enable zerogpu with decorator

app.py

@@ -9,6 +9,7 @@ import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer
 from huggingface_hub import hf_hub_download
 from duckduckgo_search import DDGS
+import spaces
 
 # ------------------------------
 # Global Cancellation Event
@@ -107,6 +108,7 @@ def retrieve_context(query, max_results=6, max_chars_per_result=600):
 # ------------------------------
 # Chat Response Generation (Simulated Streaming) with Cancellation
 # ------------------------------
+@spaces.GPU
 def chat_response(user_message, chat_history, system_prompt, enable_search,
                   max_results, max_chars, model_name, max_tokens, temperature, top_k, top_p, repeat_penalty):
     # Reset the cancellation event.

requirements.txt

@@ -1,7 +1,3 @@
---extra-index-url https://www.piwheels.org/simple
---extra-index-url https://github.com/GoogleCloudPlatform/gcloud-python-wheels/raw/refs/heads/master/wheelhouse
---extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu
-
 wheel
 jieba
 docopt
@@ -10,3 +6,4 @@ duckduckgo_search
 gradio
 torch
 transformers
+spaces