Git Product home page Git Product logo

p-multimodal-dataset-toolbox's Introduction

Pinterest Multimodal Dataset ToolBox

Created by Junhua Mao

Introduction

This is a toolbox to download and manage the released part of the Pinterest40M multimodal dataset introduced in the paper Training and Evaluating Multimodal Word Embeddings with Large-scale Web Annotated Images. More information can be found on the [Project Page](http://www.stat. ucla.edu/~junhua.mao/multimodal_embedding.html).

Cite

If you find this dataset or toolbox useful in your research, please cite:

@inproceedings{mao2016training,
  title={Training and Evaluating Multimodal Word Embeddings with Large-scale Web Annotated Images},
  author={Mao, Junhua and Xu, Jiajing and Jing, Yushi and Yuille, Alan},
  booktitle={NIPS},
  year={2016}
}

Toolbox Installation and Data Downloading

Download and setup meta files.

Suppose that toolkit is install on $PATH_PTool:

cd $PATH_PTool
bash download_meta.sh

Download images.

You can easily download images in parallel (12 workers by default) and resize the downloaded images to 224x224:

cd $PATH_PTool
python download_images.py

There are ~5 million images in the dataset. The download process can take days.

The script allows you to resume your downloading at any time. Just re-run download_images.py if your downloading is shutted down unexpectedly. It is possible that you failed to access some of the urls at the first time. Re-run download_images.py to have another try.

You are welcome to read download_images.py and py_utils.py for personalized and advanced downloading settings (e.g. see the docstring of py_utils.PinDataset.download_images).

Demo

View demo.ipynb for how to use this toolbox.

Use pin_2016_v1_0000.npy to pin_2016_v1_0097.npy as the training set.

Use pin_2016_v1_0098.npy as the validation set.

Use pin_2016_v1_0099.npy as the test set.

License

The copyright of the annotations and the images belongs to the original source. This meta data file can be used for research proposes only.

This toolbox is licensed under a Creative Commons Attribution 4.0 International License.

p-multimodal-dataset-toolbox's People

Contributors

mjhucla avatar

Watchers

 avatar  avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.