This is the official implementation of our ICML 2024 paper "MultiMax: Sparse and Multi-Modal Attention Learning""
-
Updated
Nov 6, 2026 - Python
#
activation-function-exploration
Here are 43 public repositories matching this topic...
A structured, hands-on guide to neural networks and deep learning fundamentals, covering forward propagation, activation functions, backpropagation, optimization, TensorFlow, Keras, PyTorch, and CIFAR-10 image classification.
deep-learning tensorflow optimization keras pytorch neural-networks gradient-descent keras-neural-networks cifar10 loss-functions backward-propagation forward-propagation activation-function-exploration
-
Updated
Sep 29, 2026 - Python
An Artificial Neural Network (ANN) built using PyTorch to classify images in the Fashion MNIST dataset, utilizing GPU support for improved training performance.
optimizer pytorch classification ann backpropagation loss-functions forwardpropagation activation-function-exploration
-
Updated
May 26, 2026 - Jupyter Notebook
Open ideas. Code less. Publish more.
frequency machine-learning computer-vision deep-learning neural-network pytorch artificial-intelligence fusion yolo attention sampling sequence convolutional-neural-network normalization modular-design paper-implementation detr ultralytics yolo11 activation-function-exploration
-
Updated
Feb 2, 2026 - Python
This repository contains the official implementation of the paper: "DAPA: Distribution Aware Piecewise Activation Functions for On-Device Transformer Inference and Training" (Accepted at DAC 2026).
-
Updated
Jul 23, 2026 - Python
This repository features hands-on Jupyter Notebooks, covering everything from fundamental concepts to advanced neural network architectures.
python machine-learning deep-learning neural-networks from-scratch neural-networks-from-scratch activation-function-exploration
-
Updated
May 8, 2025 - Jupyter Notebook
By Ikansh Mahajan, 3C53, 102303754, COE'27
deep-learning mlp backpropagation-learning-algorithm optimization-algorithms mlp-regressor mlp-classifier activation-function-exploration
-
Updated
May 2, 2026 - Jupyter Notebook
NN from scatch, self-coded
-
Updated
May 30, 2024 - Python
Interactive deep learning project to predict IPL innings scores from match context such as teams, venue, batsman, bowler, runs, wickets, and overs.
-
Updated
Mar 19, 2026 - Jupyter Notebook
Just Fling It! This is a custom activation function known as Fling, for deep learning networks. Read more about it below.
-
Updated
Nov 11, 2026 - Python
A mini supervised learning framework
machine-learning framework model machine-learning-algorithms optimization-algorithms loss-functions activation-function-exploration
-
Updated
Sep 17, 2026 - Python
A fast, smooth, algebraic alternative to Mish for TinyML and Edge AI. LiteMish uses only addition and multiplication, resulting in improved computational efficiency while preserving learning performance.
machine-learning deep-learning embedded-systems neural-networks tiny-dnn neural-control neural-network-inference inference-optimization embedded-ai edge-ai mish inference-acceleration tiny-ml tiny-machine-learnig real-time-ai efficient-ai energy-efficient-ai neural-network-controller activation-function-exploration tiny-deep-learning
-
Updated
Oct 4, 2026 - Jupyter Notebook
-
Updated
Dec 28, 2025 - Jupyter Notebook
We introduce two novel hybrid activation functions: S3 (Sigmoid-Softsign) and its improved version S4 (Smoothed S3)
python science machine-learning research deep-learning neural-networks sigmoid softsign activation-function-exploration
-
Updated
Oct 31, 2026 - Python
MITSU (三つ巴): a parameter-free neural-network activation, s(s−1)/(s+2) with s = softplus(u), derived from a symmetric three-variable function. NumPy and PyTorch implementations, MitsuGLU, an interactive lab, and pre-registered small-model experiments against GELU, SiLU and SwiGLU.
benchmarking machine-learning ai deep-learning experimental neural-network reproducible-research numpy transformers pytorch neural-networks experimental-research activation-function-exploration
-
Updated
Nov 13, 2026 - Python
This is a Fake and AI image prediction using Transfer Learning
-
Updated
Apr 21, 2025 - Python
Un petit projet de classe réalisé en guise de TP. Exploration de la librairie PyTorch pour le Deep Learning.
deep-learning neural-network pytorch basis backpropagation class-project forward-propagation fully-connected-neural-network backpropagation-neural-network activation-function-exploration
-
Updated
Nov 12, 2026 - Jupyter Notebook
This is a custom-built neural network that detects handwritten numbers from image inputs. It uses ReLU activation in the hidden layers and a softmax activation function in the output layer for classification. The model is trained using backpropagation with a loss function to minimize prediction errors, achieving over 99% accuracy when predicting
-
Updated
Apr 11, 2025 - Python
A production-ready Deep Learning project demonstrating an Artificial Neural Network (ANN) trained on the classic Iris dataset, served via a high-performance FastAPI backend with a modern, responsive animated web UI.
deep-learning neural-network models optimizer perceptron ann loss-functions dense tensorflow-keras activation-function-exploration
-
Updated
Sep 2, 2026 - Jupyter Notebook
Linear and Non Linear Activation Functions : Linear, ReLU, Sigmoid, Softmax, Tanh
-
Updated
Jul 16, 2025 - Jupyter Notebook
Add this topic to your repo
To associate your repository with the activation-function-exploration topic, visit your repo's landing page and select "manage topics."