update

app.py

@@ -108,6 +108,12 @@ def generate_response(input_data, chat_history, max_new_tokens, temperature, top
     # Add current user message
     messages.append({"role": "user", "content": input_data})
 
+    # Validation للقيم عشان ما تحصل CUDA errors
+    temperature = max(0.1, min(2.0, temperature))  # 0.1 to 2.0
+    top_p = max(0.1, min(1.0, top_p))  # 0.1 to 1.0
+    top_k = max(1, min(100, top_k))  # 1 to 100
+    repetition_penalty = max(1.0, min(1.5, repetition_penalty))  # 1.0 to 1.5
+
     # استخدام ChatPipeline المخصص مع streaming
     streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
 
@@ -134,10 +140,10 @@ demo = gr.ChatInterface(
     fn=generate_response,
     additional_inputs=[
         gr.Slider(label="الحد الأقصى للكلمات الجديدة", minimum=64, maximum=4096, step=1, value=2048),
-        gr.Slider(label="درجة الحرارة", minimum=0.1, maximum=2.0, step=0.1, value=0.7),
-        gr.Slider(label="Top-p", minimum=0.05, maximum=1.0, step=0.05, value=0.9),
+        gr.Slider(label="درجة الحرارة", minimum=0.1, maximum=2.0, step=0.1, value=1.0),
+        gr.Slider(label="Top-p", minimum=0.1, maximum=1.0, step=0.05, value=0.9),
         gr.Slider(label="Top-k", minimum=1, maximum=100, step=1, value=50),
-        gr.Slider(label="عقوبة التكرار", minimum=1.0, maximum=2.0, step=0.05, value=1.0)
+        gr.Slider(label="عقوبة التكرار", minimum=1.0, maximum=1.5, step=0.05, value=1.2)
     ],
     examples=[
         [{"text": "النت عندي معطل من الصبح، تقدر تساعدني؟"}],