ResNet-3 Image Classification

PyTorch CNN with residual connections for image classification (CIFAR-10 dataset).

A ResNet-3-style CNN built in PyTorch for image classification (CIFAR-10 dataset).

It demonstrates residual connections, a clean training loop, and a minimal project you can run end-to-end on a laptop/GPU.


Methodology

Architecture diagram of ResNet-style backbone (3 residual stages).

Key pieces

  • Residual blocks: 3-stage stack.
  • Training loop: SGD/Adam; optional LR scheduler.
  • Evaluation: Accuracy and loss.

We used the Standard Cross Entropy loss:

\[\mathcal{L}_{\mathrm{CE}} = -\frac{1}{N}\sum_{n=1}^{N}\sum_{c=1}^{C} y_{n,c}\,\log p_{n,c}, \quad \text{where} \quad p_{n,c}=\frac{e^{z_{n,c}}}{\sum_{k=1}^{C} e^{z_{n,k}}}. \tag{1}\label{eq:ce}\]

Results

Accuracy Plot.

Repository

GitHub Repository – Cifar10-ResNet3-Classifier

How to run:
Create env, install torch + torchvision as usual, then:

python proj1.py