How do CNNs and Transformers differ in model architecture and training for computer vision at Berkshire Grey?
