You are developing an image recognition model using PyTorch based on the ResNet50 architecture. Your code is running successfully on your local laptop using a small subsample of the data. However, your full dataset contains 200,000 labeled images, and you need to efficiently scale your training workload while minimizing costs. Considering that you plan to use 4 V100 GPUs for this task, which of the following approaches would be the most effective? | Google Professional Machine Learning Engineer Quiz