Skip to content

A Git-like Version Control File System for Datasets Management in the Era of AI.

License

Notifications You must be signed in to change notification settings

PreResearch-Labs/jzfs-datasets-ai-vcf

 
 

Repository files navigation

JZFS

Golang implementation of JZFS: version control file system for datasets management in the era of AI.



JZFS is an industry-leading Data-Centric Version Control File System, helps ensure Responsible AI Engineering by improving Data Versioning, Provenance, and Reproducibility.

Note:

  • The name JZFS pays tribute to the world's earliest paper money: Song Dynasty JiaoZi.
  • JZFS is yet another implementation of IPFS (InterPlanetary File System) as JZFS will be compatible with the implementation requirements of IPFS.
  • As a filesystem of data versioning at scale, although JZFS is built for machine learning, It has a wide range of use scenarios (refer A Universe of Uses) and can be seamlessly integrated into all your data stack.

Data-centric AI is about the practice of iterating and collaborating on data, used to build AI systems, programmatically. Machine learning pioneer Andrew Ng argues that focusing on the quality of data fueling AI systems will help unlock its full power.


Features

In production systems with machine learning components, updates and experiments are frequent. New updates to models(data products) may be released every day or every few minutes, and different users may see the results of different models as part of A/B experiments or canary releases.

  • Version Everything: Data scientists are often criticized for being less disciplined with versioning their experiments(versioning of data, pipeline, code, and models), especially when using computational notebooks.
  • Track Data Provenance: This applies to all processing steps in an AI/ML pipeline, including data collection/acquisition, data merging, data cleaning, feature extraction, learning, or deployment.
  • Reproducibility: A final question of AI/ML that is often relevant for debugging, audits, and also science more broadly is to what degree data, models, and decisions can be reproduced.

Getting Started

Requirement

  1. To build JZFS, you need a working installation of Go 1.22.0 or higher
  2. JZFS use postgres to store running data, you can install at postgres install installation guide

Build And Running

  1. clone and build
git clone https://github.com/GitDataAI/jzfs.git
cd jzfs
make build

After following the above steps, you should be able to see an executable file named "jzfs."

  1. init program and running
./jzfs init  --db postgres://<username>:<password>@localhost:5432/jiaozifs?sslmode=disable
./jzfs daemon

run with docker

docker run -v <data>:/app -p 34913:34913 gitdatateam/jzfs:latest  --db "postgres://<user>:<password>@192.168.1.16:5432/jiaozifs?sslmode=disable" --bs_path /app/data --listen http://0.0.0.0:34913 --config /app/config.toml

Cloud

Try without installing

Note: storage config for IPFS backend storage as you create a new repository in JZFS Console.

 {"type":"ipfs","ipfs":{"url":"/dns/kubo-service.ipfs.svc.cluster.local/tcp/5001"}}

Examples

Build AL/ML pipeline over JZFS
Face detection and recognition inference pipeline


Documentation

Official Documentation


Users and Partners

Lighthouse Permanent Storage
MesoReef DAO: Decentralized Science for Regenerating
LunCo
Artizen Fund
HaAI Labs


Contributors


License

Dual-licensed under MIT + Apache 2.0

About

A Git-like Version Control File System for Datasets Management in the Era of AI.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Go 99.9%
  • Other 0.1%