pipe = DiffusionPipeline.from_pretrained(model_key_base, torch_dtype=torch.float16, use_safetensors=True, variant="fp16", use_auth_token=access_token)

pipe.enable_model_cpu_offload()
# pipe.to("cuda")

# if using torch < 2.0
# pipe.enable_xformers_memory_efficient_attention()

# pipe.unet = torch.compile(pipe.unet, mode="reduce-overhead", fullgraph=True)

if enable_refiner:
    print("Loading model", model_key_refiner)
    pipe_refiner = DiffusionPipeline.from_pretrained(model_key_refiner, torch_dtype=torch.float16, use_safetensors=True, variant="fp16", use_auth_token=access_token)
    pipe_refiner.enable_model_cpu_offload()
    # pipe_refiner.to("cuda")

    # if using torch < 2.0
    # pipe_refiner.enable_xformers_memory_efficient_attention()

    # pipe_refiner.unet = torch.compile(pipe_refiner.unet, mode="reduce-overhead", fullgraph=True)

# NOTE: we do not have word list filtering in this gradio demo

is_gpu_busy = False
.
.
.