Equalized

app.py

@@ -22,8 +22,7 @@ model = AutoModelForCausalLM.from_pretrained("TinyLlama/TinyLlama-1.1B-Chat-v1.0
 # Using CUDA for an optimal experience
 
 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-model = model.to(device).eval()
-
+model = model.to(device)
 
 # Defining a custom stopping criteria class for the model's text generation
 
@@ -52,11 +51,11 @@ def generate_response(user_input, history):
     generate_kwargs = dict(
         **model_inputs,
         streamer=streamer,
-        max_new_tokens=256,
-        do_sample=False,
-        # top_p=0.95,
-        # top_k=50,
-        # temperature=0.7,
+        max_new_tokens=1024,
+        do_sample=True,
+        top_p=0.95,
+        top_k=50,
+        temperature=0.7,
         num_beams=1,
         stopping_criteria=StoppingCriteriaList([stop])
     )
@@ -389,5 +388,4 @@ with gr.Blocks(css=css, fill_width=True, title="LogicLinkV5") as demo:
         queue=False
     )
 
-demo.queue().launch(share=True, debug=True)
-model.eval()
+demo.queue().launch(share=True, debug=True)