Fine-Tuning-VGG-16

Fine Tuned For CIFAR-100

Purpose Of This Project

• To learn and Hands On about Pre-Tranied Models.
• To Learn About Pytorch 🔦

Difficulty Faced

• As Vgg-16 is a very Deep Neural Network Model Training even 1 Epoch in Google Colab Takes a Long Time even with T4 GPU.
• Can't Train the Model after 1 epoch as Google Colab was Frequently Shutting Down Operation

Pre-Trained Model Variations

1. Froze Convulation Layers Weights Update
2. Froze Fully Connected Layers Weights Update
3. Developed Model From Scratch
4. Trained Using Pre-trained Model Weights

Dataset

https://www.kaggle.com/datasets/fedesoriano/cifar100

Clone The Repo

https://github.com/secrakib/Fine-Tuning-VGG-16.git

Licenses

Apache License 2.0