HackerEarth Deep Learning Challenge: 'Tis STILL the season to be jolly

Problem statement

You work for a social media platform. Your task is to create a solution using deep learning to discern whether a post is holiday-related in an effort to better monetize the platform.

Task

You are given the following six categories. You are required to classify the images in the dataset based on these categories.

• Miscellaneous
• Christmas_Tree
• Jacket
• Candle
• Airplane
• Snowman

Data Description

This data set consists of the following two columns:

Data Format

The data folder consists of two folders and one .csv file. The details are as follows: train: Contains 6469 images for 6 classes test: Contains 3489 images train.csv: 3489 x 2

Submission Format

You are required to write your predictions in a .csv file and upload it by clicking the Upload File button.

Sample Submission

Image,Class image3476.jpg,Miscellaneous image5198.jpg,Candle image4183.jpg,Snowman image1806.jpg,Miscellaneous image7831.jpg,Miscellaneous

Evaluation Matrice

Acknowledgement

Link To Contest --> https://www.hackerearth.com/challenges/competitive/hackerearth-deep-learning-challenge-holidays/

---

My Solutions (Rank: 120, Top 3%)

1. Attempt 1
   • SCORE: 88.90237
   • Used transfer learning in Keras and fine-tuned an InceptionResNetV2 model
2. Attempt 2
   • SCORE: 90.94382
   • Used transfer learning in Keras and fine-tuned an InceptionV3 model