Consider using MLflow to help manage model saving and loading. https://mlflow.org/docs/latest/deep-learning/pytorch/guide

Saving the model

import mlflow.pytorch

with mlflow.start_run():
    mlflow.pytorch.log_model(model, "model")

Loading the model later:

import mlflow.pytorch

model = mlflow.pytorch.load_model("runs:/<run_id>/model")