Update app.py

app.py

@@ -79,9 +79,9 @@ def generate_response(message, temperature, max_new_tokens, top_p, task):
     print(f"Prompt: {prompt}")
     start_time = time.time()
     inputs = tokenizer.encode(prompt, return_tensors="pt").to(model.device)
-    outputs = model.generate(input_ids=inputs, max_new_tokens=max_new_tokens, use_cache=True, temperature=temperature, min_p=top_p, pad_token_id=tokenizer.eos_token_id)
+    outputs = model.generate(input_ids=inputs, do_sample=True, max_new_tokens=max_new_tokens, use_cache=True, temperature=temperature, min_p=top_p, pad_token_id=tokenizer.eos_token_id)
     response = tokenizer.decode(outputs[0], skip_special_tokens=True)
-    processed_response = response.split("RESPONSE :assistant")[-1].strip()
+    processed_response = response.split("RESPONSE :")[-1].strip()
     end_time = time.time()
     time_taken = end_time - start_time
     print(f"Output: {processed_response}")
@@ -90,7 +90,7 @@ def generate_response(message, temperature, max_new_tokens, top_p, task):
     current_time_pst = datetime.now(pst_timezone).strftime("%Y-%m-%d %H:%M:%S %Z%z")
     print(f"Current timestamp (PST): {current_time_pst}")
     return processed_response
-with gr.Blocks() as demo:
+with gr.Blocks(theme='1024m/1024m-1') as demo:
     gr.Markdown("# Phi-4-Hindi Demo")
     with gr.Row():
         with gr.Column():