Many ML models experience offline training, and some online training is fine-tuning a model that was pretrained offline. Hence, I see no particular reason to wo

Many ML models experience offline training, and some online training is fine-tuning a model that was pretrained offline. Hence, I see no particular reason to worry about doing this "in prod."