Update app.py

app.py

@@ -6,30 +6,29 @@ import json
 import gradio as gr
 import numpy as np
 from PIL import Image
-import spaces
 import torch
 from diffusers import StableDiffusionXLPipeline, EulerAncestralDiscreteScheduler
 
-if not torch.cuda.is_available():
-    DESCRIPTION += "\n<p>Running on CPU 🥶 This demo may not work on CPU.</p>"
+DESCRIPTION = "A Stable Diffusion XL demo running on CPU."
 
 MAX_SEED = np.iinfo(np.int32).max
-CACHE_EXAMPLES = torch.cuda.is_available() and os.getenv("CACHE_EXAMPLES", "1") == "1"
+CACHE_EXAMPLES = os.getenv("CACHE_EXAMPLES", "1") == "1"
 MAX_IMAGE_SIZE = int(os.getenv("MAX_IMAGE_SIZE", "4096"))
 USE_TORCH_COMPILE = os.getenv("USE_TORCH_COMPILE", "0") == "1"
 ENABLE_CPU_OFFLOAD = os.getenv("ENABLE_CPU_OFFLOAD", "0") == "1"
 
-device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
+# Set device to CPU explicitly
+device = torch.device("cpu")
 
-if torch.cuda.is_available():
-    pipe = StableDiffusionXLPipeline.from_pretrained(
-        "sd-community/sdxl-flash",
-        torch_dtype=torch.float16,
-        use_safetensors=True,
-        add_watermarker=False
-    )
-    pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config)
-    pipe.to("cuda")
+# Load pipeline and scheduler for CPU
+pipe = StableDiffusionXLPipeline.from_pretrained(
+    "sd-community/sdxl-flash",
+    torch_dtype=torch.float32,  # Use float32 for CPU
+    use_safetensors=True,
+    add_watermarker=False
+)
+pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config)
+pipe.to(device)  # Move the model to CPU
 
 def save_image(img):
     unique_name = str(uuid.uuid4()) + ".png"
@@ -41,7 +40,6 @@ def randomize_seed_fn(seed: int, randomize_seed: bool) -> int:
         seed = random.randint(0, MAX_SEED)
     return seed
 
-@spaces.GPU(duration=30, queue=False)
 def generate(
     prompt: str,
     negative_prompt: str = "",
@@ -55,23 +53,22 @@ def generate(
     use_resolution_binning: bool = True,
     progress=gr.Progress(track_tqdm=True),
 ):
-    pipe.to(device)
     seed = int(randomize_seed_fn(seed, randomize_seed))
-    generator = torch.Generator().manual_seed(seed)   
+    generator = torch.Generator(device=device).manual_seed(seed)
 
     options = {
-        "prompt":prompt,
-        "negative_prompt":negative_prompt,
-        "width":width,
-        "height":height,
-        "guidance_scale":guidance_scale,
-        "num_inference_steps":num_inference_steps,
-        "generator":generator,
-        "use_resolution_binning":use_resolution_binning,
-        "output_type":"pil",
-
+        "prompt": prompt,
+        "negative_prompt": negative_prompt if use_negative_prompt else None,
+        "width": width,
+        "height": height,
+        "guidance_scale": guidance_scale,
+        "num_inference_steps": num_inference_steps,
+        "generator": generator,
+        "use_resolution_binning": use_resolution_binning,
+        "output_type": "pil",
     }
-    
+
+    # Generate images
     images = pipe(**options).images
 
     image_paths = [save_image(img) for img in images]
@@ -79,17 +76,15 @@ def generate(
 
 
 css = '''
-.gradio-container{max-width: 700px !important}
-h1{text-align:center}
-footer {
-    visibility: hidden
-}
+.gradio-container { max-width: 700px !important; }
+h1 { text-align: center; }
+footer { visibility: hidden; }
 '''
+
 with gr.Blocks(css=css) as demo:
-  
     gr.Markdown("""
 <div style="text-align: center; font-weight: bold; font-size: 2em;">
-    Womener AI
+    Womener AI (CPU Mode)
 </div>
 """)
 
@@ -154,7 +149,6 @@ with gr.Blocks(css=css) as demo:
                 value=8,
             )
 
-
     use_negative_prompt.change(
         fn=lambda x: gr.update(visible=x),
         inputs=use_negative_prompt,
@@ -185,4 +179,4 @@ with gr.Blocks(css=css) as demo:
     )
 
 if __name__ == "__main__":
-    demo.queue(max_size=50).launch()
+    demo.queue(max_size=50).launch()