X-PEFT: eXtremely Parameter-Efficient Fine-Tuning for Extreme Multi-Profile Scenarios
This addresses memory scaling issues in multi-profile fine-tuning for language models, though it is incremental as it builds on existing PEFT methods.
The paper tackles the problem of parameter inefficiency in adapter tuning for multiple profiles by introducing X-PEFT, which uses compact tensors as binary masks to select adapters, achieving comparable or better performance while reducing memory per profile by 10,000 times.
Parameter-efficient fine-tuning (PEFT) techniques, such as adapter tuning, aim to fine-tune a pre-trained language model (PLM) using a minimal number of parameters for a specific task or profile. Although adapter tuning provides increased parameter efficiency compared to full-model fine-tuning, it introduces a small set of additional parameters attached to a PLM for each profile. This can become problematic in practical applications with multiple profiles, particularly when a significant increase in the number of profiles linearly boosts the total number of additional parameters. To mitigate this issue, we introduce X-PEFT, a novel PEFT method that leverages a multitude of given adapters by fine-tuning an extremely small set of compact tensors for a new profile, which serve as binary masks to adaptively select the given adapters. To efficiently validate our proposed method, we implement it using a large number of trained or untrained (random) adapters. We evaluate the performance of X-PEFT through LaMP and GLUE tasks and demonstrate that it either matches or surpasses the effectiveness of conventional adapter tuning, despite reducing the memory requirements per profile by a factor of 10,000 compared to it.