slime-rl-training

Automate LLM post-training with slime and Megatron-LM integration.

1|1|Updated May 25, 2026
One-click install
npx skills add https://github.com/aayushsoam/clawbot-agent --skill slime-rl-training-aayushsoam
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: slime-rl-training
Source: https://github.com/aayushsoam/clawbot-agent/tree/main/optional-skills/mlops/slime
Command: npx skills add https://github.com/aayushsoam/clawbot-agent --skill slime-rl-training-aayushsoam

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires sglang-router>=0.2.3, ray, torch>=2.0.0, transformers>=4.40.0, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates the post-training process for large language models (LLMs) using the slime framework, providing a streamlined approach to reinforcement learning (RL) and scaling Megatron-LM for RL applications.

Core Features & Use Cases

  • LLM Post-Training: Facilitates the use of the slime framework for post-training of LLMs, enabling RL enhancements.
  • Megatron-LM Integration: Offers seamless integration with Megatron-LM for RL scaling.
  • Custom Data Generation: Supports custom data generation workflows for flexible prompt management and sample storage.
  • Use Case: Ideal for training GLM models, implementing custom data generation, or needing tight Megatron-LM integration for RL scaling.

Quick Start

Run the following command to start a slime training session:

python train.py --actor-num-nodes 1 --actor-num-gpus-per-node 4 --rollout-num-gpus 4 --advantage-estimator grpo --use-kl-loss --kl-loss-coef 0.001 --rollout-batch-size 32 --n-samples-per-prompt 8 --global-batch-size 256 --num-rollout 3000 --prompt-data /path/to/data.jsonl

Frequently Asked Questions about slime-rl-training

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

FAQPage Schema
How do I automate LLM post-training with reinforcement learning and Megatron-LM?

Automate LLM post-training with RL by using the slime framework to scale Megatron-LM, streamlining reinforcement learning integration and custom data generation workflows.

How do I scale Megatron-LM for reinforcement learning training?

Scale Megatron-LM for RL training through the slime framework, which provides seamless integration for distributed execution alongside Ray and sglang-router.

Do I need Ray and sglang-router to run slime reinforcement learning training?

Yes, running slime RL training requires Ray and sglang-router, alongside Torch and Transformers, to manage distributed execution and model serving.

Can I use custom data generation for GLM model reinforcement learning?

Yes, custom data generation is supported for GLM model RL training, allowing flexible prompt management and sample storage directly within the workflow.

What is the best way to configure GRPO advantage estimation for LLM post-training?

Configure GRPO advantage estimation for LLM post-training by passing advantage estimator parameters and KL loss coefficients directly in the slime launch command.