fix : change to false

Fix: Add detailed stacktrace output to predictive error UI for easier debugging

Fix: Move model.to('cuda') globally for ZeroGPU interception

Fix: Remove global CUDA checks and pipeline initialization to avoid ZeroGPU interception failure

Fix: Move spaces import to the top of app.py to resolve CUDA initialization issue

Configure Gradio UI and optimized ZeroGPU quota efficiency

initial commit