In September, a new version of OpenDataLab was launched, supporting users to independently publish original data sets. At the same time, the CLI/SDK tools and data set details page were upgraded to make open source AI data sets more convenient and display clearer. There is also an activity to receive gifts for your creations, come and take a look!
(Attention! The old version of the CLI tool will cease operation and maintenance in the near future. Please install the latest version in time to avoid affecting use)
1. Easily open source your data set
Choosing an easy-to-use AI data publishing platform will get twice the result with half the effort in promoting open source results.
Use OpenDataLab to publish original data sets with one click and enjoy 3 major advantages:
● No need for website development and maintenance, saving various operation and maintenance costs;
● Standardized review mechanism and open source process to protect data and copyright security;
● Ultra-large storage capacity and network Accelerate, no need to surf the Internet scientifically, and instantly enjoy the high-speed transmission experience at home and abroad.
Now, it only takes 4 simple steps: register an author account → create a data set warehouse → upload data → submit and make it public , and easily complete the publishing and sharing of your original data set.
For detailed steps, please see the document:
https://openxlab.org.cn/docs/datasets/dataset creation process.html
For sensitive data that is restricted from opening, the platform has launched an approval function, allowing users to submit applications to authors, sign a use agreement, and open downloads after the author agrees. (If you need to approve the settings, please contact the OpenDataLab assistant to set them manually)
The platform has a strict "algorithm + manual" review mechanism embedded in it, which can effectively identify and process high-risk and highly sensitive data, ensuring that data can be uploaded and used safely and standardizedly to the greatest extent. At the same time, everyone is welcome to actively correct errors and report.
2. CLI/SDK to obtain and share data with one click
The OpenDataLab command line interface (CLI, Command Line Interface) is a very convenient tool that allows users to download public data sets on OpenDataLab . It supports Windows, Linux and Mac platforms and is highly praised by users.
In order to improve the experience, a new CLI command line and Python SDK tools have been released this time, with new data set upload and management functions. You can use commands to view, create, upload, download, and edit open source data sets with one click. The breakpoint resume function is added to make data transmission more stable and faster.
Install the latest version of CLI/SDK immediately, complete the corresponding authentication configuration, and you are ready to use. Download instructions for each dataset are available on the details page.
Dataset CLI (command line tool) Detailed description:
https://openxlab.org.cn/docs/developers/dataset/dataset CLI (command line tool).html
Dataset Python SDK detailed description:
https://openxlab.org.cn/docs/developers/dataset/datasetpython%20SDK.html
(Dataset details page, download instructions and instructions)
3. The details page displays more abundantly
The newly upgraded data set details page has new data set introduction, data set details, and settings sections in addition to the original data set label bar and release information bar.
(Indicated on the data set details page)
●The author of "Dataset Introduction"
can edit the introduction information flexibly and personalizedly in Markdown format, add cover pages, citations, statistical charts, URL links, etc., and build a unique data set display page;
● The image format files uploaded in "Data Details" can be automatically parsed, and the data details can be used as sample previews and statistics to make the data structure clear at a glance;
●The "data file"
author can choose to upload data in three different forms from the web page: file, folder, compressed package, etc.; if you select "compressed package", the system will automatically decompress your file after uploading;
● "Settings"
Data set authors can freely set the status of the data set, "private" or "public". The created data set defaults to "private" status and is only visible to the author. It needs to be manually made public before users can access it to facilitate maintenance and management.
4. Create and receive gifts
Contact the assistant and reply "Sign up" to participate in the event.
Before 12:00 on September 8, the first 20 people who successfully submit the link to the original data set
will receive a gift package worth 100 yuan. Come and sign up for
more public data sets. Welcome to visit the official website of OpenDataLab View and download: https://opendatalab.org.cn/