Skills
.Work. You
Rest

debug:tensorflow

Community

Systematically debug TensorFlow apps and models.

Software Engineering#debugging#gpu#tensorflow#cuda#keras#tensorboard#tf.debugging
AuthorSnakeO
Version1.0.0
Installs0

System Documentation

What problem does it solve?

This Skill provides a structured approach to diagnosing and resolving TensorFlow and Keras issues that hinder model development, training stability, or deployment.

Core Features & Use Cases

  • Shape and dtype validation: quickly verify tensor shapes and data types match model expectations.
  • GPU/CUDA diagnostics and memory management: detect device availability, memory growth, and compatibility.
  • NaN/Inf loss handling: identify and mitigate numerical instability in loss and metrics.
  • Gradient analysis: inspect gradient flow and clipping to diagnose training stalls.
  • SavedModel loading and compatibility: ensure correct serialization formats and custom objects.
  • Data pipeline profiling: identify bottlenecks in tf.data pipelines and optimize throughput.

Quick Start

Run a minimal diagnostic on a TensorFlow project by enabling eager execution and performing a small-scale check on synthetic data and a tiny model.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: debug:tensorflow
Download link: https://github.com/SnakeO/claude-debug-and-refactor-skills-plugin/archive/main.zip#debug-tensorflow

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 223,000+ vetted skills library on demand.