GPU Memory Footprint in Intermediate-Task vs. Target-Language Fine-Tuning for Zero-Shot Transfer Performance on XTREME-R PAWS-X

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 GPU memory footprint during intermediate-task fine-tuning on English NLI datasets compare to target-language fine-tuning when evaluating zero-shot transfer performance on the XTREME-R PAWS-X subset?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 9.1/10.