Update app.py

app.py

@@ -23,7 +23,7 @@ provider = TransformersProvider(model2, tokenizer, device)
 strategy = CreativeWritingStrategy(provider, 
                                    top_p_flat = 0.65,
                                    top_k_threshold_flat = 9,
-                                   eos_penalty = 0.8)
+                                   eos_penalty = 0.75)
 creative_sampler = BacktrackSampler(provider, strategy)
 
 def create_chat_template_messages(history, prompt):
@@ -44,13 +44,13 @@ def generate_responses(prompt, history):
 
     async def custom_sampler_task():
         generated_list = []
-        generator = creative_sampler.generate(wrapped_prompt, max_length=1024, temperature=1)
+        generator = creative_sampler.generate(wrapped_prompt, max_new_tokens=1024, temperature=1)
         for token in generator:
             generated_list.append(token)
         return tokenizer.decode(generated_list, skip_special_tokens=True)
         
     custom_output = asyncio.run(custom_sampler_task())
-    standard_output = model1.generate(inputs, max_length=1024, temperature=1)
+    standard_output = model1.generate(inputs, max_new_tokens=1024, temperature=1)
     standard_response = tokenizer.decode(standard_output[0][len(inputs[0]):], skip_special_tokens=True)
 
     return standard_response.strip(), custom_output.strip()