Which of the following statements correctly describe the advantages of the Transformer architecture? Select all that apply.

Transformers allow many words to be processed at once, so training can be much faster than with older recurrent models. Their self‑attention mechanism lets the model look at every other word in a sentence, making it easy to capture long‑distance relationships. Because the operations are mostly matrix multiplications, the architecture fits well on modern GPUs and can be scaled to huge data sets. These features give Transformers better speed, better handling of long‑range context, and easier parallel computing. For example, a transformer can read an entire paragraph in one pass, while a traditional RNN would read it word by word.