ParlAI Documentation

ParlAI is a one-stop-shop for dialog research.

Tutorials & Explanations

• ParlAI Quick-start
• Intro to ParlAI
• Tasks and Datasets in ParlAI
• Worlds, Sharing & Batching
• Using Torch Generator Agent
• Using Torch Ranker Agent
• Understanding and adding metrics
• Speeding up training
• Mutators
• Running crowdsourcing tasks
• Using Chat Services
• Tests in ParlAI

FAQ & Troubleshooting

• Frequently Asked Questions
• Tips and Tricks

Component Reference

• Standard Agents
• Tasks
• Mutators
• Model Zoo

Scripts Reference

• Command Line Usage
  • display_data
  • display_model
  • eval_model
  • interactive
  • safe_interactive
  • self_chat
  • train_model
• Advanced Scripts
  • build_candidates
  • build_dict
  • convert_to_json
  • convert_to_parlai
  • convo_render
  • data_stats
  • detect_offensive
  • eval_wordstat
  • extract_image_feature
  • interactive_web
  • multiprocessing_eval
  • multiprocessing_train
  • party
  • profile_interactive
  • profile_train
  • token_stats
  • torchscript
  • vacuum
  • verify_data
• Writing Your Own Script
  • Custom ParlAI Script
• Opt Presets
  • List of presets

API Reference

• parlai.chat_service
  • Chat Service Core
  • Services
  • Utilities
• parlai.core
  • parlai.core.agents
  • parlai.core.build_data
  • parlai.core.dict
  • parlai.core.loader
  • parlai.core.message
  • parlai.core.metrics
  • parlai.core.mutators
  • parlai.core.opt
  • parlai.core.params
  • parlai.core.script
  • parlai.core.teachers
  • parlai.core.torch_agent
  • parlai.core.worlds
• parlai.utils
  • parlai.utils.bpe
  • parlai.utils.conversations
  • parlai.utils.data
  • parlai.utils.distributed
  • parlai.utils.fp16
  • parlai.utils.logging
  • parlai.utils.misc
  • parlai.utils.pickle
  • parlai.utils.safety
  • parlai.utils.strings
  • parlai.utils.testing
  • parlai.utils.torch
  • parlai.utils.typing
  • parlai.utils.world_logging