Upload app.py

app.py

@@ -90,21 +90,21 @@ import torch.nn.functional as F
 
 
 # ---------------------------------------------------------------------------
-# INTELLIGENT MEMORY MANAGEMENT - Works on ZeroGPU
+# ZEROGPU MEMORY MANAGEMENT
 # ---------------------------------------------------------------------------
 def setup_intelligent_memory_pipeline(pipe, apply_quantization=True):
-    """Intelligent memory management for ZeroGPU"""
+    """Optimized for ZeroGPU - keep everything on GPU"""
     print("Configuring pipeline for ZeroGPU...")
     
-    # Enable VAE optimizations
+    # Enable VAE optimizations only
     print("Enabling VAE slicing and tiling...")
     if hasattr(pipe, 'vae'):
         pipe.vae.enable_slicing()
         pipe.vae.enable_tiling()
     
-    # Use CPU offloading for ZeroGPU
-    print("Using model CPU offloading for ZeroGPU")
-    pipe.enable_model_cpu_offload()
+    # Move to GPU - no CPU offloading for ZeroGPU
+    print("Moving pipeline to GPU...")
+    pipe = pipe.to("cuda")
     return pipe
 
 print("loading base pipeline architecture...")
@@ -404,7 +404,7 @@ body, .gradio-container {
 }
 """
 
-with gr.Blocks(css=css) as demo:
+with gr.Blocks() as demo:
     with gr.Column(elem_id="col-container"):
         with gr.Row(elem_id="preset-row"):
             preset_dropdown = gr.Dropdown(
@@ -1021,4 +1021,4 @@ with gr.Blocks(css=css) as demo:
     )
 
 if __name__ == "__main__":
-    demo.launch(server_name="0.0.0.0", server_port=7860, share=False)
+    demo.launch(server_name="0.0.0.0", server_port=7860, share=False, css=css)