Configure with gradient clipping helper operations. Auto-activating skill for ML Training.
SkillScore100Maintained
Automatic
◆AI Engineer pack
Last verified 8/30/2026
What it changes
Passed 22 of 22 test cases with the skill, and 21 of 22 without it.
Without skill95%
With skill100%
Checks it passed that the model alone missed
Recommends checking torch.isnan or torch.isinf on the computed norm before applying updates
1 fixed21 passed either way0 still failing0 broke
Pass-rate lift
+5 pp
Tokens
+29%
Agent turns
0%
Response time
+11%
Model gemini-3.6-flashMeasured 8/30/2026v122 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
A test case the skill fixed. Same model, same prompt, with and without the skill.
Prompt
When logging unclipped gradient norms in PyTorch, calling clip_grad_norm_ directly on infinite gradients can produce NaN weights. What property or check should be evaluated on the total norm before performing the optimization step?
What was checked, without and with the skill
Checked
Without skill
With skill
Recommends checking torch.isnan or torch.isinf on the computed norm before applying updates
Missed
Passed
case-13 · gemini-3.6-flash · 8/30/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md60 linessingle file
Gradient Clipping Helper
Overview
This skill provides automated assistance for gradient clipping helper tasks within the ML Training domain.
When to Use
This skill activates automatically when you:
Mention "gradient clipping helper" in your request
Ask about gradient clipping helper patterns or best practices
Need help with machine learning training skills covering data preparation, model training, hyperparameter tuning, and experiment tracking.
Instructions
Provides step-by-step guidance for gradient clipping helper
Follows industry best practices and patterns
Generates production-ready code and configurations
Validates outputs against common standards
Examples
Example: Basic Usage
Request: "Help me with gradient clipping helper"
Result: Provides step-by-step guidance and generates appropriate configurations