Introduction
Welcome to the latest developer documentation of Datature!
Datature
Datature is the fastest and easiest way for individuals, teams, and enterprises to build end-to-end computer vision pipelines - without a single line of code!
Our mission is to democratize AI vision capabilities and streamline features that are often blockers in orchestrating AI projects. With the rise of citizen data scientists, deep tech companies, and enterprises looking to adopt deep-learning, we want to make sure individuals, teams, and companies are equipped with the right tools to build their own AI capabilities easily and quickly 🤝
What We Do 🤨
Datature simplifies every single step of a deep learning pipeline, from annotation to model deployment. Users can manage datasets, annotate, generate synthetic data, train and deploy models - all in a single, secure cloud-based platform.
How We Do It 💡
Our users simply have to build a visual workflow using our drag-and-drop UI, and our platform will transpile the visual workflow into over thousands of lines of conformant machine learning code!
Why We Do It 🥰
Our greatest joy comes from seeing our users spend less time coding and debugging, and more time making real impact with their research and product.
Quick Demo 📽▶️
Read our Datature Pipeline documentation or watch a 5 minutes introduction to our platform Nexus👇, to learn how we revolutionize the way teams build computer vision models.
Nexus
Nexus is Datature's no-code AI platform that simplifies the development and deployment of computer vision solutions for users. Users can create an end-to-end computer vision pipeline - from onboarding data, annotation, model training, model evaluation, and model exporting - all on a single no-code platform.
What can I do with Nexus?
Nexus provides a broad range of services, from annotation to model training to model deployment - To better understand the overall pipeline, go to Datature Pipeline.
To look at individual services:
- Managing Your Projects
- Onboarding Data and Labels
- Creating Annotations
- Creating a Training Workflow
- Running the Workflow
- Monitoring Training Process
- Evaluating Model Performance
- Managing Trained Models
Portal
Portal is Datature's no-code interface designed to help teams, engineers, and product managers interactively test their trained models from both Nexus and custom trained models in an accessible manner. Portal provides tools for visualizing inference with customizable inference settings, test dataset management, trained model management, and easy customizability for developers.
What can I do with Portal?
- Registering and Deregistering a Model
- Loading and Unloading a Model
- Adding and Removing Images and Videos
- Running Inferences
- Using Your Own Custom Models
See also
- Datature Quickstart Guide - to get started with Datature
Slack Community
Join our slack community to ask questions, share knowledge, and connect with the community! We are a developer-and-researcher-first company and we believe that the product should constantly evolve to fit more use cases. We are open to suggestions and hope to build a community around the product. To chat with us and submit ideas, join Datature Community Slack today 🚀
Datature Developer Hub
The Datature Developer Hub contains comprehensive guides and documentation to help you start working with Datature as quickly as possible, as well as support if you get stuck.
Tutorials
If you are more of a visual and auditory learner, our Tutorial Videos will be a great place for you to get more information about how to build your computer vision models with Datature. We create and upload tutorial videos regularly to introduce and demonstrate platform features.
Blog
The Datature Blog is the best place to get the latest information on product features, developments, industry trends and best practices. We publish articles on a wide range of topics for readers from different industries with different levels of tech expertise.
Github Repositories
You can find many useful scripts and Jupyter notebooks on our Github - For instance, this repository contains resources used in our tutorials and guides ⚡️
Updated 2 months ago