MaPLe: Multi-modal Prompt Learning [CVPR 2023]

MaPLe: Multi-modal Prompt Learning
Muhammad Uzair Khattak, Hanoona Rasheed, Muhammad Maaz, Salman Khan, Fahad Shahbaz Khan

Official implementation of the paper "MaPLe: Multi-modal Prompt Learning".

Base-to-novel generalization:

Domain Generalization:

🚀 News

• (Feb 28, 2023)
  • Paper accepted at CVPR 2023 🎉
• (Oct 06, 2022)
  • Training and evaluation codes for MaPLe, along with pretrained models are released.
  • The repository also supports CoOp, Co-CoOp, Deep Vision Prompting, Deep Language Prompting, and Independent V-L Prompting architectures.

Highlights

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 models will be publicly released.

Main Contributions

1. Multi-modal prompt learning: Adapt CLIP using a novel prompting technique which prompts both the vision and language branch of CLIP.
2. Vision and Language Prompt Coupling: Explicitly condition vision prompts on their language counterparts and act as a bridge between the two modalities by allowing mutual propagation of gradients to promote synergy.
3. Vision and Language Deep Prompting: Learn multi-modal prompts across multiple transformer blocks in both vision and language branches to progressively learn the synergistic behaviour of both modalities.

☑️ Supported Methods

Results

MaPLe in comparison with existing methods

Results reported below show accuracy for base and novel classes for across 11 recognition datasets averaged over 3 seeds.

Installation

For installation and other package requirements, please follow the instructions detailed in INSTALL.md.

Data preparation

Please follow the instructions at DATASETS.md to prepare all datasets.

Model Zoo

Vision-Language prompting methods

Training and Evaluation

Please refer to the RUN.md for detailed instructions on training, evaluating and reproducing the results using our pre-trained models.

Citation

If you use our work, please consider citing:

@inproceedings{khattakMaPLe,
    title={MaPLe: Multi-modal Prompt Learning},
    author={khattak, Muhammad Uzair and Rasheed, Hanoona and Maaz, Muhammad and Khan, Salman and Khan, Fahad Shahbaz},
    booktitle={The IEEE/CVF Conference on Computer Vision and Pattern Recognition},
    year={2023}
}

Contact

If you have any questions, please create an issue on this repository or contact at [email protected] or [email protected].

Acknowledgements

Our code is based on Co-CoOp and CoOp repository. We thank the authors for releasing their code. If you use our model and code, please consider citing these works as well.