New Baseline with [0.71 on LB] - One Small Change

Hello :wave: Everyone,

Since supervised learning is a small part of the challenge and the main focus is to work on the “Purchasing phase”. I thought maybe now everyone should make their baseline strong to advance in the competition.
I have curated a notebook that will help you achieve 71% accuracy on LeaderBoard, making it a new baseline for your experiments, HOW COOL IS IT, right? :sunglasses: :sunglasses: :sunglasses:

In this notebook you will learn the following things:-

  1. A Secret :face_with_peeking_eye: to make a new baseline with 0.71 accuracy :scream:
  2. Tracking the progress on WandB

Notebook: AIcrowd | A New Baseline With 0.71 accuracy on LB | Posts

If you need help with an end-to-end solution, please check out the additional resources which will help you to:

  1. Understanding :face_with_monocle: the problem statement
  2. Knowing what data do you have?
  3. Understanding the starter kit codebase :file_folder: and what each file does
  4. Forking/Importing the starter kit to your GitLab account
  5. Setting up SSH key :old_key: on your system
  6. Setting up local development python environment/sandbox :hammer_and_pick:
  7. Write your first PyTorch ResNet model (CUDA and CPU compatible) in run.py class
  8. Learning WandB (A MLOps tool) and how to setup
  9. Making a submission :rocket: :rocket: :rocket: :rocket:

Additional resources:

  1. Learn How To Make Your First Submission [🎥 Tutorial]
  2. Learn How To Make First Baseline Model With 0.44+ Accuracy on LeaderBoard [🎥 Tutorial]
  3. Learn How To Setup MLOps (WandB) Pipeline To Track Your Experiments In 15 Minutes

You are just 30 minutes away from making your new successful baseline submission which will put you in a strong and confident position… :tada: :tada::tada: :tada:
Please do let me know any improvements or questions in the comments below, I would be glad to help you.
Click on :heart: if this post was of any help

6 Likes