You should know what's in your model. Train locally.

• Custom neural network implementation with:

  • Configurable layer architecture via JSON
  • Multiple activation functions (ReLU, Leaky ReLU, ELU, Sigmoid, Tanh, Softmax)
  • Various optimizers (SGD, Adam, AdaGuard)
  • Support for different loss functions (MSE, Categorical/Binary Cross-entropy)
  • L2 regularization
  • Early stopping
  • Network visualization tools

• Flask web interface for:

  • Model training
  • File uploads
  • Status monitoring

1. Clone the repository:

git clone https://github.com/gcox32/neuralnet.git

2. Install dependencies:

pip install -r requirements.txt

from src import DenseLayer, NeuralNetwork

# Load architecture from JSON file
with open('data/architecture.json', 'r') as file:
    architecture = json.load(file)

# Create neural network instance
net = NeuralNetwork(architecture=architecture, loss='categorical crossentropy', optimizer='sgd')

# Train the network
net.train(X_train, y_train, iterations=1000, learning_rate=0.01, decay=0.0, momentum=0.9)

# Make predictions
predictions = net.predict(X_test)

# Visualize the network
net.visualize_network()

# Visualize the validation accuracy
net.visualize_validation()  

# Visualize the predictions
net.visualize_predictions(X_test, y_test)

# Save the model
net.save('data/model.pkl')

# Load the model
net = NeuralNetwork.load('data/model.pkl')

yea, this is not even close to being done.

python app.py

The network supports:

• Activation Functions: ReLU, Leaky ReLU, ELU, Sigmoid, Tanh, Softmax
• Loss Functions: MSE, Categorical Cross-entropy, Binary Cross-entropy
• Optimizers: SGD, Adam, AdaGuard
• Training Modes: Batch, Mini-batch, Stochastic
• Regularization: L2 regularization
• Early Stopping: Based on validation loss

seems like a good idea.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

This project is licensed under the MIT License - see the LICENSE file for details.

Grant Cox - @Grant07549375

Project Link: https://github.com/gcox32/neuralnet

• How do I do specifically so and so?
  • No problem! Just do this.

Reach out to me at one of the following places!

• Website at letmedemo.com
• Twitter at @Grant07549375