skills.rest
anilcan-kara
anilcan-karaCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

slime-rl-training

Run end-to-end reinforcement learning post-training for large language models with slime.

Updated Jun 17, 2026
Search Tags:#llm#reinforcement-learning#slime#rl#sglang#megatron-lm
One-click install
npx skills add https://github.com/anilcan-kara/nozich-agent --skill slime-rl-training-anilcan-kara
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: slime-rl-training
Source: https://github.com/anilcan-kara/nozich-agent/tree/main/optional-skills/mlops/slime
Command: npx skills add https://github.com/anilcan-kara/nozich-agent --skill slime-rl-training-anilcan-kara

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

slime-rl-training provides a structured framework and practical guidance for performing reinforcement learning post-training of large language models using slime, enabling scalable experimentation and production-grade workflows.

Core Features & Use Cases

  • End-to-end RL post-training pipeline combining Megatron-LM training, SGLang rollout, and data buffering to scale experiments.
  • Multi-turn agentic training, tool-use scenarios, and evaluation workflows for robust RL deployments.
  • Pre-configured model scripts, data handling, and troubleshooting guidance that accelerate setup and iteration.

Quick Start

Launch the slime RL training workflow by selecting a pre-configured model script and running the training command with your dataset and required options.

Frequently Asked Questions about slime-rl-training

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I set up reinforcement learning post-training for LLMs using slime?▼

Reinforcement learning post-training for LLMs with slime involves selecting a pre-configured model script and running the training command with your dataset and required options. This launches the end-to-end workflow combining Megatron-LM training and SGLang rollout.

Can I use slime for multi-turn agentic training and tool-use scenarios?▼

Multi-turn agentic training and tool-use scenarios are supported by slime for robust LLM deployments. It provides structured workflows that handle complex interactions and evaluation pipelines for reinforcement learning.

Does slime RL training work with Megatron-LM and SGLang?▼

Slime RL training works directly with Megatron-LM for model training and SGLang for rollout generation. This combination enables scalable experimentation and production-grade workflows for large language models.

What is the best way to manage data buffers during RL post-training?▼

Data buffers during RL post-training are managed through slime's integrated data buffering system. This system handles data handling and custom function hooks to maintain efficient training cycles.

Are there pre-configured model scripts available for slime RL training?▼

Pre-configured model scripts are available for slime RL training to accelerate setup and iteration. These scripts include model configuration options and troubleshooting guidance for streamlined experimentation.

Why does my slime RL training workflow need custom function hooks?▼

Custom function hooks are needed in slime RL training to satisfy requirements for a modular, documented workflow. They allow you to configure arguments and adapt the data buffer system for specific training scenarios.

Related Skills

Explore More Software Engineering Skills→
blueskies1818blueskies1818
1

slime-rl-training

Guide LLM post-training with RL using Megatron-LM and SGLang.

Community
Advanced
ovachieverovachiever
52

grpo-rl-training

Apply GRPO/RL training with TRL to optimize language models.

Community
Advanced
zangjeicyzangjeicy

slime-rl-training

Train LLMs with GRPO and PPO using Megatron-LM and SGLang.

Community
Advanced
ovachieverovachiever
52

openrlhf-training

Train large language models with RLHF using Ray and vLLM acceleration.

Community
Advanced