Domains
Forms
Mechanism
Breaks down how the skill works and what it produces, so you can quickly judge whether it fits your scenario and value.
When writing or reviewing a LoRA/QLoRA training config, choosing rank/alpha/target modules, or deciding between LoRA, QLoRA, and full FT after finetuning-method-selection routes to SFT.
A validated, runnable LoRA/QLoRA adapter configuration with best-practice hyperparameters, reducing config errors and training failures.
Use Cases
An ML engineer prepares to fine-tune a 7B model on a custom support dataset. They use this skill to set target_modules to all-linear, choose rank=32 with alpha=64, and apply Unsloth defaults like lora_dropout=0. The validated config avoids fp16 divergence and module-trimming mistakes, letting the training script run correctly on the first try.
A team wants to adapt a 65B-class model on a 48GB GPU. The skill shows QLoRA with NF4 base is the right choice, warns that QLoRA may OOM on DGX Spark and points to memory remediation. They produce a QLoRA config with bf16=True and rank sized to task, preventing silent divergence and wasted runs.
Skill Relationships
Dependency relationships read as "the upper tier points to the lower tier." The current Skill sits in the middle tier — above are Skills that depend on it, below are Skills it depends on.
Tier 1 · These Skills Use Me
Tier 2 · Current Skill
Tier 3 · I Use These Skills
Browse this skill's relationships within its skillset. Click a node to switch the side panel; use the search box to jump to any skill.
Skill File
