llm client hf updated

translator/llm_client.py

@@ -2,13 +2,21 @@ from huggingface_hub import InferenceClient
 from openai import OpenAI
 import os
 
-HF_MODEL_DEFAULT = "mistralai/Mixtral-8x7B-Instruct"
+HF_MODEL_DEFAULT = "mistralai/Mistral-7B-Instruct-v0.3"
 OPENAI_MODEL_DEFAULT = "gpt-4o-mini"
 
+
 def generate_with_hf(prompt, model=HF_MODEL_DEFAULT):
     client = InferenceClient(model=model, token=os.getenv("HF_TOKEN"))
-    response = client.text_generation(prompt, max_new_tokens=1024)
-    return response
+    response = client.chat_completion(
+        messages=[
+            {"role": "system", "content": "You are a Unity→Unreal translation expert."},
+            {"role": "user", "content": prompt},
+        ],
+        max_tokens=1024,
+        temperature=0.7,
+    )
+    return response.choices[0].message["content"]
 
 def generate_with_openai(prompt, model=OPENAI_MODEL_DEFAULT):
     client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))