← all repositories

chrisranderson/beholder

A TensorBoard plugin that visualizes neural network training by displaying tensors as video frames in real time.

463 stars Python ML FrameworksLLMOps · Eval
beholder
Velocity · 7d
+0.1
★ / day
Trend
steady
star history

Beholder is a TensorBoard plugin for visualizing arbitrary tensors during neural network training. It renders network parameters, gradients, convolutional layer activations, and other arrays as video frames directly within TensorBoard. The plugin integrates with the TensorFlow ecosystem and has been merged into the official TensorBoard project.

heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.