Welcome to pykoi!

pykoi (/py’koi/) is an open-source python library for ML scientists and practitioners. Pykoi lets you easily collect and understand user data and feedback, train RLHF on that data, and compare different models to each other.

Getting Started

• Installation
• Quick Tour

Features

• Collecting User Feedback via a Chat Interface
  • Sharing Your Model
  • Feedback options
  • Usage Dashboard
• Comparing Models
• Reinforcement Learning with Human Feedback
  • Step 1: Supervised Finetuning
  • Step 2: Reward Finetuning
  • Step 3: Reinforcement Learning
• Data Labeling & Annotation
• Finetuning & Knowledge Retrieval
• Observability
• Vista

Code

• pykoi
  • pykoi package
    • Subpackages
      • pykoi.chat package
      • pykoi.rlhf package
      • pykoi.component package
      • pykoi.interactives package
      • pykoi.telemetry package
      • tests package
    • Submodules
    • pykoi.application module
      • Application
      • ComparatorInsertRequest
      • InferenceRankingTable
      • ModelAnswer
      • RankingTableUpdate
      • UpdateQATable
      • UserInDB
    • pykoi.state module
      • State
      • Store
    • Module contents

Social

• Community

Indices and tables

• Index

• Module Index

• Search Page