VerifAI initiative to build open-source easy-to-deploy generative question-answering engine that can reference and verify answers for correctness (using posteriori model)
I have creating the first part for the frontend implementation, so far the Application is able to retrieve the documents and providing an answer (the verification part has not been inserted yet in backend).
From Fron-end side we have the application where we can set the parameter for the search, the answer is streaming token by token to guarantee an high quality user experience.
To launch the application from local side, you need just to go in front end folder:
client-gui/verifai-ui and run the command npm start
If from the client side the command npm start does not work, you will need to set $env:NODE_OPTIONS = "--openssl-legacy-provider" (you can just run it on terminal before to execute frontend part)
The frontend will be automatically connect to the backend on the Instance where the models are deployed, sometimes could not work because I am working on it and maybe I will need to stop the process sometimes.
I have creating the first part for the frontend implementation, so far the Application is able to retrieve the documents and providing an answer (the verification part has not been inserted yet in backend). From Fron-end side we have the application where we can set the parameter for the search, the answer is streaming token by token to guarantee an high quality user experience. To launch the application from local side, you need just to go in front end folder:
client-gui/verifai-uiand run the commandnpm startnpm startdoes not work, you will need to set$env:NODE_OPTIONS = "--openssl-legacy-provider"(you can just run it on terminal before to execute frontend part)