Prevision.io — Documentation

Value proposition

Prevision.io brings powerful AI management capabilities to data science users so more AI projects make it into production and stay in production. Our purpose-built AI Management platform was designed by data scientists for data scientists and citizen data scientists to scale their value, domain expertise, and impact. The platform manages the hidden complexities and burdensome tasks that get in the way of realizing the tremendous productivity and performance gains AI can deliver across your business

Requirements

Prevision.IA is a SAAS platform optimized for Firefox and Chrome navigators. The cloud version can be accessed online at https://cloud.prevision.io, or it can be deployed on-premise or in your private cloud. Please visit us at https://prevision.io if you have any questions regarding our deployment capabilities.

Conditions

Please read the general terms and conditions available on following link : https://cloud.prevision.io/terms

contacts

If you have any questions about using Prevision.io platform please contact us using the chat button on the Prevision.io store interface or by email at the following contact address :

support@prevision.io

Getting started

There are to ways to get a Prevision.io account :

1. Go to https://cloud.prevision.io and open an account ( or link one of your OAuth2 enabled account )

2. Deploy a dedicated computing service on Google Cloud

In order to deploy Prevision.io on Google Cloud, go to the dedicated Guide

Account creation

By clicking to the following address, you will land on the connection page which allows you to create an account or sign in if you already have a Prevision.io account.

https://cloud.prevision.io

In order to create a new account, just click on the sign up button next to the log in button. You will access the following account creation form.

Signup screen

Once you have filled the needed information, you will have a 30 days free but limited access to the Prevision.io platform. In order to upgrade your free trial account into a full access one, please contact us on following email (support@prevision.io)

Once done you can follow our complete guide to release a model

Connection

Once your account has been created, you will be able to access the prevision’s Studio and Store and start creating models and deploying them.

Please note that SSO using you google/linkedin/GitHub account is available.

login with SSO

Cloud & freetrial limitations

If you are using our cloud platform using a free trial account, some limitations are set up. Here’s a quick view of limitations for free accounts:

Freetrial Limitation

Entity	Action	Limitation
PROJECT	Create Project	Free trial users can create 2 Limited Projects
DATASETS	Add dataset from file / datasource	10 Datasets max + 1GB per dataset
IMAGE FOLDER	Add / Update / Delete image Folder in project	1 Image Folder
DATA SOURCE	Add / Update / Delete datasource in project	1 Datasource max
CONNECTOR	Add / Update / Delete connector in project	1 Connector max
USECASE	Add / Update / Delete experiment in project	5 Use Cases max
USECASE VERSION	Add / Update experiment version in project	3 Concurrent experiment versions
PREDICTION	Add / Update prediction in experiment version	2 Concurrent predictions
STORE APP	Deploy Apps	5 Concurrent deployed apps

Reference Documentation

• Studio
  • Projects
    • List my Projects
    • Create a new project
    • Project navigation
      • collaboration into a project
      • Project roles
    • Edit a project
    • Delete a project
  • Data
    • Datasets
      • Create a new dataset
      • Analyse dataset
      • Edit and delete
      • Compute embeddings
    • Image folders
      • Create a new image folder
      • Edit and remove
    • Connectors
      • Create Connector
    • Data sources
      • Table as a datasource
      • Query as a datasource
      • S/FTP server file as datasource
      • S3 Bucket
      • Datasource from GCP
      • Import your datas
    • Exporters
      • Create an exporter
      • Using an exporter - export your dataset
  • Experiments
    • List your experiments
    • Create a new experiment
      • AutoML Engine and ONNX import
      • Data type
    • Training type
      • Structured Data
      • Computer Vision Model
      • Time series
      • Natural Language Processing
    • Import External models into your experiment
    • Inspect your experiment and evaluate your models
      • Models
      • Features
      • Predictions
      • Versions
      • Tasks
      • Reports
  • Pipelines
    • Pipeline components
      • Building you own component
    • Pipeline templates
    • Scheduled runs
      • When to used scheduled runs ?
      • Create a new scheduled running
      • See all your scheduled run
      • See run state, trigger a run
      • Edit a run
      • Delete a Scheduled run
  • Deployments
    • Create a new Deployments
    • Inspect and monitor a deployed experiment
      • General
      • Monitoring
      • Predictions
      • Versions
      • Usage
    • Edit and remove a deployment
  • Contributors
    • Roles & rules
    • Add & delete
    • Project settings
  • Notebooks
    • Introduction
    • Jupyter (python)
    • Jupyter (R studio)
  • Settings
    • Connecting my code
• API
  • Using The API
• SDK
  • Using the Python SDK
    • Getting started
    • Standard Machine Learning Worflow with Prevision SDK
      • External Model
    • API reference
    • Source code
  • Using the R SDK
  • Using the Prevision Quantum NN SDK
    • Getting started
    • API reference
    • Source code
    • Blog post
• Guides and How-to
  • Datascience Guide
    • How and when to do data embedding exploration
      • What is this guide about ?
      • Data exploration in the Machine Learning pipeline
      • Data exploration technic focus : Data Embedding
      • Data Embedding exploration in Prevision Platform
      • Getting and Working with the Embedding Vector
      • Going Further and use the API
    • Full ML pipeline: From data collection to deploying using Prevision.io
      • What’s in this guide ?
      • Data acquisition
      • Feature engineering
      • Define the problem
      • Experiment
      • Evaluate
      • Deploy
      • Schedule
      • Monitoring
      • Conclusion
    • Cross validating your machine learning models
      • Cross Validation Strategies
      • Validation technique: Know about the future performance
      • Common techniques to prevent performance collapse
      • The impact of a cross-validation strategy
    • Text Classification with Prevision.io Data Platform
      • Text classification experiment with Prevision.io’s UI
      • Text classification experiment with Prevision.io’s SDK
      • Conclusion
    • Synthetic Data for machine Learning : A guide
      • Usage of synthetic Data
      • Performance of synthetic Data Generator evaluation
      • Statistical properties of Generated Synthetic Data
      • Generating good Synthetic Data
      • Machine Learning efficiency
      • Going Further
  • A list of some useful Dataset to explore Machine Learning
    • Tabular Dataset
    • Images
    • NLP
    • Externals models