Skip to content

aravindsairam/Pytorch_classification-CatsvsDogs

Repository files navigation

Pytorch_classification-CatsvsDogs

Classification of Cats and Dogs with Pytorch

This is the code for Kaggle's Dogs vs. Cats Redux competition.

Training dataset = 25,000 images and Testing dataset = 12,500 images

Network details:

  • CNN Archiecture : Pretrained Resnet-34
  • Learning rate : 0.0001 with Learning rate scheduler(Reduce on Plateau)
  • Optimizer : Adam
  • Batch size : 64
  • epochs : 10
  • augmentations : random horizontal flip, shift, scale and rotate.
  • Removes 51 wrong/incorrect images from train set which are not cats nor dogs (or both cats and dogs together).

Validation Loss : 0.095 (average of 5folds).

Achieved Public and Private score : 0.06599 Log Loss (mean ensemble of 5 models)

About

Classification of Cats and Dogs with Pytorch

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published