pytorch-visualizer 0.1.1

A neural network visualization engine, built for pytorch

Neural Network Visualization Tool

This tool is designed to provide a dynamic visualization of neural network training. It integrates with a PyTorch model's forward pass as a decorator, allowing you to visualize how neural network training occurs.

Features:

• Supported Loss Functions:

  1. NLLLoss
  2. CrossEntropyLoss

• Layer Compatibility:

  • Currently supports fully connected (dense) layers.

Interactive Visualization:

• View the gradient, bias, and weight histories for both nodes and lines.
• Click on any node or line to inspect its parameters.

Installation

pip install pytorch_visualizer

How the API is used

Below is an example showing how the API can be used:

from python_visualizer.nn_visualizer import visualize
class SimpleNN(nn.Module):
    def __init__(self):
        super(SimpleNN, self).__init__()
        self.fc1 = nn.Linear(2, 5) 
        self.fc2 = nn.Linear(5, 3)
        self.relu = nn.ReLU()
    
    @visualize(epochs=5, labels=None, loss_func=None, backward=False)
    def forward(self, x):
        x = self.fc1(x) 
        x = self.relu(x)
        x = self.fc2(x) 
        return x

check demo.ipynb for a full implementation with a training loop

License

MIT