learning-rate-scheduler

Learning Rate Scheduler - Auto-activating skill for ML Training. Triggers on: learning rate scheduler, learning rate scheduler Part of the ML Training skill category.

1.02x

Quality

Does it follow best practices?

Impact

93%

1.02x

Average score across 3 eval scenarios

Securityby

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./planned-skills/generated/07-ml-training/learning-rate-scheduler/SKILL.md

Evaluation results

97%

11%

Fine-Tuning Training Script with Learning Rate Scheduling

Production LR scheduler implementation with validation

Criteria

Without context

With context

Standard framework scheduler

83%

Named schedule type

83%

91%

Schedule validation output

100%

LR values printed during training

100%

Parameterized configuration

100%

Scheduler stepped correctly

100%

Warmup or decay phase

100%

Python implementation

100%

Trains for multiple epochs

100%

No large downloads required

100%

Without context: $0.5962 · 2s · 1 turns · 3 in / 33 out tokens

With context: $0.4845 · 5m 15s · 29 turns · 60 in / 6,024 out tokens

84%

-5%

Learning Rate Schedule Comparison for Model Training

LR schedule hyperparameter comparison study

Criteria

Without context

With context

At least 3 schedulers compared

100%

Uses standard framework scheduler API

100%

LR values captured per run

100%

Final loss captured

100%

Same random seed used

30%

60%

LR curves visualization

100%

Recommendation documented

100%

Hyperparameters named

50%

100%

Structured step-by-step approach

100%

No large downloads

100%

Without context: $0.5121 · 3s · 1 turns · 3 in / 126 out tokens

With context: $0.9351 · 3s · 1 turns · 3 in / 49 out tokens

100%

Reusable Training Configuration Module with LR Scheduling

Step-by-step training pipeline with LR scheduling

Criteria

Without context

With context

Modular trainer design

100%

Standard scheduler API used

100%

Scheduler stepped in correct position

100%

LR logged per epoch

100%

Step-by-step summary document

100%

Validation statement present

100%

Hyperparameters documented

100%

Schedule includes decay or warmup

100%

Python and pip-installable libraries

100%

Trains at least 5 epochs

100%

No large model downloads

100%

Without context: $0.2917 · 3m 49s · 17 turns · 18 in / 5,143 out tokens

With context: $0.4791 · 4m 6s · 26 turns · 60 in / 6,391 out tokens

Repository: jeremylongshore/claude-code-plugins-plus-skills
Commit: f17dd51

Evaluated: 17 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

Fine-Tuning Training Script with Learning Rate Scheduling Learning Rate Schedule Comparison for Model Training Reusable Training Configuration Module with LR Scheduling

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.