Update app.py

app.py

@@ -6,7 +6,9 @@ import logging
 from typing import Tuple, Literal
 import functools
 import spaces
-from cashews import cache
+from cachetools import TTLCache
+from cachetools.func import ttl_cache
+import time
 # Set up logging
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
@@ -16,8 +18,8 @@ MODEL_NAME = "davanstrien/Smol-Hub-tldr"
 model = None
 tokenizer = None
 device = None
-
-cache.setup("mem://", size_limit="4gb")
+CACHE_TTL = 6 * 60 * 60  # 6 hours in seconds
+CACHE_MAXSIZE = 100
 
 def load_model():
     global model, tokenizer, device
@@ -87,9 +89,9 @@ def _generate_summary_gpu(card_text: str, card_type: str) -> str:
 
     return summary
 
-@cache(ttl="6h")
+@ttl_cache(maxsize=CACHE_MAXSIZE, ttl=CACHE_TTL)
 def generate_summary(card_text: str, card_type: str) -> str:
-    """Cached wrapper for generate_summary."""
+    """Cached wrapper for generate_summary with TTL."""
     return _generate_summary_gpu(card_text, card_type)
 
 def summarize(hub_id: str = "", card_type: str = "model", content: str = "") -> str: