The CIFAR-10 dataset classification

Description

In this repo, some basic models of deep learning were compared in the classification:

The CIFAR-10 dataset.

The models are:

Model	Number of Parameters	Best accuracy (%)
Simple CNN model	62k	62
More complex CNN model with drop-out	1.1M	79
ResNet18	11.1M	85
VGG11	28.4M	86

Models architecture was coded in models file.

The training dataset was augmented with transform_pipeline

Examples of transfromations are:

You can read about the project and it's results in:

main file(main.ipynb)

Set up and configuration

You can download the repository via the command

git clone https://github.com/DzmitryPihulski/LLM_question_and_answer_system_with_RAG.git

I used python version 3.11.4.