Update src/insight.py

src/insight.py

@@ -10,8 +10,10 @@ model_id = "google/gemma-3n-E4B-it"
 
 
 hf_token = os.environ.get("HUGGINGFACE_TOKEN")
-tokenizer = AutoTokenizer.from_pretrained(model_id, token=hf_token)
-model = AutoModelForCausalLM.from_pretrained(model_id, token=hf_token)
+cache_dir = "/tmp/hf_cache"
+tokenizer = AutoTokenizer.from_pretrained(model_id, token=hf_token, cache_dir=cache_dir)
+model = AutoModelForCausalLM.from_pretrained(model_id, token=hf_token, cache_dir=cache_dir)
+
 
 def call_llm(prompt):
     inputs = tokenizer(prompt, return_tensors="pt", truncation=True)