Scribd
Upload
1 views2 pages

02 Technical Requirements-2

The document outlines the technical requirements and grading rubric for the MLOps course's computer vision project, emphasizing the need to integrate concepts from previous projects. It details submission guidelines, including the format for project documentation and necessary components like GitHub links and project descriptions. Additionally, it specifies requirements for model serving, load testing, and adherence to best practices in MLOps for grading purposes.

Uploaded by

aditya426690
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
1 views2 pages

02 Technical Requirements-2

The document outlines the technical requirements and grading rubric for the MLOps course's computer vision project, emphasizing the need to integrate concepts from previous projects. It details submission guidelines, including the format for project documentation and necessary components like GitHub links and project descriptions. Additionally, it specifies requirements for model serving, load testing, and adherence to best practices in MLOps for grading purposes.

Uploaded by

aditya426690
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

MLOps Course: Computer Vision Projects Part 2

This document provides more details about the technical requirements for the computer
vision project.

Caveat Emptor
• All the projects in the course (SR, CV, NLP) will be cumulaDvely comprehensive. In
other words, you will be using all the concepts you have learned so far in the course in
each of the project. So, in the final NLP project, you will be using everything you
learned in the course. In other words, everything in this document will also be needed
for the second and third project.
• So please do a good job so that you can easily copy/paste it for your future course and
non-course projects :)

Grading Rubric
• Deadline: March 20 23:59
• The project will be graded along three dimensions.
• (40 points) FuncDon requirements + Demo
o The first check is about whether you have done a competent job in
implemenDng the funcDonal requirements.
o You will be doing a demo to the class. Do NOT prepare any slides. It is just a 5-
minute demo of the project. Once you do the demo, a random student (or the
instructor) will test your system.
• (40 points) Technical requirements
o This document will give lot of comments on how to implement the project.
o How well you implement these features will determine the next 40 points.
• (20 points) Miscellaneous requirements
o The instructor will also go over the code line-by-line and evaluate the project
based on aestheDcs.
o In general, MLOps is all about how to make a project more durable, scalable
and easy to use. So, the last 20 points will measure how well your project
adheres to these ideals beyond what is asked in the technical requirements.
o The instructor will also give detailed feedback that you can use for other
projects in this course.

Submission Guidelines
• The submission should be a single PDF file containing the following informaDon
o Project name
o Project members: Student ids, names and email ids
o GitHub Link to your project
o Brief 1-2 para descripDon of your project.
• Your code should be self-contained and easy to run. Specifically, I should be able to run
it with the following three commands. No other commands are allowed :)
o git clone <github url>
o just setup # this should do the setup such as installing relevant libraries
o just run # this should start the project
• Please share the project with the GitHub Id saravanan-thirumuruganathan.

Requirements from Speech RecogniCon Project


Your computer vision project has to saDsfy the following requirements from the speech
recogniDon project. Please refer to that document for more details.

• Git requirements
• Unified logging
• Support for two clients: GUI and API based
• CustomizaDon using config files (what to configure is up to you). Simplest will be to
configure FastAPI service.
• Data validaDon with PydanDc
• Setup with uv and Just
• Project documentaDon with mkdocs (see load tesDng below)
• Ruff code linDng

Model Serving Requirements


You have to serve your model using two methods: one using FastAPI and one using model
serving infrastructure. Your project should have a drop-down that determines which of these
two API sources will be called. You can assume that this value will not change during the
program execuDon. In other words, I can set this value at the start and will not change it later.
You can use any model serving infra that we will discuss in the class (BentoML, Torch Serve,
and Ray Serve).

Load TesCng Requirements


You should do load tesDng with Locust in a realisDc manner and record the results in the
project documentaDon.

You might also like