Skills
.Work. You
Rest

finding-lobbench-results

Community

Find existing LOBbench results before benching.

Data & Analytics#pipeline#experiments#results#wandb#slurm#bench#lobbench
AuthorKangOxford
Version1.0.0
Installs0

System Documentation

What problem does it solve?

LOBbench results for a given checkpoint, SLURM job, wandb run, or experiment are often already available in pipeline directories. Re-running benches wastes GPU hours and time. This Skill helps you locate existing results quickly to avoid unnecessary computations.

Core Features & Use Cases

  • Search-first discipline: Always check agent_outputs, kangli lob_pipeline, and aramis pipelines for existing results.
  • Quick verification: Identifies matching checkpoints, wandb runs, or experiments and returns exact paths and metrics.
  • Use case: Before launching a new LOBbench, run this Skill to confirm whether a prior bench already produced comparable results.

Quick Start

Provide a target identifier (checkpoint, SLURM job, wandb run, or experiment) to locate existing LOBbench results.

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: finding-lobbench-results
Download link: https://github.com/KangOxford/auto-quant-research/archive/main.zip#finding-lobbench-results

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.