[RFC0109] All the inputs or inferences of Monlam AI Model need to be stored

[tenkus47]

RFC0109 All the inputs or inferences of Monlam AI Model need to be stored

Named Concepts

inference: input for model and Output data that are produced by the Monlam AI models.

Summary

here we need to store all the input text and raw base64 speech in database. for that we need button on MT for submitting and not debounce .

conditions: the inference data will be stored when

• whenever we get response from model api.
• whenever we submit the edited output data , edited should have a separate column in database.
• like and dislike will increase and decrease rating count accordingly .

Dependencies

we need a way to get model version , either document/hosted .

Infrastructures

Include all the infrastructure required for running the task, such as S3 bucket, EC2 server, etc.

Design Illustrations

Required column in database

Justification

Add justification for the strategies you have proposed above.

Why was the currently proposed design selected over alternatives?

What would be the impact of going with one of the alternative approaches?

Testing

Describe the kind of testing procedures that are needed as part of fulfilling this request.

Implementation Steps

List all the steps involved during implementation.

• [x] model schema design on prisma + migration
  Estimated time: 0.5 day Actual time:

• [x] like / dislike rating implement Estimated time: 0.5 day Actual time:

• [x] post API design for each models Estimated time: 2 day Actual time:

Reviewed By

@kaldan007

[tenkus47]

the data are stored on database for 3 models ,ocr is not complete yet.