Update app.py

app.py

@@ -19,33 +19,39 @@ def load_model():
         repo_id=MODEL_REPO,
         filename=MODEL_FILE,
         cache_dir=CACHE_DIR,
-        force_download=False  # Set to True to bypass cache
+        force_download=False
     )
     
     return Llama(
-        model_path=model_path,  # Now a valid path string
-        n_ctx=2048,             # Context window size
-        n_threads=4,            # CPU threads for faster inference
-        verbose=False           # Disable debug logs
+        model_path=model_path,
+        n_ctx=2048,
+        n_threads=4,
+        verbose=False
     )
 
 # Load model at startup
 llm = load_model()
 
-# Generation function with parameters
+# Generation function with anti-repetition
 def generate_text(prompt, max_tokens=MAX_TOKENS, temp=0.7, top_p=0.95):
-    """Generate text using GGUF model with parameter control"""
+    """Generate text with repetition prevention and error handling"""
+    if not prompt.strip():
+        return "Please enter a valid prompt."
+        
     try:
         output = llm(
             prompt=prompt,
             max_tokens=max_tokens,
             temperature=temp,
             top_p=top_p,
-            echo=False  # Don't repeat input in output
+            echo=False,
+            # Anti-repetition parameters
+            repeat_penalty=1.2,
+            no_repeat_ngram_size=3
         )
-        return output["choices"][0]["text"]
+        return output["choices"][0]["text"].strip()
     except Exception as e:
-        return f"Error generating text: {str(e)}"
+        return f"⚠️ Error generating text: {str(e)}"
 
 # UI Components
 with gr.Blocks(theme="soft") as demo:
@@ -60,7 +66,8 @@ with gr.Blocks(theme="soft") as demo:
             prompt = gr.Textbox(
                 label="Input Prompt", 
                 placeholder="Enter your prompt here...",
-                lines=5
+                lines=5,
+                min_length=10
             )
             max_tokens = gr.Slider(
                 minimum=50, 
@@ -72,14 +79,14 @@ with gr.Blocks(theme="soft") as demo:
             temp = gr.Slider(
                 minimum=0.1, 
                 maximum=1.0, 
-                value=0.7, 
+                value=0.85, 
                 step=0.1,
                 label="Creativity (Temperature)"
             )
             top_p = gr.Slider(
                 minimum=0.1, 
                 maximum=1.0, 
-                value=0.95, 
+                value=0.9, 
                 step=0.05,
                 label="Top-p Sampling"
             )