[Project] Stanford-Cars with fastai v1

Deep Learning
43

94.79% with EfficientNet-b7 + Mish + Ranger

  • Achieved 94.79% (standard deviation of 0.094) 5-run, 40epoch, mean test set accuracy on Stanford Cars using Mish EfficientNet-b7 + Ranger (no mixup)
  • Matched the EfficientNet paper EfficientNet-b7 result of 94.7% (current SOTA is: 94.8% 96.2% from Domain Adaptive Transfer Learning with Specialist Models)

17
Screenshot 2019-10-02 10.17.17.png1562×1138 88 KB

Code here: https://github.com/morganmcg1/stanford-cars

4 Likes