data juicer

Data processing for and with foundation models! 🍎 🍋 🌽 ➡️ ➡️🍸 🍹 🍷

modelscope
3706
209
Python

[中文主页] | [DJ-Cookbook] | [OperatorZoo] | [API] | [Awesome LLM Data]

Data Processing for and with Foundation Models

Data-Juicer



pypi version
Docker version

DataModality
Usage
ModelScope- Demos
HuggingFace- Demos

Document_List
文档列表
OpZoo
Paper
Paper

Data-Juicer is a one-stop system to process text and multimodal data for and with foundation models (typically LLMs).
We provide a playground with a managed JupyterLab. Try Data-Juicer straight away in your browser! If you find Data-Juicer useful for your research or development, please kindly support us by starting it (then be instantly notified of our new releases) and citing our works.

Platform for AI of Alibaba Cloud (PAI) has cited our work and integrated Data-Juicer into its data processing products. PAI is an AI Native large model and AIGC engineering platform that provides dataset management, computing power management, model tool chain, model development, model training, model deployment, and AI asset management. For documentation on data processing, please refer to: PAI-Data Processing for Large Models.

Data-Juicer is being actively updated and maintained. We will periodically enhance and add more features, data recipes and datasets. We welcome you to join us (via issues, PRs, Slack channel, DingDing group, …), in promoting data-model co-development along with research and applications of foundation models!

News

  • new [2025-02-05] We propose a new data selection method, DaaR, which is theoretically informed, via treating diversity as a reward, achieves better overall performance across 7 benchmarks when post-training SOTA LLMs. See more details in Diversity as a Reward: Fine-Tuning LLMs on a Mixture of Domain-Undetermined Data.
  • new [2025-01-11] We release our 2.0 paper, Data-Juicer 2.0: Cloud-Scale Adaptive Data Processing for Foundation Models. It now can process 70B data samples within 2.1h, using 6400 CPU cores on 50 Ray nodes from Alibaba Cloud cluster, and deduplicate 5TB data within 2.8h using 1280 CPU cores on 8 Ray nodes.
  • new [2025-01-03] We support post-tuning scenarios better, via 20+ related new OPs, and via unified dataset format compatiable to LLaMA-Factory and ModelScope-Swift.
  • new [2025-12-17] We propose HumanVBench, which comprises 17 human-centric tasks with synthetic data, benchmarking video-MLLMs’ capabilities from views of inner emotion and outer manifestations. See more details in our paper, and try to evaluate your models with it.
  • new [2024-11-22] We release DJ v1.0.0, in which we refactored Data-Juicer’s Operator, Dataset, Sandbox and many other modules for better usability, such as supporting fault-tolerant, FastAPI and adaptive resource management.
  • [2024-08-25] We give a tutorial about data processing for multimodal LLMs in KDD’2024.
History News: >
  • [2024-08-09] We propose Img-Diff, which enhances the performance of multimodal large language models through contrastive data synthesis, achieving a score that is 12 points higher than GPT-4V on the MMVP benchmark. See more details in our paper, and download the dataset from huggingface and modelscope.
  • [2024-07-24] “Tianchi Better Synth Data Synthesis Competition for Multimodal Large Models” — Our 4th data-centric LLM competition has kicked off! Please visit the competition’s official website for more information.
  • [2024-07-17] We utilized the Data-Juicer Sandbox Laboratory Suite to systematically optimize data and models through a co-development workflow between data and models, achieving a new top spot on the VBench text-to-video leaderboard. The related achievements have been compiled and published in a paper, and the model has been released on the ModelScope and HuggingFace platforms.
  • [2024-07-12] Our awesome list of MLLM-Data has evolved into a systemic survey from model-data co-development perspective. Welcome to explore and contribute!
  • [2024-06-01] ModelScope-Sora “Data Directors” creative sprint—Our third data-centric LLM competition has kicked off! Please visit the competition’s official website for more information.
  • [2024-03-07] We release Data-Juicer v0.2.0 now!
    In this new version, we support more features for multimodal data (including video now), and introduce DJ-SORA to provide open large-scale, high-quality datasets for SORA-like models.
  • [2024-02-20] We have actively maintained an awesome list of LLM-Data, welcome to visit and contribute!
  • [2024-02-05] Our paper has been accepted by SIGMOD’24 industrial track!
  • [2024-01-10] Discover new horizons in “Data Mixture”—Our second data-centric LLM competition has kicked off! Please visit the competition’s official website for more information.
  • [2024-01-05] We release Data-Juicer v0.1.3 now!
    In this new version, we support more Python versions (3.8-3.10), and support multimodal dataset converting/processing (Including texts, images, and audios. More modalities will be supported in the future).
    Besides, our paper is also updated to v3.
  • [2023-10-13] Our first data-centric LLM competition begins! Please
    visit the competition’s official websites, FT-Data Ranker (1B Track, 7B Track), for more information.

Table of Contents

Why Data-Juicer?

  • Systematic & Reusable:
    Empowering users with a systematic library of 100+ core OPs, and 50+ reusable config recipes and
    dedicated toolkits, designed to
    function independently of specific multimodal LLM datasets and processing pipelines. Supporting data analysis, cleaning, and synthesis in pre-training, post-tuning, en, zh, and more scenarios.

  • User-Friendly & Extensible:
    Designed for simplicity and flexibility, with easy-start guides, and DJ-Cookbook containing fruitful demo usages. Feel free to implement your own OPs for customizable data processing.

  • Efficient & Robust: Providing performance-optimized parallel data processing (Aliyun-PAI\Ray\CUDA\OP Fusion),
    faster with less resource usage, verified in large-scale production environments.

  • Effect-Proven & Sandbox: Supporting data-model co-development, enabling rapid iteration
    through the sandbox laboratory, and providing features such as feedback loops and visualization, so that you can better understand and improve your data and models. Many effect-proven datasets and models have been derived from DJ, in scenarios such as pre-training, text-to-video and image-to-text generation.
    Data-in-the-loop

DJ-Cookbook

Curated Resources

Coding with Data-Juicer (DJ)

Use Cases & Data Recipes

  • Data Recipe Gallery
    • Data-Juicer Minimal Example Recipe
    • Reproducing Open Source Text Datasets
    • Improving Open Source Pre-training Text Datasets
    • Improving Open Source Post-tuning Text Datasets
    • Synthetic Contrastive Learning Image-text Datasets
    • Improving Open Source Image-text Datasets
    • Basic Example Recipes for Video Data
    • Synthesizing Human-centric Video Benchmarks
    • Improving Existing Open Source Video Datasets
  • Data-Juicer related Competitions
    • Better Synth, explore the impact of large model synthetic data on image understanding ability with DJ-Sandbox Lab and multimodal large models
    • Modelscope-Sora Challenge, based on Data-Juicer and EasyAnimate framework, optimize data and train SORA-like small models to generate better videos
    • Better Mixture, only adjust data mixing and sampling strategies for given multiple candidate datasets
    • FT-Data Ranker (1B Track, 7B Track), For a specified candidate dataset, only adjust the data filtering and enhancement strategies
    • Kolors-LoRA Stylized Story Challenge, based on Data-Juicer and DiffSynth-Studio framework, explore Difussion model fine-tuning
  • DJ-SORA
  • Based on Data-Juicer and AgentScope framework, leverage agents to call DJ Filters and call DJ Mappers

Interactive Examples

Installation

Prerequisites

  • Recommend Python>=3.9,<=3.10
  • gcc >= 5 (at least C++14 support)

From Source

  • Run the following commands to install the latest basic data_juicer version in
    editable mode:
cd <path_to_data_juicer>
pip install -v -e .
  • Some OPs rely on some other too large or low-platform-compatibility third-party libraries. You can install optional dependencies as needed:
cd <path_to_data_juicer>
pip install -v -e .  # Install minimal dependencies, which support the basic functions
pip install -v -e .[tools] # Install a subset of tools dependencies

The dependency options are listed below:

Tag Description
. or .[mini] Install minimal dependencies for basic Data-Juicer.
.[all] Install all dependencies except sandbox.
.[sci] Install all dependencies for all OPs.
.[dist] Install dependencies for distributed data processing. (Experimental)
.[dev] Install dependencies for developing the package as contributors.
.[tools] Install dependencies for dedicated tools, such as quality classifiers.
.[sandbox] Install all dependencies for sandbox.
  • Install dependencies for specific OPs

With the growth of the number of OPs, the dependencies of all OPs become very heavy. Instead of using the command pip install -v -e .[sci] to install all dependencies,
we provide two alternative, lighter options:

  • Automatic Minimal Dependency Installation: During the execution of Data-Juicer, minimal dependencies will be automatically installed. This allows for immediate execution, but may potentially lead to dependency conflicts.

  • Manual Minimal Dependency Installation: To manually install minimal dependencies tailored to a specific execution configuration, run the following command:

    # only for installation from source
python tools/dj_install.py --config path_to_your_data-juicer_config_file

# use command line tool
dj-install --config path_to_your_data-juicer_config_file

Using pip

  • Run the following command to install the latest released data_juicer using pip:
pip install py-data-juicer
  • Note:
    • only the basic APIs in data_juicer and two basic tools
      (data processing and analysis) are available in this way. If you want customizable
      and complete functions, we recommend you install data_juicer from source.
    • The release versions from pypi have a certain lag compared to the latest version from source.
      So if you want to follow the latest functions of data_juicer, we recommend you install from source.

Using Docker

  • You can

    • either pull our pre-built image from DockerHub:

      docker pull datajuicer/data-juicer:<version_tag>

    • or run the following command to build the docker image including the
      latest data-juicer with provided Dockerfile:

      docker build -t datajuicer/data-juicer:<version_tag> .

    • The format of <version_tag> is like v0.2.0, which is the same as the release version tag.

Installation check

import data_juicer as dj
print(dj.__version__)

For Video-related Operators

Before using video-related operators, FFmpeg should be installed and accessible via the $PATH environment variable.

You can install FFmpeg using package managers(e.g. sudo apt install ffmpeg on Debian/Ubuntu, brew install ffmpeg on OS X) or visit the official ffmpeg link.

Check if your environment path is set correctly by running the ffmpeg command from the terminal.

🔼 back to index

Quick Start

Data Processing

  • Run process_data.py tool or dj-process command line tool with your config as the argument to process
    your dataset.
# only for installation from source
python tools/process_data.py --config configs/demo/process.yaml

# use command line tool
dj-process --config configs/demo/process.yaml

  • Note: For some operators that involve third-party models or resources that are not stored locally on your computer, it might be slow for the first running because these ops need to download corresponding resources into a directory first.
    The default download cache directory is ~/.cache/data_juicer. Change the cache location by setting the shell environment variable, DATA_JUICER_CACHE_HOME to another directory, and you can also change DATA_JUICER_MODELS_CACHE or DATA_JUICER_ASSETS_CACHE in the same way:

  • Note: When using operators with third-party models, it’s necessary to declare the corresponding mem_required in the configuration file (you can refer to the settings in the config_all.yaml file). During runtime, Data-Juicer will control the number of processes based on memory availability and the memory requirements of the operator models to achieve better data processing efficiency. When running with CUDA environments, if the mem_required for an operator is not declared correctly, it could potentially lead to a CUDA Out of Memory issue.

# cache home
export DATA_JUICER_CACHE_HOME="/path/to/another/directory"
# cache models
export DATA_JUICER_MODELS_CACHE="/path/to/another/directory/models"
# cache assets
export DATA_JUICER_ASSETS_CACHE="/path/to/another/directory/assets"
  • Flexible Programming Interface:
    We provide various simple interfaces for users to choose from as follows.
#... init op & dataset ...

# Chain call style, support single operator or operator list
dataset = dataset.process(op)
dataset = dataset.process([op1, op2])
# Functional programming style for quick integration or script prototype iteration
dataset = op(dataset)
dataset = op.run(dataset)

Distributed Data Processing

We have now implemented multi-machine distributed data processing based on RAY. The corresponding demos can be run using the following commands:

# Run text data processing
python tools/process_data.py --config ./demos/process_on_ray/configs/demo.yaml
# Run video data processing
python tools/process_data.py --config ./demos/process_video_on_ray/configs/demo.yaml
  • To run data processing across multiple machines, it is necessary to ensure that all distributed nodes can access the corresponding data paths (for example, by mounting the respective data paths on a file-sharing system such as NAS).
  • The deduplication operators for RAY mode are different from the single-machine version, and all those operators are prefixed with ray, e.g. ray_video_deduplicator and ray_document_deduplicator.
  • More details can be found in the doc for distributed processing.

Users can also opt not to use RAY and instead split the dataset to run on a cluster with Slurm. In this case, please use the default Data-Juicer without RAY.
Aliyun PAI-DLC supports the RAY framework, Slurm framework, etc. Users can directly create RAY jobs and Slurm jobs on the DLC cluster.

Data Analysis

  • Run analyze_data.py tool or dj-analyze command line tool with your config as the argument to analyze your dataset.
# only for installation from source
python tools/analyze_data.py --config configs/demo/analyzer.yaml

# use command line tool
dj-analyze --config configs/demo/analyzer.yaml

# you can also use auto mode to avoid writing a recipe. It will analyze a small
# part (e.g. 1000 samples, specified by argument `auto_num`) of your dataset 
# with all Filters that produce stats.
dj-analyze --auto --dataset_path xx.jsonl [--auto_num 1000]
  • Note: Analyzer only computes stats for Filters that produce stats or other OPs that produce tags/categories in meta. So other OPs will be ignored in the analysis process. We use the following registries to decorate OPs:
    • NON_STATS_FILTERS: decorate Filters that DO NOT produce any stats.
    • TAGGING_OPS: decorate OPs that DO produce tags/categories in meta field.

Data Visualization

  • Run app.py tool to visualize your dataset in your browser.
  • Note: only available for installation from source.
streamlit run app.py

Build Up Config Files

  • Config files specify some global arguments, and an operator list for the
    data process. You need to set:
    • Global arguments: input/output dataset path, number of workers, etc.
    • Operator list: list operators with their arguments used to process the dataset.
  • You can build up your own config files by:
    • ➖：Modify from our example config file config_all.yaml which includes all ops and default
      arguments. You just need to remove ops that you won’t use and refine
      some arguments of ops.
    • ➕：Build up your own config files from scratch. You can refer our
      example config file config_all.yaml, op documents, and advanced Build-Up Guide for developers.
    • Besides the yaml files, you also have the flexibility to specify just
      one (of several) parameters on the command line, which will override
      the values in yaml files.
python xxx.py --config configs/demo/process.yaml --language_id_score_filter.lang=en

  • The basic config format and definition is shown below.

    Basic config example of format and definition

Sandbox

The data sandbox laboratory (DJ-Sandbox) provides users with the best practices for continuously producing data recipes. It features low overhead, portability, and guidance.

  • In the sandbox, users can quickly experiment, iterate, and refine data recipes based on small-scale datasets and models, before scaling up to produce high-quality data to serve large-scale models.
  • In addition to the basic data optimization and recipe refinement features offered by Data-Juicer, users can seamlessly use configurable components such as data probe and analysis, model training and evaluation, and data and model feedback-based recipe refinement to form a complete one-stop data-model research and development pipeline.

The sandbox is run using the following commands by default, and for more information and details, please refer to the sandbox documentation.

python tools/sandbox_starter.py --config configs/demo/sandbox/sandbox.yaml

Preprocess Raw Data (Optional)

  • Our Formatters support some common input dataset formats for now:
    • Multi-sample in one file: jsonl/json, parquet, csv/tsv, etc.
    • Single-sample in one file: txt, code, docx, pdf, etc.
  • However, data from different sources are complicated and diverse. Such as:
    • Raw arXiv data downloaded from S3 include thousands of tar files and even more gzip files in them, and expected tex files are embedded in the gzip files so they are hard to obtain directly.
    • Some crawled data include different kinds of files (pdf, html, docx, etc.). And extra information like tables, charts, and so on is hard to extract.
  • It’s impossible to handle all kinds of data in Data-Juicer, issues/PRs are welcome to contribute to processing new data types!
  • Thus, we provide some common preprocessing tools in tools/preprocess for you to preprocess these data.
    • You are welcome to make your contributions to new preprocessing tools for the community.
    • We highly recommend that complicated data can be preprocessed to jsonl or parquet files.

For Docker Users

  • If you build or pull the docker image of data-juicer, you can run the commands or tools mentioned above using this docker image.
  • Run directly:
# run the data processing directly
docker run --rm \  # remove container after the processing
  --privileged \
  --shm-size 256g \
  --network host \
  --gpus all \
  --name dj \  # name of the container
  -v <host_data_path>:<image_data_path> \  # mount data or config directory into the container
  -v ~/.cache/:/root/.cache/ \  # mount the cache directory into the container to reuse caches and models (recommended)
  datajuicer/data-juicer:<version_tag> \  # image to run
  dj-process --config /path/to/config.yaml  # similar data processing commands
  • Or enter into the running container and run commands in editable mode:
# start the container
docker run -dit \  # run the container in the background
  --privileged \
  --shm-size 256g \
  --network host \
  --gpus all \
  --rm \
  --name dj \
  -v <host_data_path>:<image_data_path> \
  -v ~/.cache/:/root/.cache/ \
  datajuicer/data-juicer:latest /bin/bash

# enter into this container and then you can use data-juicer in editable mode
docker exec -it <container_id> bash

🔼 back to index

License

Data-Juicer is released under Apache License 2.0.

Contributing

We are in a rapidly developing field and greatly welcome contributions of new
features, bug fixes, and better documentation. Please refer to
How-to Guide for Developers.

Acknowledgement

Data-Juicer is used across various foundation model applications and research initiatives, such as industrial scenarios in Alibaba Tongyi and Alibaba Cloud’s platform for AI (PAI).
We look forward to more of your experience, suggestions, and discussions for collaboration!

Data-Juicer thanks many community contributors and open-source projects, such as
Huggingface-Datasets, Bloom, RedPajama, Arrow, Ray, …

References

If you find Data-Juicer useful for your research or development, please kindly cite the following works, 1.0paper, 2.0paper.

@inproceedings{djv1,
  title={Data-Juicer: A One-Stop Data Processing System for Large Language Models},
  author={Daoyuan Chen and Yilun Huang and Zhijian Ma and Hesen Chen and Xuchen Pan and Ce Ge and Dawei Gao and Yuexiang Xie and Zhaoyang Liu and Jinyang Gao and Yaliang Li and Bolin Ding and Jingren Zhou},
  booktitle={International Conference on Management of Data},
  year={2024}
}

@article{djv2,
  title={Data-Juicer 2.0: Cloud-Scale Adaptive Data Processing for Foundation Models},
  author={Chen, Daoyuan and Huang, Yilun and Pan, Xuchen and Jiang, Nana and Wang, Haibin and Ge, Ce and Chen, Yushuo and Zhang, Wenhao and Ma, Zhijian and Zhang, Yilei and Huang, Jun and Lin, Wei and Li, Yaliang and Ding, Bolin and Zhou, Jingren},
  journal={arXiv preprint arXiv:2501.14755},
  year={2024}
}
More related papers from the Data-Juicer Team: >

🔼 back to index