Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Manage mixed precision trainer operations. Auto-activating skill for ML Training. Triggers on: mixed precision trainer, mixed precision trainer Part of the ML Training skill category. Use when working with mixed precision trainer functionality. Trigger with phrases like "mixed precision trainer", "mixed trainer", "mixed".
.claude/skills/dicklesworthstone-mixed-precision-trainer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 54% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 57% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 19% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 9% | 0% |
This skill provides automated assistance for mixed precision trainer tasks within the ML Training domain.
This skill activates automatically when you:
Example: Basic Usage Request: "Help me with mixed precision trainer" Result: Provides step-by-step guidance and generates appropriate configurations
| Error | Cause | Solution | |-------|-------|----------| | Configuration invalid | Missing required fields | Check documentation for required parameters | | Tool not found | Dependency not installed | Install required tools per prerequisites | | Permission denied | Insufficient access | Verify credentials and permissions |
Part of the ML Training skill category. Tags: ml, training, pytorch, tensorflow, sklearn
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 16,426 | 24,240 | +48% | 1 | 1 | 0% | 2,948 | 3,494 | +19% | 0 | 0 | — |
case-02 | pass→pass | 13,974 | 13,260 | -5% | 1 | 1 | 0% | 2,778 | 3,041 | +9% | 0 | 0 | — |
case-03 | fail→pass | 16,223 | 12,975 | -20% | 1 | 1 | 0% | 2,153 | 2,899 | +35% | 0 | 0 | — |
case-04 | fail→pass | 14,700 | 21,917 | +49% | 1 | 1 | 0% | 1,934 | 2,973 | +54% | 0 | 0 | — |
case-05 | fail→fail | 14,735 | 10,453 | -29% | 1 | 1 | 0% | 1,941 | 2,406 | +24% | 0 | 0 | — |
case-06 | pass→pass | 14,920 | 18,771 | +26% | 1 | 1 | 0% | 1,905 | 2,541 | +33% | 0 | 0 | — |
case-07 | pass→pass | 11,580 | 6,871 | -41% | 1 | 1 | 0% | 1,323 | 1,608 | +22% | 0 | 0 | — |
case-08 | pass→pass | 12,947 | 14,894 | +15% | 1 | 1 | 0% | 1,611 | 1,969 | +22% | 0 | 0 | — |
case-09 | pass→pass | 11,423 | 13,057 | +14% | 1 | 1 | 0% | 1,195 | 1,931 | +62% | 0 | 0 | — |
case-10 | pass→pass | 16,611 | 17,498 | +5% | 1 | 1 | 0% | 2,465 | 2,767 | +12% | 0 | 0 | — |
case-11 | pass→pass | 16,180 | 15,579 | -4% | 1 | 1 | 0% | 2,180 | 2,466 | +13% | 0 | 0 | — |
case-12 | pass→pass | 22,823 | 29,829 | +31% | 1 | 1 | 0% | 3,233 | 3,947 | +22% | 0 | 0 | — |
case-13 | pass→pass | 20,212 | 20,192 | -0% | 1 | 1 | 0% | 2,779 | 3,061 | +10% | 0 | 0 | — |
case-14 | pass→pass | 22,285 | 21,679 | -3% | 1 | 1 | 0% | 3,161 | 3,406 | +8% | 0 | 0 | — |
case-15 | pass→pass | 11,160 | 13,822 | +24% | 1 | 1 | 0% | 2,315 | 3,273 | +41% | 0 | 0 | — |
case-16 | pass→pass | 16,644 | 18,127 | +9% | 1 | 1 | 0% | 2,161 | 2,803 | +30% | 0 | 0 | — |
case-17 | fail→pass | 19,732 | 12,209 | -38% | 1 | 1 | 0% | 1,119 | 1,759 | +57% | 0 | 0 | — |
case-18 | pass→pass | 9,738 | 14,611 | +50% | 1 | 1 | 0% | 1,558 | 2,118 | +36% | 0 | 0 | — |
case-19 | pass→pass | 11,860 | 18,913 | +59% | 1 | 1 | 0% | 2,387 | 3,092 | +30% | 0 | 0 | — |
case-20 | pass→pass | 10,780 | 16,867 | +56% | 1 | 1 | 0% | 2,108 | 2,567 | +22% | 0 | 0 | — |
case-21 | pass→pass | 7,507 | 14,007 | +87% | 1 | 1 | 0% | 1,531 | 2,110 | +38% | 0 | 0 | — |
case-22 | pass→pass | 9,364 | 10,595 | +13% | 1 | 1 | 0% | 723 | 1,385 | +92% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 21 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +14 percentage points is the difference between those two pass rates over the 21 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.