Kernel size was 7 and stride 2. Accuracy was considerably low because a 7x7 is not very suitable for Resnet models.
Kernel size made to 3x3 and an acceptable accuracy is achieved.
-
Notifications
You must be signed in to change notification settings - Fork 0
Comparison between a pre-trained ResNet18 model ImageNet tested gradually increased dataset on dataset CIFAR100 and ResNet50 model rtrained on CIFAR100 dataset. While maintaining an acceptable accuracy, a huge time decrease on training time is achieved due to smaller, but distilled dataset that allows model to be general enough on unseen data.
RhinoCoder/ResNet50-vs-ResNet18-Gradual-Increase
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
About
Comparison between a pre-trained ResNet18 model ImageNet tested gradually increased dataset on dataset CIFAR100 and ResNet50 model rtrained on CIFAR100 dataset. While maintaining an acceptable accuracy, a huge time decrease on training time is achieved due to smaller, but distilled dataset that allows model to be general enough on unseen data.
Topics
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published