Tensorboard

Solid wrapper around Google's TensorBoard for tracking ML experiments directly from Claude. Covers the essentials: logging scalars during training, visualizing model graphs, comparing hyperparameter runs, and using the embedding projector for high-dimensional data. The skill gives you working examples for both PyTorch and TensorFlow, which is helpful since the APIs differ. The hyperparameter comparison feature is genuinely useful when you're running multiple experiments and need to see which config actually worked. Main value is having the boilerplate ready to go instead of looking up the SummaryWriter API for the hundredth time. Best for anyone doing iterative model development who needs to see what's happening during training without cluttering up notebooks with matplotlib.

skills.sh installs: 1K
GitHub stars: 14.4K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add orchestra-research/ai-research-skills --skill tensorboard --agent claude-code

Installs into .claude/skills of the current project.

Files

TensorBoard: Visualization Toolkit for ML

When to Use This Skill

Use TensorBoard when you need to:

Visualize training metrics like loss and accuracy over time
Debug models with histograms and distributions
Compare experiments across multiple runs
Visualize model graphs and architecture
Project embeddings to lower dimensions (t-SNE, PCA)
Track hyperparameter experiments
Profile performance and identify bottlenecks
Visualize images and text during training

Users: 20M+ downloads/year | GitHub Stars: 27k+ | License: Apache 2.0

Installation

# Install TensorBoard
pip install tensorboard

# PyTorch integration
pip install torch torchvision tensorboard

# TensorFlow integration (TensorBoard included)
pip install tensorflow

# Launch TensorBoard
tensorboard --logdir=runs
# Access at http://localhost:6006

Quick Start

PyTorch

from torch.utils.tensorboard import SummaryWriter

# Create writer
writer = SummaryWriter('runs/experiment_1')

# Training loop
for epoch in range(10):
    train_loss = train_epoch()
    val_acc = validate()

SKILL.md on GitHub