mml-pypi 0.0.4.2

My Machine Learning (MML) Library. A hybrid backend (numpy or torch) machine learning and deep learning framework coding from scratch. Write `import mml` to use it.

MML (mml-pypi)

My Machine Learning (MML) Library (developed byNathmath/DOF Studio), identified as mml-pypi on PyPI.

A hybrid backend (numpy or torch) machine learning and deep learning framework coding from scratch.

It is a fantastic toolkit for machine learning teaching, learning, quick application with production level performance.

Another shining feature is its AutoNeuralNetwork framework - automatically build, train, fine-tune, and validate a neural network especially designed for non-professional groups.

How to Install

pip install mml-pypi==0.0.4.2

import mml

Version

MML 0.0.4.2 Released.

License

Open sourced with Apache 2.0 License

What's Inside?

What's inside? See below.

Containers using Mixed Backends

• Matrix (For ML Algorithms) (numpy √ torch √)
• Tensor (For NN Framework) (numpy √ torch √)

ML Algorithms from Scratch

• Linear Models (OLS and FGLS)
• Generalized Linear Models (FGLS with Actvation)
• Time Series Models (TS)
• Principal Component Analysis (PCA)
• Support Vector Machine (SVM)
• Classification And Regression Tree (CART)
• Linear Regression Tree Wrapper (LRTW)
• Random Forest (RF)
• Gradient Boosting Machine (GBM)
• Extreme Gradient Boosting Machine (XGBM)
• ...

Neural Network Framework from Scratch

• Basic Module (nn_Module)
• Dense Layer (Dense)
• Dropout Layer (Dropout)
• Flatten Layer (Flatten)
• Stacked RNN Layer (StackedRNN)
• Stacked LSTM Layer (StackedLSTM)
• Loss Functions (MSE, RMSE, MAE, BinaryCrossEntropy, MultiCrossEntropy)
• Optimizers (SGD, Adam, AdamW)
• Easy Interface for Evaluation (Evaluator)
• ...

Utils from Scratch

• Regression, Binary Classification, Multi Classification Metrics
• Train-Test Split, Train-Test Split for Time Series
• Data Scaler
• Data Wrangling Toolkits
• Easy Save and Load Interface
• Generic Optimizer
• Generic Bootstrap Sampler
• ..