https://github.com/deshwalmahesh/ml-models-from-scratch

Repo for ML Models built from scratch such as Self-Attention, Linear +Logistic Regression, PCA, LDA. CNN, LSTM, Neural Networks using Numpy only
https://github.com/deshwalmahesh/ml-models-from-scratch

base cnn cnn-numpy lstm-numpy ml model numpy scratch scratch-implementation self-attention

Last synced: about 1 month ago
JSON representation

Repo for ML Models built from scratch such as Self-Attention, Linear +Logistic Regression, PCA, LDA. CNN, LSTM, Neural Networks using Numpy only

• Host: GitHub
• URL: https://github.com/deshwalmahesh/ml-models-from-scratch
• Owner: deshwalmahesh
• Created: 2021-02-28T09:34:30.000Z (almost 4 years ago)
• Default Branch: main
• Last Pushed: 2024-11-17T05:15:25.000Z (about 1 month ago)
• Last Synced: 2024-11-17T05:19:03.211Z (about 1 month ago)
• Topics: base, cnn, cnn-numpy, lstm-numpy, ml, model, numpy, scratch, scratch-implementation, self-attention
• Language: Jupyter Notebook
• Homepage:
• Size: 38.4 MB
• Stars: 28
• Watchers: 2
• Forks: 4
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# ML-Models-from-Scratch

Repo for ML Models  built from scratch. Includes 

1. Data Structures in Python

2. Traditional ML:

   - Linear Regression

   - Logistic Regression

   - SVM

   - PCA

   - LDA

3. Deep Learning:

   - Loss with Backprop

   - Activations with BackProp

   - Layers: Dense (Linear), CNN, LSTM, Attention, MultiHead Attention

4. Metrics

5. Architectures:

   - Transformers

   - Encoder-Decoder

   - LSTM

   - DeepLab

   - UNET

   - YOLO (need to finish)

   - MAMBA (WIP)

This code contains the traditional and simpole ML models built using `numpy`, `pandas` and `python 3`. Once in a while `pytorch` is also used for some tasks. There might be some issues but will be solved over time. Not the best of the models but gives good understanding of the problems and helps in deep understanding of the concepts.