Skip to content

Yyuxin0822/LLM_EcoCircuit

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

LLM_EcoCircuit (In Development)

EcoCircuit is an LLM-based web application for generative flow diagrams. As an educative tool for environmental engineering, it generates flow / circular economy diagram and allows freedom for users to doodle and modify.

This realease realizes all the below functions of Sherbot. The interactivity of the web will be improved in the next release

Init

  1. Activate virtual environment and python dependencies
  2. cd "static", install node modules in package-lock, this project uses leader-line and plain-draggable package by anseki
  3. The webapp is powered by gpt-4 and dalle2, you can adapt the code to the model you selected accordingly
  4. Run python app.py, the webpage will open

GENERATE FLOW

  1. "GENERATE FLOW" is the button to send new TEXT-TO-FLOW DIAGRAM request.
  2. First, Input your environmental description here, click "GENERATE FLOW"

text-to-flow

  1. The web will generate a compacted view of input-to-output flow.

text-to-flow

text-to-flow

text-to-flow

text-to-flow

text-to-flow

  1. You can toggle image on and off by "Explore > Show Creative Image" and "Explore > Hide Creative Image"

text-to-flow

ASK SHERBOT

  1. "ASK SHERBOT" is the button to send new request based on the element selections.
Trial 1
  1. Try, right-click to select one element, such as "OXYGEN"

text-to-flow

  1. Then, click "ASK SHERBOT"

text-to-flow

  1. Now, go back to by clicking "Explore > Lookup Simple Flow"
Trial 2
  1. Try, right-click to select one flow, such as right-click "OXYGEN" --> select "Select Output Flow"--> "FUEL SOURCE"

text-to-flow

  1. Then, click "ASK SHERBOT"
  2. The tool will expand on the internal process for generating flow This process can be revisited by click "Explore > Lookup Complex Flow"

text-to-flow

  1. Now, go back to by clicking "Explore > Lookup Simple Flow"
Trial 3
  1. Try, right-click on multiple element and multiple flow, like below

text-to-flow

  1. Then, click "ASK SHERBOT"
  2. It will generate a complex flow view, delinating multiple processes for going from input to output

text-to-flow

  1. Navigate to "Explore > Lookup Element Co-optimization" It will generate a networked flow view, delinating how all the elements can be connected

text-to-flow

  1. Now, go back to by clicking "Explore > Lookup Simple Flow"
Trial 4
  1. Select all element and flow

text-to-flow

  1. Then, click "ASK SHERBOT"
  2. Then, click "Explore > Lookup Complex Flow"

text-to-flow

  1. Then, click "Explore > Element Co-optimization"

text-to-flow

As ChatGPT will never generate the same result, this is another result generated with this prompt and process

text-to-flow

This tool contains many features for human customization

  1. Including Regenerate Image / Drag Image / Draw / Edit Image Size / Show Notes / Modify input and output / Add New Tag

Below are some examples.

text-to-flow

text-to-flow

text-to-flow

text-to-flow

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages