tinynn-py 1.0.2

A tiny neural network library.

Tinynn

Tinynn is a lightweight neural network library built entirely in NumPy, designed to help developers understand the nitty-gritty details of neural networks. With its clean and concise code, Tinynn offers a great learning platform for those who want to deepen their understanding of the inner workings of neural networks.

Despite its simplicity, Tinynn has all the key features of a neural network library, including support for feedforward and recurrent networks, various activation functions, and common optimization algorithms. Whether you're a student, researcher, or machine learning enthusiast, Tinynn is an excellent tool for experimenting with neural network architectures and training algorithms.

Key Features

• Lightweight and easy-to-understand implementation
• Support for feedforward and recurrent networks
• Various activation functions to choose from
• Common optimization algorithms
• Built entirely in NumPy for efficient numerical computations

Installation

To install Tinynn, you can use pip:

pip install tinynn-py

Usage

Here's a basic example demonstrating how to create a feedforward neural network using Tinynn:

from tinynn.models import Sequential
from tinynn.layers import Dense
import numpy as np

model = Sequential()

X = np.array([[1,2,3],
     [5,4,3],
     [2,3,4]])
y = np.array([0, 0, 1])

model.add(Dense(3,64))
model.add(Dense(64,2,activation='softmax'))

model.compile_model(learning_rate=0.01, optimizer='adagrad') #Also available params: decay_rate,momentum(only for sgd) and optimizer = sgd,adagrad
model.fit(X,y,epochs=1000)

Contributions

Contributions to Tinynn are welcome! Whether you want to report a bug, suggest a new feature, or submit a pull request, please feel free to do so. Let's make Tinynn better together.

License

Tinynn is released under the MIT License. See the LICENSE file for more details.