Upload app.py

app.py

@@ -11,9 +11,10 @@ from huggingface_hub import InferenceClient
 # CONFIG
 
 # You can hard-code the model here, or override with MODEL_ID as a Space variable.
-MODEL_ID = os.environ.get("MODEL_ID", "meta-llama/Meta-Llama-3.1-8B-Instruct")
-# Public model → no token required
-HF_TOKEN = os.environ.get("HF_TOKEN")  # optional, can stay empty
+# Use a free, public chat model that doesn't need a novita api_key
+MODEL_ID = os.environ.get("MODEL_ID", "TinyLlama/TinyLlama-1.1B-Chat-v1.0")
+HF_TOKEN = os.environ.get("HF_TOKEN")  # can stay empty
+
 
 REPO_PATH = "/data/questions.json"  # where we store generated questions