hard20 pts · first answer countssecuritymodel-registry

Loading a model from a public hub

A teammate wants to load a pickled PyTorch checkpoint downloaded from an unknown public repository in a production pipeline. What is the main risk and safer approach?

A.Pickle files are slower to load, so convert to CSVB.There is no risk because model files only contain numbersC.The model might be too accurateD.Unpickling can execute arbitrary code; prefer weight-only formats such as safetensors, verify source and checksums or signatures, and scan artifacts before use

← Model affects its own training data Where to use tensor parallelism →