runtime

app.py

@@ -4,6 +4,7 @@ from model import *
 DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
 
 model = GPTLanguageModel().to(DEVICE)
+
 model.load_state_dict(torch.load("mini-gpt.pth",map_location=DEVICE), strict=False)
 model.eval()
 answer = decode(model.generate(context, max_new_tokens=3000)[0].tolist())