Post Training Slime Curated Resource and Installation Guide CookApps CookApps Marketplace

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

bashcode

# 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

bashcode

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

Quick Start: GRPO Training

bashcode

# 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

pythoncode

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

Or with chat format:

pythoncode

{
    "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:

bashcode

# 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

bashcode

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

bashcode

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

bashcode

--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

pythoncode

# 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

bashcode

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):

bashcode

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

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

bashcode

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

3. slime Arguments:

bashcode

# 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

pythoncode

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)

pythoncode

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:

bashcode

# 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:

bashcode

# 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:

bashcode

# 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:

bashcode

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

# Use streaming dataset
--streaming-data

---

Supported Models

Model Family	Configurations
GLM	GLM-4.5, GLM-4.6, GLM-4.7, GLM-Z1-9B
Qwen	Qwen3 (4B, 8B, 30B-A3B), Qwen3-MoE, Qwen2.5
DeepSeek	V3, V3.1, R1
Llama	Llama 3 (8B, 70B)
Others	Kimi 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:

bashcode

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

Custom Reward Model

pythoncode

# 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()

bashcode

--custom-rm-path custom_rm.py

Evaluation Multi-Task

bashcode

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

---

Resources

Documentation: https://thudm.github.io/slime/

GitHub: https://github.com/THUDM/slime

Blog: https://lmsys.org/blog/2025-07-09-slime/

Examples: See examples/ directory for 14+ worked examples

Post Training Slime

Interactive Terminal Locked

slime: LLM Post-Training Framework for RL Scaling

When to Use slime

Key Features

Architecture Overview

Installation

From Source

Quick Start: GRPO Training

Workflow 1: Standard GRPO Training

Prerequisites Checklist

Step 1: Prepare Data

Step 2: Configure Model

Step 3: Launch Training

Step 4: Monitor Training

Workflow 2: Asynchronous Training

When to Use Async

Launch Async Training

Async-Specific Parameters

Workflow 3: Multi-Turn Agentic Training

Prerequisites

Step 1: Define Custom Generate Function

Step 2: Launch with Custom Function

Configuration Reference

Three Argument Categories

Key Constraints

Data Buffer System

Basic Data Source

Buffered Data Source (Off-Policy)

Common Issues and Solutions

Issue: SGLang Engine Crash

Issue: Weight Sync Timeout

Issue: OOM During Training

Issue: Slow Data Loading

Supported Models

Advanced Topics

Co-location Mode

Custom Reward Model

Evaluation Multi-Task

Resources

How to Use & Getting Started

Installation Runbook & Walkthrough Locked

Unlock full access to Post Training Slime

More in research

Awesome ChatGPT Prompts

Awesome Obsidian

Awesome AI Tools