Update CODE/vulnscan.py

The condition logic is incorrect. If torch.cuda.is_available() returns False, then torch.backends.cudnn.is_available() will also return False, making this condition always False. Consider simplifying to just check if CUDA is not available but the system has CUDA capabilities.

Co-authored-by: Copilot <[email protected]>
Signed-off-by: Shahm Najeeb <[email protected]>

Author: DefinetlyNotAI

CODE/vulnscan.py

@@ -60,7 +60,7 @@ def _load_model(self) -> None:
                     map_location=torch.device("cuda" if torch.cuda.is_available() else "cpu"),
                     weights_only=False
                 )
-                if not torch.cuda.is_available() and torch.version.cuda and torch.backends.cudnn.is_available():
+                if not torch.cuda.is_available() and torch.version.cuda:
                     log.warning(
                         "NVIDIA GPU detected but CUDA is not available. Check your PyTorch and CUDA installation to utilise as much power as possible.")
                 log.debug(f"Model using device: {torch.device('cuda' if torch.cuda.is_available() else 'cpu')}")