Quick Start

Welcome to the Quick Start or Getting Started section, where you can find guides to help you start using DagsHub and get up and running quickly.

In the Doc home, you may have seen a simple DagsHub flow diagram. Below is a detailed version of that same diagram. It shows how you might go from 0 to trained model deployed to a production environment with DagsHub.

DagsHub Workflow Diagram

You can break down the DagsHub diagram into 5 main steps (See these in the sidebar too):

1. Creating a DagsHub repository (or connecting an existing repository)
2. Uploading, connecting or versioning your data
3. Building or curating your dataset
4. Tracking your experiments
5. Managing your models

A Notebook Tutorial

If you want to learn DagsHub through a notebook tutorial - we have you covered with the Hello World Colab option below.

•  Hello World Colab

  ---

  A step-by-step computer vision tutorial, zero local setup needed. For other data types, see our tutorials section

The tutorial covers all the steps above without any local setup required.

Learn DagsHub Step-By-Step

If you want to learn step-by-step, or focus on a specific part of the flow, we have you covered with a series of video tutorials.

•  1. Creating a DagsHub repository (choose one):

  ---

  • Create a new repository
  • Connect an existing repository

•  2. Uploading, connecting or versioning your data (choose one):

  ---

  • Upload your data
  • Connect an existing repository
  • Version your data

•  3. Building or curating your dataset:

  ---

  • Create a dataset
  • Add metadata
  • Query your dataset
  • Annotate your dataset
    • Auto-Label your dataset
  • Export & augment your dataset

•  4. Tracking your experiments:

  ---

  • Monitoring your training runs
  • Logging artifacts & models
  • Comparing experiment results

•  5. Managing your models:

  ---

  • Registering your models
  • Deploying your models