Update app.py

app.py

@@ -6,10 +6,11 @@ import torch
 app = FastAPI()
 
 # Load model and tokenizer once at startup
-model_name = "gpt2"  # change this to your own model
+model_name = "./gpt2"  # ← path to the local directory
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForCausalLM.from_pretrained(model_name)
 
+
 class PromptRequest(BaseModel):
     prompt: str
     max_new_tokens: int = 50