Model performance and tuning analysis conducted on the CIFAR10 and CIFAR100 datasets. Convolutional Neural Network (CNN), Gated Multilayer Perceptron (gMLP), and Vision Transformer (ViT) model architectures are utilized. The study is built using PyTorch, PyTorch Lightning for clean and concise code and Optuna for hyperparameter tuning.

Simplified PyTorch implementation of image classification, support multi-gpu training and validating, automatic mixed precision training etc.

A PyTorch implementation of "EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks".

Robust Transformer with Locality Inductive Bias and Feature Normalization (JESTECH 2023)

This project optimizes the CIFAR-10 dataset for improved model performance through data exploration, augmentation, and training a CNN. It includes data loading, preprocessing, exploratory data analysis (EDA), and model training in a streamlined pipeline, showcasing the importance of data preparation in achieving better classification accuracy.

This repository includes official implementation and model weights of Data-Efficient Multi-Scale Fusion Vision Transformer.

Deep Learning Project on Diffusion Models for Image Generation

This repository contains a Python-based image recognition project using TensorFlow and Keras. It leverages a pre-trained Convolutional Neural Network (CNN) model on the CIFAR-10 dataset to classify objects in images. The project supports image recognition from both local files and URLs.

Image transformations and deep learning using PyTorch

Self-training variants using PyTorch

Classification with PyTorch.

Super-Convergence on CIFAR10

Study of the impact of a Randomized ReLU function on the performance of a neural network model.

Implementation of Federated Learning on a simple CNN architecture on CIFAR10 dataset.

A lightweight and extensible toolbox for image classification

Official PyTorch Implementation for the "Distilling Datasets Into Less Than One Image" paper.

Quick guide to submitting and monitoring jobs on Ozerlabs slurm cluster.

A coding-free framework built on PyTorch for reproducible deep learning studies. 🏆25 knowledge distillation methods presented at CVPR, ICLR, ECCV, NeurIPS, ICCV, etc are implemented so far. 🎁 Trained models, training logs and configurations are available for ensuring the reproducibiliy and benchmark.

This is the official repository of L-DAWA: Layer-wise Divergence Aware Weight Aggregation in Federated Self-Supervised Visual Representation Learning github

use CNN to fit cifar10,but overfitting