danielkornev
commented
3 years ago High-Level Goals (for Improvement)
Pri 0
Learn
Library
How Do I - individual components
- How Do I (with Google Colab Tutorials) (Run in Colab | View Notebook)
How Do I - advanced scenarios
- Multiskill AI Assistant
- Complex NLP Pipeline (e.g. like for Pro//Chtenie)
Agent
- Go-Bot Skill Builder
- FAQ Skill Builder
- 3rd Party Skills
- Integrations
Dream
- Provided Skills
API
Library
- Reference (+ old docs before migration)
Agent
- Reference
Goals
- Emphasis on How-Do-Is (many How-Do-Is)/Learn
- Emphasis on Google Colab Tutorials (Paragraph Level) (where applicable)
- Emphasis on High-Performance Computing (once available)
- Move Integration components into separate areas: integrating thirdparty skills (e.g., RASA bots), integrating with different systems (e.g., Google Assistant)
- Move Deployment into a separate area
Pri 1
- Better distinction between supported languages (spoken)
- Maybe put models into a separate place and make them updateable (like HF Transformers?)
Pri 2
- Put demos into the docs (like in RASA and AllenNLP docs)
- Editable docs using markdown
- Docs (articles) separate from API ref (so that we could edit articles independent of Library releases, yet could see all of those docs in the same place)
Moved to internal Trello