skills.rest
kengzkengzCommunity·1 Agent Skills Included

SLM-Lab

Train and benchmark deep reinforcement learning agents with JSON configs

Trains deep reinforcement learning agents using ready-made algorithms like PPO, SAC, DQN, and CrossQ across 70+ game and robotics environments. Defines every experiment in simple JSON spec files, eliminating manual code changes and guaranteeing reproducible results. Automates cloud GPU training, hyperparameter search, performance plots, and benchmark tracking so experiments run faster end to end.
1,362|290
↗
npx skills add kengz/SLM-Lab --all -g -y
Available:

Instructs the AI agent to act as an orchestrating lead engineer that delegates work to agent teams, enforces code quality and testing rules, and follows SLM-Lab's architecture, spec system, and benchmark workflows when running RL experiments.

All Skills in This Repository (1)

Pure Emerald Level Indicators
📦 In Repo
kengzkengz

slm-lab-benchmark

Execute deep reinforcement learning benchmarks in the SLM-Lab framework.

Community
Advanced

Frequently Asked Questions

FAQPage Schema
How to install SLM-Lab?▼

Run `npx skills add kengz/SLM-Lab --all -g -y` in your terminal to install all tools in this suite globally.

What is SLM-Lab used for?▼

SLM-Lab is a modular deep reinforcement learning framework in PyTorch for training and benchmarking RL agents. It supports algorithms like PPO, SAC, and DQN across Atari, MuJoCo, and classic control environments.

How do I run an RL experiment without writing code?▼

Every experiment is fully defined in a JSON spec file, so you just pick an algorithm and environment and run one command. No Python changes are needed to train, search hyperparameters, or generate plots.

Can SLM-Lab train on cloud GPUs?▼

Yes. It integrates with dstack to launch cloud GPU training jobs and automatically syncs results and plots to HuggingFace.

Does SLM-Lab work with AI coding agents?▼

Yes. It ships a CLAUDE.md orchestrator and benchmark skill that let AI agents launch runs, monitor jobs, extract scores, and update benchmark tables autonomously.

Related Repositories in Software Engineering

View All in Software Engineering→
openclawopenclaw
388.4k

openclaw

Run a personal AI assistant across your devices and chat apps

Official📦 120 Skills
obraobra
280.0k

superpowers

Gives coding agents a disciplined workflow from idea to merged code

Community📦 15 Skills
facebookfacebook
248.4k

react

AI agent skills for building, testing, and porting React core

Official📦 14 Skills