Update app/vectorizer.py

app/vectorizer.py

@@ -9,6 +9,10 @@ from pinecone import Pinecone, ServerlessSpec
 # Disable parallelism for tokenizers
 os.environ['TOKENIZERS_PARALLELISM'] = 'false'
 
+# Ensure the cache directory exists
+os.makedirs('/app/cache/hub', exist_ok=True)
+os.environ['HF_HOME'] = '/app/cache/hub'
+
 # Configure logging
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)