caretech-owl / gerd

Generating and evaluating relevant documentation (GERD)
https://towardsdatascience.com/running-llama-2-on-cpu-inference-for-document-q-a-3d636037a3d8
MIT License
4 stars 0 forks source link

[UC]: GEN: Assisted (continous) writing #25

Closed aleneum closed 2 months ago

aleneum commented 10 months ago

Summary

continue backend api and frontend; As a user I want to continue writing a previously generated raw text document. The process consists of text passage recommendations from an LLM and the user's refinement until the document is deemed complete.

Rationale

The gnerated document of the generation service must most likely be refined and extended. A 'continue' API interface could assist this process.

Level

user goal

Actors

User, GenService, Gradio

Preconditions

Postconditions

Basic Flow

  1. User adds or refines text
  2. User submits text
  3. System processed text and extends it with n tokens
  4. Text is returned to the frontend and shown to the user
  5. The text is deemed ready

Alternative Paths

No response

Visualisation

flowchart LR;
  1-->2
  2-->3
  3-->4
  4-->1
  4-->5

Other related issues, use cases, features

No response