Create app.py

app.py

@@ -0,0 +1,61 @@
+import gradio as gr
+import torch
+import spaces
+from diffusers import FluxPipeline
+
+# 1. Initialize the Pipeline
+model_id = "black-forest-labs/FLUX.1-schnell"
+
+# Load the model with bfloat16 to save memory and improve speed
+pipe = FluxPipeline.from_pretrained(
+    model_id, 
+    torch_dtype=torch.bfloat16
+)
+
+# 2. Define the Generation Function with ZeroGPU decorator
+# The @spaces.GPU decorator handles the dynamic GPU allocation on Hugging Face
+@spaces.GPU(duration=60)
+def generate_image(prompt, seed, width, height, steps):
+    pipe.to("cuda") # Moves model to GPU only during execution
+    
+    generator = torch.Generator("cuda").manual_seed(seed)
+    
+    image = pipe(
+        prompt=prompt,
+        width=width,
+        height=height,
+        num_inference_steps=steps, 
+        generator=generator,
+        guidance_scale=0.0 # Schnell version works best with 0 guidance
+    ).images[0]
+    
+    return image
+
+# 3. Create the Gradio Interface
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown("# 🎨 FLUX.1-schnell Image Generator")
+    gr.Markdown("Generating high-quality images on Hugging Face ZeroGPU.")
+
+    with gr.Row():
+        with gr.Column():
+            prompt = gr.Textbox(label="Enter your prompt", placeholder="A neon-lit cyberpunk cat...")
+            generate_btn = gr.Button("Generate", variant="primary")
+            
+            with gr.Accordion("Settings", open=False):
+                seed = gr.Slider(0, 1000000, label="Seed", value=42, step=1)
+                width = gr.Slider(512, 1024, label="Width", value=1024, step=64)
+                height = gr.Slider(512, 1024, label="Height", value=1024, step=64)
+                steps = gr.Slider(1, 4, label="Inference Steps", value=4, step=1)
+        
+        with gr.Column():
+            output_image = gr.Image(label="Generated Image")
+
+    generate_btn.click(
+        fn=generate_image, 
+        inputs=[prompt, seed, width, height, steps], 
+        outputs=output_image
+    )
+
+# 4. Launch the App
+if __name__ == "__main__":
+    demo.launch()