Skip to content

Latest commit

 

History

History

ofa

OFA

OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework

Abstract

In this work, we pursue a unified paradigm for multimodal pretraining to break the scaffolds of complex task/modality-specific customization. We propose OFA, a Task-Agnostic and Modality-Agnostic framework that supports Task Comprehensiveness. OFA unifies a diverse set of cross-modal and unimodal tasks, including image generation, visual grounding, image captioning, image classification, language modeling, etc., in a simple sequence-to-sequence learning framework. OFA follows the instruction-based learning in both pretraining and finetuning stages, requiring no extra task-specific layers for downstream tasks. In comparison with the recent state-of-the-art vision & language models that rely on extremely large cross-modal datasets, OFA is pretrained on only 20M publicly available image-text pairs. Despite its simplicity and relatively small-scale training data, OFA achieves new SOTAs in a series of cross-modal tasks while attaining highly competitive performances on uni-modal tasks. Our further analysis indicates that OFA can also effectively transfer to unseen tasks and unseen domains.

How to use it?

Use the model

from mmpretrain import inference_model

result = inference_model('ofa-base_3rdparty-finetuned_caption', 'demo/cat-dog.png')
print(result)
# {'pred_caption': 'a dog and a kitten sitting next to each other'}

Test Command

Prepare your dataset according to the docs.

Test:

python tools/test.py configs/ofa/ofa-base_finetuned_refcoco.py https://download.openmmlab.com/mmclassification/v1/ofa/ofa-base_3rdparty_refcoco_20230418-2797d3ab.pth

Models and results

Image Caption on COCO

Model Params (M) BLEU-4 CIDER Config Download
ofa-base_3rdparty-finetuned_caption* 182.24 42.64 144.50 config model

Models with * are converted from the official repo. The config files of these models are only for inference. We haven't reproduce the training results.

Visual Grounding on RefCOCO

Model Params (M) Accuracy (testA) Accuracy (testB) Config Download
ofa-base_3rdparty-finetuned_refcoco* 182.24 90.49 83.63 config model

Models with * are converted from the official repo. The config files of these models are only for inference. We haven't reproduce the training results.

Visual Question Answering on VQAv2

Model Params (M) Accuracy Config Download
ofa-base_3rdparty-finetuned_vqa* 182.24 78.00 config model
ofa-base_3rdparty-zeroshot_vqa* 182.24 58.32 config model

Models with * are converted from the official repo. The config files of these models are only for inference. We haven't reproduce the training results.

Citation

@article{wang2022ofa,
  author    = {Peng Wang and
               An Yang and
               Rui Men and
               Junyang Lin and
               Shuai Bai and
               Zhikang Li and
               Jianxin Ma and
               Chang Zhou and
               Jingren Zhou and
               Hongxia Yang},
  title     = {OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence
               Learning Framework},
  journal   = {CoRR},
  volume    = {abs/2202.03052},
  year      = {2022}
}