MAGMA -- Multimodal Augmentation of Generative Models through Adapter-based Finetuning

Paper: https://arxiv.org/abs/2112.05253

Abstract

Large-scale pretraining is fast becoming the norm in Vision-Language (VL) modeling. However, prevailing VL approaches are limited by the requirement for labeled data and the use of complex multi-step pretraining objectives. We present MAGMA - a simple method for augmenting generative language models with additional modalities using adapter-based finetuning. Building on Frozen, we train a series of VL models that autoregressively generate text from arbitrary combinations of visual and textual input. The pretraining is entirely end-to-end using a single language modeling objective, simplifying optimization compared to previous approaches. Importantly, the language model weights remain unchanged during training, allowing for transfer of encyclopedic knowledge and in-context learning abilities from language pretraining. MAGMA outperforms Frozen on open-ended generative tasks, achieving state of the art results on the OKVQA benchmark and competitive results on a range of other popular VL benchmarks, while pretraining on 0.2% of the number of samples used to train SimVLM.

Usage

from magma import Magma

from huggingface_hub import hf_hub_url, cached_download

checkpoint_url = hf_hub_url(repo_id="osanseviero/magma", filename="model.pt")
checkpoint_path = cached_download(checkpoint_url)

model = Magma.from_checkpoint(
    config_path = "configs/MAGMA_v1.yml",
    checkpoint_path = checkpoint_path,
    device = 'cuda:0'
)