Versions
Open Source and Enterprise
Overview
Here we explain a bit about what are the differences on the open source & enterprise editions. The features are divided into three major sub areas, Annotation, Catalog, and Workflow.
License
See License
Open Source Features
- Core Platform
- Development System
- Annotate All Media: Image, Video, 3D, Text, Geo, Audio, and more Annotation Summary
- User Friendly
- Deep Data Lake Feature Store (Data Catalog)
- Unlimited Annotations and Automations
- Unlimited API & Python SDK/CLI Access
- Auto-Annotate (ML Automation)
- Automated task distribution
- Collaboration across teams
- Performance dashboards
- Robust QA & collaboration capabilities
- Model-assisted labeling (Pre-Label)
- All special tools (Video segmentation, medical, tiled)
- All Attributes and Geometric Templates
- Project Management & Data Curation
- Workflows
- Data Explorer (model diagnostics)
- Role Based Access Control (RBAC)
- Long 4k Videos & Ultra-High Resolution
- DICOM Native (Coming Soon)
- All Cloud Integrations (AWS, GCP, Azure)
- Issues & Comments
- Identity (OIDC, SAML, SSO, LDAP, more)
- Compliance (SOC 2, GDPR, NIST, more)
- Security
- Support: Quality Docs, Global Slack Community
More Included Features
- Streamlined Annotation UI suitable both from "First Time" Subject Matter Experts, and powerful options for Professional Full Time Annotators
- Many User Labeling - Designed for many users from Day 1.
- Scale to Mega Projects with sophisticated organizational concepts.
- Ingest prediction data with less scripts.
- Import Wizard saves you hours having to map your data (pre-labels, QA, debug etc.).
- All-Cloud Integrated File Browser
- Scalable pipeline for massive ingestion - we have tested to 600+ hardware nodes
- Integrated pipeline hooks - newly added data auto creates tasks and more
- Collaboration across teams between machine learning, product, ops, managers, and more.
- Store virtually any scale of dataset and instantly access slices of the data to avoid having to download/unzip/load.
- Fast access to datasets from multiple machines. Have multiple Data Scientists working on the same data.
- Integrates with your tools and 3rd party workforces.
It's a database for your training data, both metadata and access of raw BLOB data (over top of your storage choice). - Quality Assurance Features Summary
- Webhooks with Actions
- Continually upgrade your data, including easily adding more depth
to existing partially annotated sets. - Skip downloading and unzipping massive datasets. Explore data instantly through the browser.
Try it now (click Dataset Explorer) - Automatic Dataset Versioning and user definable datasets.
- Collaborate share and comment on specific instances with a Diffgram Permalink.
- Compare model results visually
- Curate data and send for labeling with one click
- Uncover bad data and edge cases. Use your models to debug the human. Visually see errors.
Stream Training Data is Now Available. - Colab Notebook Example || Docs Pytorch Tensorflow
- Run models instantly with Javascript
- Make HTTP requests to your model based on user events
- Runs on your local system or cloud. Less lag, more secure, more control.
- Enforce PII & RBAC automatically across life-cycle of
training data from ingest to dataset to model predictions and back again - Task features can be used as modules within Workflow
- Automatic Per Task Review Routing, with configurable review chance
- Reporting
Enterprise
- Enterprise use license.
- Everything in Open Source
- Compliance (SOC 2, GDPR, NIST, more)
- Identity (OIDC, SAML, SSO, LDAP, more)
- Security
- Custom Deal Terms
- Custom Unlimited Users
- Custom Software Engineering
- Best Support
- Model Training (AutoML or Custom)
- 4 Hour Response SLA Standard
- 15 Minute Response SLA Available
- Up to Five 9’s of availability
- Value Add Services (Data Science, Project and Labeling Service Admin)
- AI Solution & Pipeline Engineer
- Long Term Support (LTS) Assurances
- Dedicated Success Manager
- Training
- Advanced queue customization
- Multiple workspaces
- Enterprise dashboards
- Audit & Security Reports
- Sales executive engagement
- Custom Commercial Pricing Models
- Service Level Objectives (SLO/SLA)
- Scale to billions and trillions of annotations, petabytes of Data
- Roadmap Steering Committee
In enterprise we will also work with your security team directly, do security audits, compliance audits, and assist in configuring your identity setup.
Engaging Community and Sales
In order to give you the most appropriate experience please kindly consider your intended use case:
Intended Use | Recommended Contact |
---|---|
Unsure | Community Slack Channel |
Open Source | Community Slack Channel Github |
Enterprise | Sales |
Diffgram.com Playground Limits
The free playground version that's hosted on diffgram.com is meant to try it out and has these limits:
- Up to 2 users.
- Up to 100 files per dataset.
- Video up to 100 frames.
After playing with it online:
- Install diffgram on your own hardware
- Contact us to talk about enterprise versions.
Support
Open Source | Enterprise | |
---|---|---|
Who is responsible? | You | Joint |
How can I communicate with the diffgram team? | General community channels. Response time may be greater then 1 day. No resolution promised, may help log ticket in git | Email, Dedicated slack channel, Video 2-3 business hours or next business day Video calls available |
Ticket expectation
For Community support, we see you as collaborator, and we may ask for a complete reproduceable example, specific technical error messages, and may even ask you to propose your own solution.
Contributor Notes: Open Source Code & Feature Flags
Developer Contributor Info
You may see some code on open source versions that might not be accessible or usable at run time. For example, there is code to help run diffgram.com and for enterprise teams.
-
Although all the enterprise code resides in the same repository, we purposefully limit some of our features with feature flags and licensing checks.
-
Although you might see the code of these feature, we do licensing checking on our licensing servers to see if the installations allow the usage of those features.
-
We strongly discourage any source code modifications to enable these features since they go against our terms of usage. Practically it will also make it difficult for you to receive updates if you do this and is against the law.
-
There are different limits for the diffgram.com free and open source self installed. As explained in more detail on this page, open source limits are only focused on the user limit, where as the diffgram.com version hosted by diffgram has dramatically different limits.
diffgram.com
(with the .com) is short handed for the version hosted by us.
open source
is short hand for diffgram software installed by you. (And not Enterprise).
Open Source Explained
See Open Source
Updated 5 months ago