reduce rep penality to 1.2

app.py

@@ -288,7 +288,7 @@ with gr.Blocks(title="LLM Inference with ZeroGPU") as demo:
             temp = gr.Slider(0.1, 2.0, value=0.7, step=0.1, label="Temperature")
             k = gr.Slider(1, 100, value=40, step=1, label="Top-K")
             p = gr.Slider(0.1, 1.0, value=0.9, step=0.05, label="Top-P")
-            rp = gr.Slider(1.0, 2.0, value=1.5, step=0.1, label="Repetition Penalty")
+            rp = gr.Slider(1.0, 2.0, value=1.2, step=0.1, label="Repetition Penalty")
             gr.Markdown("### Web Search Settings")
             mr = gr.Number(value=6, precision=0, label="Max Results")
             mc = gr.Number(value=600, precision=0, label="Max Chars/Result")