Saved in:
Bibliographic Details
Main Authors: Liu, Biao, Fang, Wenyi, Wu, Xiaoyu, Zheng, Yang, Hu, Zheng, Yuan, Bo
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2412.15277
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912162615656448
author Liu, Biao
Fang, Wenyi
Wu, Xiaoyu
Zheng, Yang
Hu, Zheng
Yuan, Bo
author_facet Liu, Biao
Fang, Wenyi
Wu, Xiaoyu
Zheng, Yang
Hu, Zheng
Yuan, Bo
contents Pre-trained Vision-Language (VL) models such as CLIP have demonstrated their excellent performance across numerous downstream tasks. A recent method, Context Optimization (CoOp), further improves the performance of VL models on downstream tasks by introducing prompt learning. CoOp optimizes a set of learnable vectors, aka prompt, and freezes the whole CLIP model. However, relying solely on CLIP loss to fine-tune prompts can lead to models that are prone to overfitting on downstream task. To address this issue, we propose a plug-in prompt-regularization method called PLPP (Prompt Learning with PerPlexity), which use perplexity loss to regularize prompt learning. PLPP designs a two-step operation to compute the perplexity for prompts: (a) calculating cosine similarity between the weight of the embedding layer and prompts to get labels, (b) introducing a language model (LM) head that requires no training behind text encoder to output word probability distribution. Meanwhile, we unveil that the essence of PLPP is inherently a form of self-distillation. To further prevent overfitting as well as to reduce the additional computation introduced by PLPP, we turn the hard label to soft label and choose top-$k$ values for calculating the perplexity loss. For accelerating model convergence, we introduce mutual self-distillation learning, that is perplexity and inverted perplexity loss. The experiments conducted on four classification tasks indicate that PLPP exhibits superior performance compared to existing methods.
format Preprint
id arxiv_https___arxiv_org_abs_2412_15277
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle PLPP: Prompt Learning with Perplexity Is Self-Distillation for Vision-Language Models
Liu, Biao
Fang, Wenyi
Wu, Xiaoyu
Zheng, Yang
Hu, Zheng
Yuan, Bo
Computation and Language
Artificial Intelligence
Pre-trained Vision-Language (VL) models such as CLIP have demonstrated their excellent performance across numerous downstream tasks. A recent method, Context Optimization (CoOp), further improves the performance of VL models on downstream tasks by introducing prompt learning. CoOp optimizes a set of learnable vectors, aka prompt, and freezes the whole CLIP model. However, relying solely on CLIP loss to fine-tune prompts can lead to models that are prone to overfitting on downstream task. To address this issue, we propose a plug-in prompt-regularization method called PLPP (Prompt Learning with PerPlexity), which use perplexity loss to regularize prompt learning. PLPP designs a two-step operation to compute the perplexity for prompts: (a) calculating cosine similarity between the weight of the embedding layer and prompts to get labels, (b) introducing a language model (LM) head that requires no training behind text encoder to output word probability distribution. Meanwhile, we unveil that the essence of PLPP is inherently a form of self-distillation. To further prevent overfitting as well as to reduce the additional computation introduced by PLPP, we turn the hard label to soft label and choose top-$k$ values for calculating the perplexity loss. For accelerating model convergence, we introduce mutual self-distillation learning, that is perplexity and inverted perplexity loss. The experiments conducted on four classification tasks indicate that PLPP exhibits superior performance compared to existing methods.
title PLPP: Prompt Learning with Perplexity Is Self-Distillation for Vision-Language Models
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2412.15277