FeatureVisualizer-pytorch

An implementation of feature visualizer for pytorch tensor.

Prerequisites

• python >= 3.6

• pytorch

• matplotlib

• numpy

Usage

# Initialization
FV = FeatureVisualizer()

# Load data
image = load_image(image_path)
feat = net(image)

# Visualize features
FV.viz(feat, save_path=feat_path)

# Mix images and features
FV.mix(image_path, feat_path, save_path=mix_path)

Demo

python demo.py