Master ML Debugging: Fix PyTorch & TensorFlow Training Issues

Master ML Debugging: Fix PyTorch & TensorFlow Training Issues

This AI agent skill assists developers in identifying and addressing specific training failures within machine learning models built using PyTorch, Lightning, or TensorFlow/Keras. It is designed to support resolving issues such as incomplete gradient propagation, NaNs in computations, incorrect loss values, and challenges in reproducing training runs.

What this skill does

• Investigates the symptoms of training issues and transforms them into observable failures with specified inputs.
• Preserves the existing framework, backend, and training semantics originally intended by the user.
• Records detailed information relevant to the failure, including command lines, software versions, and environment specifics.
• Assists in identifying the boundaries of failure by reviewing tensor shapes and their interaction with loss functions.
• Classifies issues related to dependency/import failures separately from training bugs.
• Focuses on maintaining the integrity of input shapes and value ranges to diagnose the root cause accurately, avoiding blanket fixes.

Who it is for

This skill benefits developers and teams using AI coding agents like Claude Code, Cursor, and Codex who need to diagnose and fix specific machine learning training issues. It is particularly useful for engineers engaged in model development and maintenance who require precise and insightful debugging capabilities.

Use cases

• Developers encountering non-deterministic behavior in their training setups who need to establish reproducibility.
• Engineers needing to diagnose specific instances of NaNs in tensor operations.
• Teams troubleshooting unexpected discrepancies in loss calculations due to incorrect tensor broadcasting.
• Ensuring machine learning deployments are reliable by confirming no silent errors exist in training workflows.

Technical details

• Tool is optimized for investigation of ML training issues using PyTorch, Lightning, and TensorFlow/Keras frameworks.
• Supports troubleshooting within CPU or eager-mode environments to isolate problems effectively.
• Records execution-specific parameters such as data order, seed handling, checkpoint and optimizer states for completeness in debugging sessions.

Source & Licence

This package is built on open-source work published by 00200200 (00200200/maintainer-skills-lab) and distributed under MIT. The original licence text and copyright notice are included in your download.

Personal and commercial use, modification and redistribution are permitted, provided the original copyright and licence notice are retained.

Your purchase covers curation, licence verification, packaging, documentation and instant delivery. It does not grant exclusive rights to the underlying open-source code, which remains available under its original licence.

Delivery & Support

• Delivery: instant — a secure download link is emailed to you as soon as payment is confirmed.
• Format: ZIP archive containing the skill files, documentation and the original licence.
• Support: support@mcpcart.com — we aim to reply within 2 business days.
• Updates: updates are included only where stated on this page.

Refunds

This is a digital product delivered immediately after purchase. By completing your order you request immediate delivery and acknowledge that, once the download has been accessed, the statutory right to cancel no longer applies to the extent permitted by law. Refund requests are handled in accordance with our published Refund Policy.

Claude, Codex, Gemini and Cursor are trademarks of their respective owners. MCP Cart is an independent marketplace and is not affiliated with, endorsed by, or sponsored by any of them. Compatibility references describe interoperability only.