Skip to content

Image Classification Model in PyTorch for Indoor Scene Recognition

License

Notifications You must be signed in to change notification settings

ashrutkumar/Indoor-scene-recognition

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 

Repository files navigation

indoor-scene-recognition

Image Classification Model in PyTorch for Indoor Scene Recognition.

This is a Google Colab Notebook. The notebook can also be found here : https://colab.research.google.com/drive/1ALGzKj3cafL0MPUNtbUNXzRc8S_C_lqM

The MIT-67 Indoor Scene Recognition Dataset is used here. The dataset has 15620 images in total distributed amongst 67 classes consiting of airport,trainstation,kitchen,library,etc. Link: http://web.mit.edu/torralba/www/indoor.html

The model used is Resnext101_32x16d. Link: https://github.com/facebookresearch/ResNeXt .

Transfer Learning was used in this task. Pre-trained weights of the model on the ImageNet dataset was used.
Along with transfer learning, data-augmentation, learning rate annealing, early stopping,etc. were also used in the training process.

This model was originally created for the kaggle challenge ( https://www.kaggle.com/c/qstp-deep-learning-2019 ) . The model acheived 3rd rank in the contest amongst 30 participants(Top 1 %).

The notebook contains the link to the weights of the final model, so that it can directly be used without training all over.

About

Image Classification Model in PyTorch for Indoor Scene Recognition

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published