MaPLe: Multi-modal Prompt Learning

TL;DR

MaPLe introduces a multi-modal prompt learning approach that jointly optimizes vision and language prompts to enhance alignment and generalization of CLIP-like models across diverse image recognition tasks.

Contribution

This work proposes a novel multi-modal prompt learning framework that improves vision-language alignment by coupling prompts and modeling stage-wise features, outperforming prior single-branch prompt methods.

Findings

Achieves 3.45% improvement on novel classes over state-of-the-art.

Improves overall harmonic mean by 2.72% across datasets.

Demonstrates effectiveness on tasks involving class generalization, dataset shifts, and domain adaptation.

Abstract

Pre-trained vision-language (V-L) models such as CLIP have shown excellent generalization ability to downstream tasks. However, they are sensitive to the choice of input text prompts and require careful selection of prompt templates to perform well. Inspired by the Natural Language Processing (NLP) literature, recent CLIP adaptation approaches learn prompts as the textual inputs to fine-tune CLIP for downstream tasks. We note that using prompting to adapt representations in a single branch of CLIP (language or vision) is sub-optimal since it does not allow the flexibility to dynamically adjust both representation spaces on a downstream task. In this work, we propose Multi-modal Prompt Learning (MaPLe) for both vision and language branches to improve alignment between the vision and language representations. Our design promotes strong coupling between the vision-language prompts to ensure mutual synergy and discourages learning independent uni-modal solutions. Further, we learn separate prompts across different early stages to progressively model the stage-wise feature relationships to allow rich context learning. We evaluate the effectiveness of our approach on three representative tasks of generalization to novel classes, new target datasets and unseen domain shifts. Compared with the state-of-the-art method Co-CoOp, MaPLe exhibits favorable performance and achieves an absolute gain of 3.45% on novel classes and 2.72% on overall harmonic-mean, averaged over 11 diverse image recognition datasets. Our code and pre-trained models are available at https://github.com/muzairkhattak/multimodal-prompt-learning.