Scaling Language Model Size in Cross-Device Federated Learning

Most studies in cross-device federated learning focus on small models, due to the server-client communication and on-device computation bottlenecks. In this work, we leverage various techniques for mitigating these bottlenecks to train larger language models in cross-device federated learning. With systematic applications of partial model training, quantization, efficient transfer learning, and communication-efficient optimizers, we are able to train a 21M parameter Transformer and 20. 2M parameter Conformer that achieve the same or better perplexity as that of a similarly sized LSTM with 10 smaller client-to-server communication cost and 11\% lower perplexity than smaller LSTMs commonly studied in literature.