run llama in the other space

app.py

@@ -12,7 +12,15 @@ from jinja2 import Environment, FileSystemLoader
 from backend.query_llm import generate_hf, generate_openai, hf_models, openai_models
 from backend.semantic_search import retrieve
 import itertools
+from gradio_client import Client
 
+client = Client("Be-Bo/llama-3-chatbot_70b")
+
+def run_llama(_, msg, *__):
+    return client.predict(
+        message=msg,
+        api_name="/chat"
+    )
 inf_models = list(hf_models.keys()) + list(openai_models)
 
 emb_models = ["bge", "minilm"]
@@ -82,8 +90,9 @@ def bot(history, model_name, oepnai_api_key,
     prompt = template.render(documents=documents, query=query)
     prompt_html = template_html.render(documents=documents, query=query)
 
-
-    if model_name in hf_models:
+    if model_name == "llama 3":
+        generate_fn = run_llama
+    elif model_name in hf_models:
          generate_fn = generate_hf
     elif model_name in openai_models:
          generate_fn = generate_openai