New to Claude Skills? Learn how to install them →

davila7 on GitHub

Slime RL Training

Free

Post-training framework for LLM reinforcement learning.

Get this skill

Free · Opens the source repo

What Slime RL Training does

Slime is a specialized framework designed for post-training reinforcement learning (RL) of large language models (LLMs). Developed by Tsinghua University's THUDM team, it integrates seamlessly with Megatron-LM for training and SGLang for efficient rollout generation. This framework is particularly beneficial for those working with models like GLM-4.5, GLM-4.6, and GLM-4.7, allowing users to leverage the full capabilities of Megatron-LM while employing SGLang for high-throughput inference.

The architecture of Slime is structured around a data buffer that manages prompt initialization and sample storage, facilitating custom data generation workflows. This capability is crucial for researchers or developers who require flexibility in their data handling processes. Additionally, Slime supports various training configurations with full parallelism, enabling users to optimize their resource utilization during model training.

Slime is an ideal choice for developers and researchers looking to implement custom data generation workflows or those needing tight integration with Megatron-LM for RL scaling. It is particularly suited for training models like GLM, Qwen3, DeepSeek V3, or Llama 3. However, users seeking enterprise-grade stability or specific backend flexibility may want to explore alternative frameworks. Overall, Slime stands out as a research-grade tool backed by production capabilities from Z.ai, making it a robust option for advanced machine learning tasks.

When to use it

Use Slime when you need to perform post-training on large language models with a focus on reinforcement learning, especially when working with Megatron-LM and SGLang.

When not to use it

Avoid Slime if you require enterprise-grade stability features or if your project demands flexible backend swapping, as other tools may be better suited for those needs.

What you can build with it

Training GLM Models

Utilize Slime to effectively train GLM models with integrated reinforcement learning techniques.

Custom Data Generation Workflows

Implement flexible data generation workflows tailored to your specific training needs.

High-Throughput Rollouts

Leverage SGLang for high-throughput response generation during model rollout.

How to install Slime RL Training

View source

1. Install with the skills CLI

npx skills add davila7/claude-code-templates/post-training-slime --agent claude-code

2. Or install it manually

Download the skill folder and drop it into ~/.claude/skills/ for all projects, or .claude/skills/ to scope it to one repo. Restart Claude Code so it picks up the new skill.

Anthropic's agentic coding CLI, and the reference implementation of Agent Skills. Drop a skill folder into ~/.claude/skills and Claude Code loads it automatically whenever a task matches the skill's description. Claude Code docs

Inside SKILL.md

Written by davila7

slime: LLM Post-Training Framework for RL Scaling

slime is an LLM post-training framework from Tsinghua's THUDM team, powering GLM-4.5, GLM-4.6, and GLM-4.7. It connects Megatron-LM for training with SGLang for high-throughput rollout generation.

When to Use slime

Choose slime when you need:

  • Megatron-LM native training with SGLang inference
  • Custom data generation workflows with flexible data buffers
  • Training GLM, Qwen3, DeepSeek V3, or Llama 3 models
  • Research-grade framework with production backing (Z.ai)

Consider alternatives when:

  • You need enterprise-grade stability features → use miles
  • You want flexible backend swapping → use verl
  • You need PyTorch-native abstractions → use torchforge

Key Features

  • Training: Megatron-LM with full parallelism support (TP, PP, DP, SP)
  • Rollout: SGLang-based high-throughput generation with router
  • Data Buffer: Flexible prompt management and sample storage
  • Models: GLM-4.x, Qwen3, DeepSeek V3/R1, Llama 3

Architecture Overview

┌─────────────────────────────────────────────────────────┐
│                    Data Buffer                          │
│ - Prompt initialization and management                  │
│ - Custom data generation and filtering                  │
│ - Rollout sample storage                                │
└─────────────┬───────────────────────────┬───────────────┘
              │                           │
┌─────────────▼───────────┐ ┌─────────────▼───────────────┐
│ Training (Megatron-LM)  │ │ Rollout (SGLang + Router)   │
│ - Actor model training  │ │ - Response generation       │
│ - Critic (optional)     │ │ - Reward/verifier output    │
│ - Weight sync to rollout│ │ - Multi-turn support        │
└─────────────────────────┘ └─────────────────────────────┘

Installation

# Recommended: Docker
docker pull slimerl/slime:latest
docker run --rm --gpus all --ipc=host --shm-size=16g \
  -it slimerl/slime:latest /bin/bash

# Inside container
cd /root/slime && pip install -e . --no-deps

From Source

git clone https://github.com/THUDM/slime.git
cd slime
pip install -r requirements.txt
pip install -e .

Quick Start: GRPO Training

# Source model configuration
source scripts/models/qwen3-4B.sh

# Launch training
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 \
    ${MODEL_ARGS[@]} ${CKPT_ARGS[@]}

Workflow 1: Standard GRPO Training

Use this workflow for training reasoning models with group-relative advantages.

Prerequisites Checklist

  • Docker environment or Megatron-LM + SGLang installed
  • Model checkpoint (HuggingFace or Megatron format)
  • Training data in JSONL format

Step 1: Prepare Data

# data.jsonl format
{"prompt": "What is 2 + 2?", "label": "4"}
{"prompt": "Solve: 3x = 12", "label": "x = 4"}

Or with chat format:

{
    "prompt": [
        {"role": "system", "content": "You are a math tutor."},
        {"role": "user", "content": "What is 15 + 27?"}
    ],
    "label": "42"
}

Step 2: Configure Model

Choose a pre-configured model script:

# List available models
ls scripts/models/
# glm4-9B.sh, qwen3-4B.sh, qwen3-30B-A3B.sh, deepseek-v3.sh, llama3-8B.sh, ...

# Source your model
source scripts/models/qwen3-4B.sh

Step 3: Launch Training

python train.py \
    --actor-num-nodes 1 \
    --actor-num-gpus-per-node 8 \
    --rollout-num-gpus 8 \
    --advantage-estimator grpo \
    --use-kl-loss \
    --kl-loss-coef 0.001 \
    --prompt-data /path/to/train.jsonl \
    --input-key prompt \
    --label-key label \
    --apply-chat-template \
    --rollout-batch-size 32 \
    --n-samples-per-prompt 8 \
    --global-batch-size 256 \
    --num-rollout 3000 \
    --save-interval 100 \
    --eval-interval 50 \
    ${MODEL_ARGS[@]}

Step 4: Monitor Training

  • Check TensorBoard: tensorboard --logdir outputs/
  • Verify reward curves are increasing
  • Monitor GPU utilization across nodes

Workflow 2: Asynchronous Training

Use async mode for higher throughput by overlapping rollout and training.

When to Use Async

  • Large models with long generation times
  • High GPU idle time in synchronous mode
  • Sufficient memory for buffering

Launch Async Training

python train_async.py \
    --actor-num-nodes 1 \
    --actor-num-gpus-per-node 8 \
    --rollout-num-gpus 8 \
    --advantage-estimator grpo \
    --async-buffer-size 4 \
    --prompt-data /path/to/train.jsonl \
    ${MODEL_ARGS[@]}

Async-Specific Parameters

--async-buffer-size 4        # Number of rollouts to buffer
--update-weights-interval 2  # Sync weights every N rollouts

Workflow 3: Multi-Turn Agentic Training

Use this workflow for training agents with tool use or multi-step reasoning.

Prerequisites

  • Custom generate function for multi-turn logic
  • Tool/environment interface

Step 1: Define Custom Generate Function

# custom_generate.py
async def custom_generate(args, samples, evaluation=False):
    """Multi-turn generation with tool calling."""
    for sample in samples:
        conversation = sample.prompt

        for turn in range(args.max_turns):
            # Generate response
            response = await generate_single(conversation)

            # Check for tool call
            tool_call = extract_tool_call(response)
            if tool_call:
                tool_result = execute_tool(tool_call)
                conversation.append({"role": "assistant", "content": response})
                conversation.append({"role": "tool", "content": tool_result})
            else:
                break

        sample.response = response
        sample.reward = compute_reward(sample)

    return samples

Step 2: Launch with Custom Function

python train.py \
    --custom-generate-function-path custom_generate.py \
    --max-turns 5 \
    --prompt-data /path/to/agent_data.jsonl \
    ${MODEL_ARGS[@]}

See examples/search-r1/ for a complete multi-turn search example.


Configuration Reference

Three Argument Categories

slime uses three types of arguments:

1. Megatron Arguments (passed directly):

--tensor-model-parallel-size 2
--pipeline-model-parallel-size 1
--num-layers 32
--hidden-size 4096

2. SGLang Arguments (prefixed with --sglang-):

--sglang-mem-fraction-static 0.8
--sglang-context-length 8192
--sglang-log-level INFO

3. slime Arguments:

# Resource allocation
--actor-num-nodes 1
--actor-num-gpus-per-node 8
--rollout-num-gpus 8
--colocate  # Share GPUs between training/inference

# Data
--prompt-data /path/to/data.jsonl
--input-key prompt
--label-key label

# Training loop
--num-rollout 3000
--rollout-batch-size 32
--n-samples-per-prompt 8
--global-batch-size 256

# Algorithm
--advantage-estimator grpo  # or: gspo, ppo, reinforce_plus_plus
--use-kl-loss
--kl-loss-coef 0.001

Key Constraints

rollout_batch_size × n_samples_per_prompt = global_batch_size × num_steps_per_rollout

Example: 32 × 8 = 256 × 1


Data Buffer System

slime's data buffer enables flexible data management:

Basic Data Source

class RolloutDataSource:
    def get_samples(self, num_samples):
        """Fetch prompts from dataset."""
        return self.dataset.sample(num_samples)

    def add_samples(self, samples):
        """Called after generation (no-op by default)."""
        pass

Buffered Data Source (Off-Policy)

class RolloutDataSourceWithBuffer(RolloutDataSource):
    def __init__(self):
        self.buffer = []

    def add_samples(self, samples):
        """Store generated samples for reuse."""
        self.buffer.extend(samples)

    def buffer_filter(self, args, buffer, num_samples):
        """Custom selection logic (prioritized, stratified, etc.)."""
        return select_best(buffer, num_samples)

Common Issues and Solutions

Issue: SGLang Engine Crash

Symptoms: Inference engine dies mid-training

Solutions:

# Enable fault tolerance
--use-fault-tolerance

# Increase memory allocation
--sglang-mem-fraction-static 0.85

# Reduce batch size
--rollout-batch-size 16

Issue: Weight Sync Timeout

Symptoms: Training hangs after rollout

Solutions:

# Increase sync interval
--update-weights-interval 5

# Use colocated mode (no network transfer)
--colocate

Issue: OOM During Training

Symptoms: CUDA OOM in backward pass

Solutions:

# Enable gradient checkpointing
--recompute-activations

# Reduce micro-batch size
--micro-batch-size 1

# Enable sequence parallelism
--sequence-parallel

Issue: Slow Data Loading

Symptoms: GPU idle during data fetch

Solutions:

# Increase data workers
--num-data-workers 4

# Use streaming dataset
--streaming-data

Supported Models

Model FamilyConfigurations
GLMGLM-4.5, GLM-4.6, GLM-4.7, GLM-Z1-9B
QwenQwen3 (4B, 8B, 30B-A3B), Qwen3-MoE, Qwen2.5
DeepSeekV3, V3.1, R1
LlamaLlama 3 (8B, 70B)
OthersKimi K2, Moonlight-16B

Each model has pre-configured scripts in scripts/models/.


Advanced Topics

Co-location Mode

Share GPUs between training and inference to reduce memory:

python train.py \
    --colocate \
    --actor-num-gpus-per-node 8 \
    --sglang-mem-fraction-static 0.4 \
    ${MODEL_ARGS[@]}

Custom Reward Model

# custom_rm.py
class CustomRewardModel:
    def __init__(self, model_path):
        self.model = load_model(model_path)

    def compute_reward(self, prompts, responses):
        inputs = self.tokenize(prompts, responses)
        scores = self.model(inputs)
        return scores.tolist()
--custom-rm-path custom_rm.py

Evaluation Multi-Task

--eval-prompt-data aime /path/to/aime.jsonl \
--eval-prompt-data gsm8k /path/to/gsm8k.jsonl \
--n-samples-per-eval-prompt 16

Resources

Frequently asked questions about Slime RL Training

Similar skills