Update app.py

app.py

@@ -1,88 +1,60 @@
 import gradio as gr
-import numpy as np
-import random
-import spaces 
-from diffusers import DiffusionPipeline
 import torch
+from diffusers import StableDiffusionPipeline
+import random
+import numpy as np
+
+# 1. Load a CPU-optimized model
+# 'segmind/tiny-sd' is much smaller and faster on CPUs than SDXL
+model_id = "segmind/tiny-sd"
 
-# 1. Load the model on CPU first (Safe for startup)
-model_repo_id = "stabilityai/sdxl-turbo" 
-pipe = DiffusionPipeline.from_pretrained(
-    model_repo_id, 
-    torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32
+# Use float32 because CPU doesn't support float16 well
+pipe = StableDiffusionPipeline.from_pretrained(
+    model_id, 
+    torch_dtype=torch.float32
 )
+pipe = pipe.to("cpu")
 
-# NOTE: We do NOT move to cuda or enable xformers here. 
-# That would cause the crash you saw.
+# Optimize for CPU speed
+pipe.set_progress_bar_config(disable=True)
 
 MAX_SEED = np.iinfo(np.int32).max
-MAX_IMAGE_SIZE = 1024
 
-@spaces.GPU(duration=60) # Magic starts here
-def infer(
-    prompt,
-    negative_prompt,
-    seed,
-    randomize_seed,
-    width,
-    height,
-    guidance_scale,
-    num_inference_steps,
-    progress=gr.Progress(track_tqdm=True),
-):
+def infer(prompt, seed, randomize_seed, width, height):
     if randomize_seed:
         seed = random.randint(0, MAX_SEED)
     
-    # 2. Move to GPU and enable speed boosts ONLY when the function runs
-    pipe.to("cuda")
-    try:
-        pipe.enable_xformers_memory_efficient_attention()
-    except Exception:
-        pass # Fallback if xformers isn't needed
-        
-    generator = torch.Generator("cuda").manual_seed(seed)
+    generator = torch.Generator("cpu").manual_seed(seed)
     
+    # We use very low steps (10-15) because CPU is slow
     image = pipe(
         prompt=prompt,
-        negative_prompt=negative_prompt,
-        guidance_scale=guidance_scale,
-        num_inference_steps=num_inference_steps,
-        width=width,
-        height=height,
         generator=generator,
+        num_inference_steps=15, 
+        guidance_scale=7.0,
+        width=width,
+        height=height
     ).images[0]
     
     return image, seed
 
-# UI Styling
-css = """#col-container { margin: 0 auto; max-width: 640px; }"""
-
-with gr.Blocks(css=css) as demo:
-    with gr.Column(elem_id="col-container"):
-        gr.Markdown(" # CodeIgnite Image Engine")
+# Simple UI
+with gr.Blocks() as demo:
+    gr.Markdown("# CodeIgnite CPU Image Engine")
+    with gr.Column():
+        prompt = gr.Textbox(label="Prompt", placeholder="A simple cat drawing")
+        run_button = gr.Button("Generate (CPU Mode)")
+        result = gr.Image(label="Result")
         
-        with gr.Row():
-            prompt = gr.Text(label="Prompt", show_label=False, placeholder="Enter your prompt", container=False)
-            run_button = gr.Button("Run", variant="primary")
-        
-        result = gr.Image(label="Result", show_label=False)
-
-        with gr.Accordion("Advanced Settings", open=False):
-            negative_prompt = gr.Text(label="Negative prompt", placeholder="Optional")
+        with gr.Accordion("Settings", open=False):
             seed = gr.Slider(label="Seed", minimum=0, maximum=MAX_SEED, step=1, value=0)
             randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
-            
-            with gr.Row():
-                width = gr.Slider(label="Width", minimum=256, maximum=MAX_IMAGE_SIZE, step=32, value=512)
-                height = gr.Slider(label="Height", minimum=256, maximum=MAX_IMAGE_SIZE, step=32, value=512)
-            
-            with gr.Row():
-                guidance_scale = gr.Slider(label="Guidance scale", minimum=0.0, maximum=10.0, step=0.1, value=0.0)
-                num_inference_steps = gr.Slider(label="Steps", minimum=1, maximum=4, step=1, value=2)
+            width = gr.Slider(label="Width", minimum=256, maximum=512, step=32, value=384)
+            height = gr.Slider(label="Height", minimum=256, maximum=512, step=32, value=384)
 
     run_button.click(
         fn=infer,
-        inputs=[prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps],
+        inputs=[prompt, seed, randomize_seed, width, height],
         outputs=[result, seed],
         api_name="predict"
     )