Training Debug & Performance Tuning
Debug and optimize model training with systematic loss curve analysis
What You Can Do
You can systematically diagnose why your models aren't converging, identify which hyperparameters matter most for your training dynamics, and generate prioritized optimization strategies. By analyzing loss curves, gradient flow patterns, and validation metrics, you'll accelerate convergence, reduce wasted compute, and achieve better generalization—without trial-and-error tuning.
Features
decode divergence, saturation, oscillation, and overfitting signals to understand what's happening during training
rank which parameters (learning rate, batch size, regularization, warmup) most affect your model's convergence
detect vanishing/exploding gradients, activation saturation, and dead neurons that stall learning
receive prioritized, actionable recommendations to reach target performance faster
systematically distinguish overfitting from underfitting and quantify generalization risk
benchmark warmup strategies, decay schedules, and adaptive methods for your architecture
evaluate computational cost vs. model quality for different configurations
systemize diagnosis of NaN losses, divergence, stalled progress, and instability
Example Output
Example 1: Loss Curve Analysis
Observations:
✓ Training loss decreases smoothly → healthy gradient flow
✗ Validation loss plateaus after epoch 15 → overfitting begins
✗ Loss spike at epoch 8 → likely learning rate too high or batch effect
Root Cause: Validation overfitting, training LR still acceptable
Recommendations:
1. Add L2 regularization (start λ=0.001)
2. Implement early stopping with patience=5
3. Increase dropout to 0.5 in dense layers
Example 2: Hyperparameter Sensitivity Ranking
Impact on Final Validation Accuracy:
1. Learning Rate (±15% variance) — CRITICAL
2. Warmup Steps (±8% variance) — HIGH
3. Batch Size (±5% variance) — MEDIUM
Top 3 Actions:
- Reduce LR from 1e-3 to 5e-4 (steeper convergence)
- Extend warmup from 1000 to 3000 steps (smoother early training)
- Increase batch size 32→64 (30% faster with better gradient estimates)
Example 3: Convergence Acceleration Plan
Current: 85% accuracy in 50 epochs → Target: 87% in 30 epochs
Proposed Changes (estimated 35% speedup):
1. Scheduler: Linear warmup (3k steps) → cosine decay
2. Regularization: L2(0.0001), increase dropout 0.3→0.4
3. Batch: 32→64 with augmentation on 80% of data
4. Expected: reach 85% by epoch 30, potential 87% by epoch 40
What's Included
- SKILL.md: Complete training diagnostics workflow with decision trees for systematic analysis
- Loss Curve Analyzer checklist: Step-by-step guide to interpret patterns and identify failure modes
- Hyperparameter sensitivity template: Framework for ranking parameters by impact on convergence
- Training dynamics worksheet: Prompts to structure your loss logs, metrics, and observations
- Convergence acceleration playbook: Prioritized strategies for your specific bottleneck
- Debugging flowchart: Decision tree to diagnose NaN, divergence, or stalled training
- Learning rate tuning guide: Warmup and decay schedule recommendations by architecture
Who It's For
- ML Engineers building and deploying production models who need faster iteration cycles
- Research Scientists tuning deep learning models and investigating convergence issues
- ML Ops Engineers optimizing training pipelines and reducing computational cost
- Data Scientists training custom models on limited budgets or hardware
- Computer Vision/NLP Specialists debugging architecture-specific training dynamics
Best For
- Debugging training failures (NaN loss, divergence, stalled progress, oscillation)
- Hyperparameter optimization without expensive grid search
- Analyzing and accelerating convergence speed
- Reducing wasted compute by identifying ineffective configurations early
- Improving model generalization and reducing overfitting
You might also like
You can accelerate your entire antenna design workflow—from initial concept to optimized specifications—by leveraging Claude's knowledge of electromagnetic theory, antenna types, and design trade-offs. Claude helps you parameterize designs, identify optimal configurations for your frequency band and application, and prepare detailed briefs for simulation and manufacturing. You'll generate design reviews, impedance-matching solutions, and validation checklists without context-switching to multiple tools.
You can convert raw API code, OpenAPI specs, GraphQL schemas, and implementation details into production-ready reference documentation. The skill generates systematically organized endpoint documentation with parameter tables, real request/response examples, error code catalogs, authentication details, rate limits, and SDK usage snippets in multiple programming languages—creating the authoritative developer reference for REST APIs, GraphQL APIs, WebSocket APIs, and webhook schemas.
Rapidly analyze production incidents by correlating observability signals—logs, metrics, traces, and events—to identify the true root cause instead of symptoms. You'll reconstruct incident timelines, quantify impact scope, and generate both immediate remediation steps and long-term prevention measures.
This skill guides you through designing Kalman filters for your specific state estimation and sensor fusion problems. You'll learn to diagnose filter divergence, tune covariance matrices (Q and R), and validate observability—turning mathematical theory into production-ready filters. The skill provides step-by-step workflows, diagnostic tools, and practical templates for common applications like GPS/IMU fusion, SLAM, and aircraft navigation.
Claude analyzes your code diffs and generates standardized commit messages that follow the Conventional Commits specification. The skill automatically determines the correct commit type, scope, and description based on the changes you've made, ensuring your messages are parseable by automation tools while remaining human-readable for code reviewers.
You'll transform high-level feature requirements into a complete development plan covering API design, frontend implementation, testing strategy, and documentation. The skill helps you identify dependencies, parallelize work across frontend and backend teams, and establish quality gates that ensure features ship bug-free with proper error handling and logging.
Transform schema and data changes into production-ready migration packages that protect data integrity and minimize downtime. You can handle column additions, renames, type conversions, and backfills using proven patterns like expand-contract and shadow tables. The skill generates explicit rollback scripts, validation checklists, and phased execution plans that keep applications stable throughout the migration.
You can rapidly design, optimize, and validate chromatographic and LC-MS/MS methods for analyzing drugs and metabolites in biological matrices. This skill guides you through analytical method development workflows—including parameter selection, sensitivity optimization, and regulatory strategy—while ensuring ICH Q2(R2) and FDA compliance throughout the development process.