skills.rest
nadicodeai
nadicodeaiOFFICIAL
Explore all skills by this author →
WebsiteGitHub

slime-rl-training

Configure RL fine-tuning pipelines for GLM models with Megatron-LM and SGLang.

Updated Apr 9, 2026
Search Tags:#reinforcement-learning#slime#rl#sglang#llm-training#megatron-lm
One-click install
npx skills add https://github.com/nadicodeai/argo-agent --skill slime-rl-training-nadicodeai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: slime-rl-training
Source: https://github.com/nadicodeai/argo-agent/tree/main/optional-skills/mlops/slime
Command: npx skills add https://github.com/nadicodeai/argo-agent --skill slime-rl-training-nadicodeai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

slime provides an end-to-end guide for RL-based fine-tuning of large language models using the slime framework, integrating Megatron-LM and SGLang to streamline data rollout, training loops, and evaluation setup.

Core Features & Use Cases

  • Guidance for configuring slime-style RL pipelines (GRPO, async, multi-turn) on GLM-family models.
  • Explanation of data buffers, rollout orchestration, and Megatron-LM + SGLang integration for scalable training.
  • Practical use cases including post-training for GLMs with high-throughput data generation and tool-enabled prompts.

Quick Start

Follow the Quick Start to run a GRPO training workflow for GLM models using Megatron-LM and SGLang.

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 GRPO training for GLM-family models using slime?▼

Set up GRPO training for GLM-family models by defining dependencies, configuration patterns, and data buffers across slime, Megatron-LM, and SGLang to orchestrate end-to-end training loops.

What is the best way to integrate Megatron-LM and SGLang for high-throughput LLM rollouts?▼

Use slime to orchestrate high-throughput LLM data rollouts by configuring Megatron-LM and SGLang integration, enabling scalable policy optimization and efficient data buffer management.

Can I use slime for multi-turn tool usage and async reinforcement learning pipelines?▼

Yes, slime supports async and multi-turn reinforcement learning pipelines, providing custom generate functions and data buffers to handle multi-turn tool usage and policy optimization.

Does slime support PPO variants alongside GRPO for LLM fine-tuning?▼

Yes, slime supports policy optimization using GRPO and PPO variants, providing configuration patterns to streamline reinforcement learning fine-tuning for large language models.

Why do I need Megatron-LM and SGLang for RL-based LLM training?▼

Megatron-LM and SGLang are integrated with slime to solve the complexity of RL-based fine-tuning, enabling scalable training loops and high-throughput rollout orchestration for large language models.

Related Skills

Explore More Data & Analytics Skills→
tadod12tadod12

slime-rl-training

Configure Megatron-LM and SGLang arguments for post-training RL workflows.

Community
Advanced
ScientiaCapitalScientiaCapital
1

model-deployment

Export fine-tuned models to GGUF and deploy with vLLM or Ollama.

Community
Advanced
eddieluengeddielueng
1

slime-rl-training

Run GRPO RL post-training for LLMs with Megatron-LM and SGLang rollouts.

Community
Advanced
jacardljacardl

slime-rl-training

Run RL post-training for large language models with Megatron-LM and SGLang.

Community
Advanced