For AI agents: visit https://documentation.picsellia.com/llms.txt for an index of all pages formatted in Markdown and endpoints in OpenAPI.
Jump to Content
User Guides
User Guides
Recipes
Python SDK Reference
Discussions
User Guides
Python SDK Reference
User Guides
Python SDK Reference
User Guides
Recipes
Python SDK Reference
Discussions
📘 Model Management
Introduction
🚀 Welcome to Picsellia's documentation
💡 Why Picsellia ?
💬 Glossary
DOCUMENTATION
📑 General Information & Administration
Access Picsellia
Homepage
Setup your profile
Python SDK installation
Picsellia Organizations
Picsellia platform structure
Track operations with Jobs
Track your plan and usage
Users Management System
📚 Data Management
Datalake - Philosophy and infrastructure
Datalake - Upload Data from local drive
Datalake - Import Data from your Cloud-based Object Storage
Datalake - Overview
Datalake - Tagging system
Datalake - Setup your own Metadata
Datalake - Projections
Datalake - Explore your Data with Embeddings
Datalake - Query Language
Datalake - Processings
Datalake - Dataset creation
Datalake - Summary & Workflow
Dataset - Dataset versioning system
Dataset - Assets overview
Dataset - Tagging system
Dataset - Set up the dataset
Dataset - Annotation import and export
Dataset - Annotation studio
Dataset - Annotation tool
Dataset - Annotation Campaign
Dataset - Processings
Dataset - Query language
DatasetVersion - Explore your Assets and Shapes with Embeddings
Dataset - Fork a Dataset Version
Dataset - Custom Shapes import from another version
🔬 Data Science
Project - Creation & collaboration
Experiment - Creation
Experiment - Experiment overview
Experiment - Launch training
Experiment - Experiment Tracking
Experiment - Evaluation
Experiment - Comparison
Experiment - Export as a model
🚀 Model Operations
Registry - Model versionning system
Registry - ModelVersion
Registry - Create a ModelVersion manually
Registry - Processings
Registry - Deploy a ModelVersion
Deployments - List of deployments
Deployments - Overview
Deployments - Dashboard metrics
Deployments - Configure and use a Pipeline
Deployments - Setup Alerts
Deployments - Inferences
Deployments - Predictions overview
Deployments - Prediction Review tool
Deployments - Prediction Review Campaign
Deployments - Prediction Review Tool
Deployments - Shadow deployment
Tutorials
⏭️ Quick Start Guide
1. Set up an account, an Organization and the Python SDK
2. Access to the documentation
3. What do you want to achieve with Picsellia ?
4. Import your Data in the Datalake
5. Create your first Dataset
6. Import your Annotations
7. Inherit a new DatasetVersion
8. Create your Project and launch Experiments
9. Deploy a ModelVersion and set the Data Pipeline up
10. Make predictions
11. Conclusion
🏁 Community Edition - Start using Picsellia !
1. Join Picsellia Community
2. Dashboard overview
3. Import your Data in the Datalake
4. Create your first Dataset
5. Create Annotations
6. Inherit a new DatasetVersion
7. Conclusion
📦 Migrate your models to Picsellia
1. Overview
2. Create your first Model
3. Make your ModelVersion Deployable
4. Make your ModelVersion Trainable
5. What's next?
🧩 Integrate Picsellia into your training scripts
1. Overview
2. Initializing Picsellia connection & retrieve the Experiment
3. Retrieve Data, files and parameters from Picsellia
4. Track your trainings with Callbacks
5. Test your ModelVersion with Picsellia Evaluation Interface
6. Store new files
Final script
🔎 Evaluate your Model Performances
1. What's an Evaluation on Picsellia
2. Evaluate a Classification Model
3. Evaluate a Detection Model
4. Evaluate a Segmentation Model
📈 Monitoring an External Model
1. Pre-requisite
2. Classification - Monitor Model Predictions
3. Object Detection - Monitor Model Predictions
4. Segmentation - Monitor Model Predictions
🤖 Create your own Processing
1. Create a Private Processing
2. Write the Processing script
3. Configuration reference
🐠 Configure Picsellia webhooks
1. What's a Webhook?
2. Events
3. How to configure Webhooks on Picsellia
☁️ Connect Picsellia to your own Cloud storage
1. Configure your bucket
1.1. AWS - Create a delegated Access for Picsellia
1.2. GCS - Create access credentials
1.3. Azure - Register a client application
2. Datalake mechanism
3. Configure your new Datalake
🔄 Use MLFlow with Picsellia
In depth explanation
Custom Objects
Namespace
Computed Metrics
AE Outlier
KS Drift (Kolmogorov-Smirnov Test)
Tide Errors (Object detection)
Powered by
📘 Model Management
Copy Page
