Comparison of mT5 Model Performance with English vs. Multilingual Intermediate-Task Fine-Tuning on XTREME-L Tasks

Intermediate-task training---fine-tuning a pretrained model on an intermediate task before fine-tuning again on the target task---often improves model performance substantially on language understanding tasks in monolingual English settings. We investigate whether English intermediate-task training is still helpful on non-English target tasks. Using nine intermediate language-understanding tasks, we evaluate intermediate-task transfer in a zero-shot cross-lingual setting on the XTREME benchmark. We see large improvements from intermediate training on the BUCC and Tatoeba sentence retrieval tas

Research goal: How does the inference latency and accuracy of mT5 models fine-tuned with English intermediate-task training compare to those using multilingual intermediate tasks on XTREME-L tasks, when measured using F1 score and tokens per second?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 9.3/10.