|
"""PMC-OA Dataset""" |
|
|
|
import os |
|
import jsonlines |
|
|
|
import datasets |
|
|
|
|
|
logger = datasets.logging.get_logger(__name__) |
|
|
|
|
|
_CITATION = """\ |
|
@article{lin2023pmc, |
|
title={PMC-CLIP: Contrastive Language-Image Pre-training using Biomedical Documents}, |
|
author={Lin, Weixiong and Zhao, Ziheng and Zhang, Xiaoman and Wu, Chaoyi and Zhang, Ya and Wang, Yanfeng and Xie, Weidi}, |
|
journal={arXiv preprint arXiv:2303.07240}, |
|
year={2023} |
|
} |
|
""" |
|
|
|
_DESCRIPTION = """\ |
|
Foundation models trained on large-scale dataset gain a recent surge in CV and NLP. In contrast, development in biomedical domain lags far behind due to data scarcity. |
|
To address this issue, we build and release PMC-OA, a biomedical dataset with 1.6M image-caption pairs collected from PubMedCentral's OpenAccess subset, which is 8 times larger than before. |
|
PMC-OA covers diverse modalities or diseases, with majority of the image-caption samples aligned at finer-grained level, i.e., subfigure and subcaption. |
|
While pretraining a CLIP-style model on PMC-OA, our model named PMC-CLIP achieves state-of-the-art results on various downstream tasks, |
|
including image-text retrieval on ROCO, MedMNIST image classification, Medical VQA, i.e. +8.1% R@10 on image-text retrieval, +3.9% accuracy on image classification. |
|
""" |
|
|
|
_HOMEPAGE = "https://weixionglin.github.io/PMC-CLIP/" |
|
|
|
_URL = "https://huggingface.co/datasets/axiong/pmc_oa/resolve/main/" |
|
_URLS = { |
|
"images": _URL + "images.zip", |
|
"train": _URL + "train.jsonl", |
|
"valid": _URL + "valid.jsonl", |
|
"test": _URL + "test.jsonl" |
|
} |
|
|
|
|
|
class PMC_OA(datasets.GeneratorBasedBuilder): |
|
"""PMC_OA Dataset.""" |
|
|
|
def _info(self): |
|
return datasets.DatasetInfo( |
|
description=_DESCRIPTION, |
|
features=datasets.Features( |
|
{ |
|
|
|
"image": datasets.Image(), |
|
"caption": datasets.Value("string"), |
|
} |
|
), |
|
supervised_keys=None, |
|
homepage=_HOMEPAGE, |
|
citation=_CITATION, |
|
) |
|
|
|
def _split_generators(self, dl_manager): |
|
data_dir = dl_manager.download_and_extract(_URLS) |
|
|
|
return [ |
|
datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": os.path.join(data_dir, "train.jsonl")}), |
|
datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={"filepath": os.path.join(data_dir, "valid.jsonl")}), |
|
datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs={"filepath": os.path.join(data_dir, "test.jsonl")}), |
|
] |
|
|
|
def _generate_examples(self, filepath): |
|
"""This function returns the examples in the raw (text) form.""" |
|
logger.info("generating examples from = %s", filepath) |
|
|
|
with jsonlines.open(filepath) as reader: |
|
_id = 0 |
|
for obj in reader: |
|
relative_image_path = obj['image'] |
|
image_path = os.path.join("images", relative_image_path) |
|
caption = obj['caption'] |
|
yield _id, { |
|
"image": { |
|
"path": image_path, |
|
"bytes": open(image_path, "rb").read(), |
|
}, |
|
"caption": caption, |
|
} |
|
_id += 1 |
|
|
|
|
