Resnet18-on-CIFAR10

Implementing Transfer Learning using Resnet18. Training and testing Resnet18 on Even numbered Classes in CIFAR10 Dataset.

Language used: Python

Libraries used: torch, torchvision, matplotlib, seaborn

Dataset: CIFAR10 dataset

Dataset Visualization:

Observations:

1. Train-Test Loss and Accuracy:

Epochs	Train Loss	Train Accuracy	Test Accuracy
1	0.360	86.067%	84.700%
2	0.224	91.647%	92.800%
3	0.184	93.267%	93.067%
4	0.172	93.860%	93.033%
5	0.137	94.807%	95.000%

2. Train-Test Accuracy Plot:

3. Confusion Matrix: