add HF endpoint

agent.py

@@ -163,9 +163,13 @@ def build_graph(provider: str = "huggingface"):
         # llm = ChatGroq(model="qwen-qwq-32b", temperature=0) # optional : qwen-qwq-32b gemma2-9b-it
     if provider == "huggingface":
         # TODO: Add huggingface endpoint
+        HUGGINGFACEHUB_API_TOKEN = os.getenv['HUGGINGFACEHUB_API_TOKEN']
+        repo_id = "llama-2-7b-chat-hf"
         llm = ChatHuggingFace(
             llm=HuggingFaceEndpoint(
-                url="https://api-inference.huggingface.co/models/Meta-DeepLearning/llama-2-7b-chat-hf",
+                # url="https://api-inference.huggingface.co/models/Meta-DeepLearning/llama-2-7b-chat-hf",
+                repo_id=repo_id,
+                huggingfacehub_api_token=HUGGINGFACEHUB_API_TOKEN,
                 temperature=0,
             ),
         )