deeo-learning

High performance, cross-platform machine learning for Unity Engine.

Improved StyleGAN Embedding: Where are the Good Latents?

Failure detection module for 3D printing using Computer Vision

Get different poses of the body in real time.

Realtime Face Mask Detection is the subject of the project proposal. We present a deep learning and computer vision-based mask face identification algorithm. As an object detection technique, deep learning was applied

Some Key Points from the Deep Learning Tuning Playbook

In silico model for predicting of Interleukin-6 inducing peptides

The Video/Audio Summarization Application transcribes and summarizes lengthy audio or video files, helping users quickly access key information. Using Wav2Vec 2.0 for accurate transcription and a summarization model, it provides concise, digestible summaries. With a user-friendly interface, it's suitable for both academic and professional use.

A clean, from-scratch PyTorch implementation of the original Transformer model for Neural Machine Translation, featuring optimized training and a web GUI for attention visualization.

User can draw a digit on a grid and then the model predicts the digit and renders the prediction | Scipy, Numpy, Scikit-Learn, TensorFlow, Keras, Pygame

From-scratch Neural Network framework in Python using NumPy. Implements dense layers, forward/backpropagation, gradient descent, ReLU/Tanh activations, MSE loss, and demonstrates learning on XOR and MNIST without TensorFlow or PyTorch.

A plant leaf disease identifier model based on a ResNet9 architecture

A web-based deep learning tool using MobileNetV2 to classify 7 types of skin diseases from the HAM10000 dataset.

DCNN: Deformable Convolutional Neural Network Implementation from scratch

Rust implementation of the Multiscreen neural language-model block. 🦀

Semi-Supervised Domain Adaptation (SSDA) with Correlation Alignment (CORAL). In this Tutorial we are using MNIST(Source Domain) and SVHN (Target Domain) datasets

Separable Convolution Layers

Deep learning without a framework: signal propagation and residual connections, convolution and a full CNN in NumPy, a character-level RNN, inside an LSTM, self-attention, autoencoders, transfer learning and augmentation, and adversarial examples. Includes a Keras CNN and ResNet50 example.

Human Activity Detection using Machine Learning algorithms from Smart-watch data.

Beginner-friendly notes and resources on Natural Language Processing (NLP), covering text preprocessing, vectorization, embeddings, models, and project pipeline.