Paper ID: 2312.00700

Generative Parameter-Efficient Fine-Tuning

Chinmay Savadikar, Xi Song, Tianfu Wu

We present Generative Parameter-Efficient Fine-Tuning (GIFT) for adapting pretrained Transformer backbones on downstream tasks. GIFT learns to generate the fine-tuned weights for a layer directly from its pretrained weights. The GIFT network is parameterized in a minimally-simple way by two linear layers (without bias terms), and is shared by different pretrained layers selected for fine-tuning (e.g., the Query layers), which result in significantly fewer trainable parameters compared to the layer-specific methods like Low-Rank Adapter (LoRA). We also show this formulation bridges parameter-efficient fine-tuning and representation fine-tuning. We perform comprehensive experiments on natural language tasks (commonsense and arithmetic reasoning, instruction tuning, and sequence classification) and computer vision tasks (fine-grained classification). We obtain the best performance and parameter efficiency among baselines on commonsense and arithmetic reasoning, and instruction following using the Llama family of models and on visual recognition benchmarks using Vision Transformers. Notably, compared to LoRA, we obtain 5.7% absolute increase in average accuracy with 14 times reduction of parameters on Commonsense170k using Llama-3 (8B), and 5.4% absolute increase in the win rate with 4 times reduction of parameters using Llama-2 (7B) during instruction tuning. Our GIFT also obtains a slightly higher win rate on instruction tuning than GPT 3.5 (Turbo 1106).

Submitted: Dec 1, 2023