Update app.py

app.py

@@ -23,15 +23,9 @@ model = AutoModelForCausalLM.from_pretrained( model_path, device_map='cuda', loa
 
 @spaces.GPU
 def generate_text(prompt, custom_prompt, temperature, max_length):
-    # Example message format
-    msg = [{
-        'title': prompt, 
-        'content': custom_prompt 
-    }]
-    
-    prompt_text = f"title: {msg[0]['title']}\ncontent: {msg[0]['content']}"
-    inputs = tokenizer(prompt_text, return_tensors='pt').to('cuda')
-    generated_ids = model.generate(**inputs, max_new_tokens=max_length, temperature=temperature, pad_token_id=tokenizer.eos_token_id)
+    structured_input = f"[Title]: {prompt}\n[Content]: {custom_prompt}"
+    inputs = tokenizer(structured_input, return_tensors='pt').to('cuda')    
+    generated_ids = model.generate(**inputs, max_new_tokens=max_length, temperature=temperature, do_sample=True, pad_token_id=tokenizer.eos_token_id)
     generated_text = tokenizer.decode(generated_ids[0], skip_special_tokens=True)
     
     return generated_text