Skills
.Work. You
Rest

perform-sweep

Community

Run ablation sweeps to optimize GRPO.

Data & Analytics#hyperparameters#experiments#GRPO#ablation#sweep#WandB
Authorbglick13
Version1.0.0
Installs0

System Documentation

What problem does it solve?

End-to-end workflow for running ablation experiments on the Diplomacy GRPO training pipeline.

Core Features & Use Cases

  • Fire-and-forget: Launch sweeps in Modal cloud and monitor progress remotely.
  • Auto-resume: If Modal times out (24hr max), sweep automatically respawns.
  • Sequential execution: Runs one training at a time (infra constraint) to preserve resources.
  • Progress tracking: State is saved after each run for easy recovery and rerun.

Quick Start

Create your sweep configuration under experiments/sweeps/<name>/sweep.yaml, then launch with: python scripts/launch_sweep.py experiments/sweeps/<name>/

Dependency Matrix

Required Modules

None required

Components

scripts

💻 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: perform-sweep
Download link: https://github.com/bglick13/diplomacy-v2/archive/main.zip#perform-sweep

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 223,000+ vetted skills library on demand.