[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M73ZSCRRMVDKDXPXPW5ZH843":3,"skill-same:01M73ZSCRRMVDKDXPXPW5ZH843":69},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":39,"commitSha":40,"skillDirectory":41,"forksCount":13,"repositoryHtmlUrl":42,"repositoryHomepageUrl":6,"ownerHtmlUrl":43,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":44,"install":45,"resources":58},"01M73ZSCRRMVDKDXPXPW5ZH843","slime-rl-training",null,"Provides guidance for LLM post-training with RL using slime, a Megatron+SGLang framework. Use when training GLM models, implementing custom data generation workflows, or needing tight Megatron-LM integration for RL scaling.","gjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN","gjbvsjjvgm-sketch","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F278511911?v=4","01M71F3JXKZQ70H2KCHHTXKZDA","01M71PENEVFN19DBZFNHJPCHQ7",0,"MIT","https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN\u002Fblob\u002Fmain\u002Fcore\u002Fhermes-agent\u002Foptional-skills\u002Fmlops\u002Fslime\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: slime-rl-training\ndescription: Provides guidance for LLM post-training with RL using slime, a Megatron+SGLang framework. Use when training GLM models, implementing custom data generation workflows, or needing tight Megatron-LM integration for RL scaling.\nversion: 1.0.0\nauthor: Orchestra Research\nlicense: MIT\ndependencies: [sglang-router>=0.2.3, ray, torch>=2.0.0, transformers>=4.40.0]\nmetadata:\n  hermes:\n    tags: [Reinforcement Learning, Megatron-LM, SGLang, GRPO, Post-Training, GLM]\n\n---\n\n# slime: LLM Post-Training Framework for RL Scaling\n\nslime 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.\n\n## When to Use slime\n\n**Choose slime when you need:**\n- Megatron-LM native training with SGLang inference\n- Custom data generation workflows with flexible data buffers\n- Training GLM, Qwen3, DeepSeek V3, or Llama 3 models\n- Research-grade framework with production backing (Z.ai)\n\n**Consider alternatives when:**\n- You need enterprise-grade stability features → use **miles**\n- You want flexible backend swapping → use **verl**\n- You need PyTorch-native abstractions → use **torchforge**\n\n## Key Features\n\n- **Training**: Megatron-LM with full parallelism support (TP, PP, DP, SP)\n- **Rollout**: SGLang-based high-throughput generation with router\n- **Data Buffer**: Flexible prompt management and sample storage\n- **Models**: GLM-4.x, Qwen3, DeepSeek V3\u002FR1, Llama 3\n\n## Architecture Overview\n\n```\n┌─────────────────────────────────────────────────────────┐\n│                    Data Buffer                          │\n│ - Prompt initialization and management                  │\n│ - Custom data generation and filtering                  │\n│ - Rollout sample storage                                │\n└─────────────┬───────────────────────────┬───────────────┘\n              │                           │\n┌─────────────▼───────────┐ ┌─────────────▼───────────────┐\n│ Training (Megatron-LM)  │ │ Rollout (SGLang + Router)   │\n│ - Actor model training  │ │ - Response generation       │\n│ - Critic (optional)     │ │ - Reward\u002Fverifier output    │\n│ - Weight sync to rollout│ │ - Multi-turn support        │\n└─────────────────────────┘ └─────────────────────────────┘\n```\n\n## Installation\n\n```bash\n# Recommended: Docker\ndocker pull slimerl\u002Fslime:latest\ndocker run --rm --gpus all --ipc=host --shm-size=16g \\\n  -it slimerl\u002Fslime:latest \u002Fbin\u002Fbash\n\n# Inside container\ncd \u002Froot\u002Fslime && pip install -e . --no-deps\n```\n\n### From Source\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002FTHUDM\u002Fslime.git\ncd slime\npip install -r requirements.txt\npip install -e .\n```\n\n## Quick Start: GRPO Training\n\n```bash\n# Source model configuration\nsource scripts\u002Fmodels\u002Fqwen3-4B.sh\n\n# Launch training\npython train.py \\\n    --actor-num-nodes 1 \\\n    --actor-num-gpus-per-node 4 \\\n    --rollout-num-gpus 4 \\\n    --advantage-estimator grpo \\\n    --use-kl-loss --kl-loss-coef 0.001 \\\n    --rollout-batch-size 32 \\\n    --n-samples-per-prompt 8 \\\n    --global-batch-size 256 \\\n    --num-rollout 3000 \\\n    --prompt-data \u002Fpath\u002Fto\u002Fdata.jsonl \\\n    ${MODEL_ARGS[@]} ${CKPT_ARGS[@]}\n```\n\n---\n\n## Workflow 1: Standard GRPO Training\n\nUse this workflow for training reasoning models with group-relative advantages.\n\n### Prerequisites Checklist\n- [ ] Docker environment or Megatron-LM + SGLang installed\n- [ ] Model checkpoint (HuggingFace or Megatron format)\n- [ ] Training data in JSONL format\n\n### Step 1: Prepare Data\n\n```python\n# data.jsonl format\n{\"prompt\": \"What is 2 + 2?\", \"label\": \"4\"}\n{\"prompt\": \"Solve: 3x = 12\", \"label\": \"x = 4\"}\n```\n\nOr with chat format:\n```python\n{\n    \"prompt\": [\n        {\"role\": \"system\", \"content\": \"You are a math tutor.\"},\n        {\"role\": \"user\", \"content\": \"What is 15 + 27?\"}\n    ],\n    \"label\": \"42\"\n}\n```\n\n### Step 2: Configure Model\n\nChoose a pre-configured model script:\n\n```bash\n# List available models\nls scripts\u002Fmodels\u002F\n# glm4-9B.sh, qwen3-4B.sh, qwen3-30B-A3B.sh, deepseek-v3.sh, llama3-8B.sh, ...\n\n# Source your model\nsource scripts\u002Fmodels\u002Fqwen3-4B.sh\n```\n\n### Step 3: Launch Training\n\n```bash\npython train.py \\\n    --actor-num-nodes 1 \\\n    --actor-num-gpus-per-node 8 \\\n    --rollout-num-gpus 8 \\\n    --advantage-estimator grpo \\\n    --use-kl-loss \\\n    --kl-loss-coef 0.001 \\\n    --prompt-data \u002Fpath\u002Fto\u002Ftrain.jsonl \\\n    --input-key prompt \\\n    --label-key label \\\n    --apply-chat-template \\\n    --rollout-batch-size 32 \\\n    --n-samples-per-prompt 8 \\\n    --global-batch-size 256 \\\n    --num-rollout 3000 \\\n    --save-interval 100 \\\n    --eval-interval 50 \\\n    ${MODEL_ARGS[@]}\n```\n\n### Step 4: Monitor Training\n- [ ] Check TensorBoard: `tensorboard --logdir outputs\u002F`\n- [ ] Verify reward curves are increasing\n- [ ] Monitor GPU utilization across nodes\n\n---\n\n## Workflow 2: Asynchronous Training\n\nUse async mode for higher throughput by overlapping rollout and training.\n\n### When to Use Async\n- Large models with long generation times\n- High GPU idle time in synchronous mode\n- Sufficient memory for buffering\n\n### Launch Async Training\n\n```bash\npython train_async.py \\\n    --actor-num-nodes 1 \\\n    --actor-num-gpus-per-node 8 \\\n    --rollout-num-gpus 8 \\\n    --advantage-estimator grpo \\\n    --async-buffer-size 4 \\\n    --prompt-data \u002Fpath\u002Fto\u002Ftrain.jsonl \\\n    ${MODEL_ARGS[@]}\n```\n\n### Async-Specific Parameters\n\n```bash\n--async-buffer-size 4        # Number of rollouts to buffer\n--update-weights-interval 2  # Sync weights every N rollouts\n```\n\n---\n\n## Workflow 3: Multi-Turn Agentic Training\n\nUse this workflow for training agents with tool use or multi-step reasoning.\n\n### Prerequisites\n- [ ] Custom generate function for multi-turn logic\n- [ ] Tool\u002Fenvironment interface\n\n### Step 1: Define Custom Generate Function\n\n```python\n# custom_generate.py\nasync def custom_generate(args, samples, evaluation=False):\n    \"\"\"Multi-turn generation with tool calling.\"\"\"\n    for sample in samples:\n        conversation = sample.prompt\n\n        for turn in range(args.max_turns):\n            # Generate response\n            response = await generate_single(conversation)\n\n            # Check for tool call\n            tool_call = extract_tool_call(response)\n            if tool_call:\n                tool_result = execute_tool(tool_call)\n                conversation.append({\"role\": \"assistant\", \"content\": response})\n                conversation.append({\"role\": \"tool\", \"content\": tool_result})\n            else:\n                break\n\n        sample.response = response\n        sample.reward = compute_reward(sample)\n\n    return samples\n```\n\n### Step 2: Launch with Custom Function\n\n```bash\npython train.py \\\n    --custom-generate-function-path custom_generate.py \\\n    --max-turns 5 \\\n    --prompt-data \u002Fpath\u002Fto\u002Fagent_data.jsonl \\\n    ${MODEL_ARGS[@]}\n```\n\nSee `examples\u002Fsearch-r1\u002F` for a complete multi-turn search example.\n\n---\n\n## Configuration Reference\n\n### Three Argument Categories\n\nslime uses three types of arguments:\n\n**1. Megatron Arguments** (passed directly):\n```bash\n--tensor-model-parallel-size 2\n--pipeline-model-parallel-size 1\n--num-layers 32\n--hidden-size 4096\n```\n\n**2. SGLang Arguments** (prefixed with `--sglang-`):\n```bash\n--sglang-mem-fraction-static 0.8\n--sglang-context-length 8192\n--sglang-log-level INFO\n```\n\n**3. slime Arguments**:\n```bash\n# Resource allocation\n--actor-num-nodes 1\n--actor-num-gpus-per-node 8\n--rollout-num-gpus 8\n--colocate  # Share GPUs between training\u002Finference\n\n# Data\n--prompt-data \u002Fpath\u002Fto\u002Fdata.jsonl\n--input-key prompt\n--label-key label\n\n# Training loop\n--num-rollout 3000\n--rollout-batch-size 32\n--n-samples-per-prompt 8\n--global-batch-size 256\n\n# Algorithm\n--advantage-estimator grpo  # or: gspo, ppo, reinforce_plus_plus\n--use-kl-loss\n--kl-loss-coef 0.001\n```\n\n### Key Constraints\n\n```\nrollout_batch_size × n_samples_per_prompt = global_batch_size × num_steps_per_rollout\n```\n\nExample: 32 × 8 = 256 × 1\n\n---\n\n## Data Buffer System\n\nslime's data buffer enables flexible data management:\n\n### Basic Data Source\n\n```python\nclass RolloutDataSource:\n    def get_samples(self, num_samples):\n        \"\"\"Fetch prompts from dataset.\"\"\"\n        return self.dataset.sample(num_samples)\n\n    def add_samples(self, samples):\n        \"\"\"Called after generation (no-op by default).\"\"\"\n        pass\n```\n\n### Buffered Data Source (Off-Policy)\n\n```python\nclass RolloutDataSourceWithBuffer(RolloutDataSource):\n    def __init__(self):\n        self.buffer = []\n\n    def add_samples(self, samples):\n        \"\"\"Store generated samples for reuse.\"\"\"\n        self.buffer.extend(samples)\n\n    def buffer_filter(self, args, buffer, num_samples):\n        \"\"\"Custom selection logic (prioritized, stratified, etc.).\"\"\"\n        return select_best(buffer, num_samples)\n```\n\n---\n\n## Common Issues and Solutions\n\n### Issue: SGLang Engine Crash\n\n**Symptoms**: Inference engine dies mid-training\n\n**Solutions**:\n```bash\n# Enable fault tolerance\n--use-fault-tolerance\n\n# Increase memory allocation\n--sglang-mem-fraction-static 0.85\n\n# Reduce batch size\n--rollout-batch-size 16\n```\n\n### Issue: Weight Sync Timeout\n\n**Symptoms**: Training hangs after rollout\n\n**Solutions**:\n```bash\n# Increase sync interval\n--update-weights-interval 5\n\n# Use colocated mode (no network transfer)\n--colocate\n```\n\n### Issue: OOM During Training\n\n**Symptoms**: CUDA OOM in backward pass\n\n**Solutions**:\n```bash\n# Enable gradient checkpointing\n--recompute-activations\n\n# Reduce micro-batch size\n--micro-batch-size 1\n\n# Enable sequence parallelism\n--sequence-parallel\n```\n\n### Issue: Slow Data Loading\n\n**Symptoms**: GPU idle during data fetch\n\n**Solutions**:\n```bash\n# Increase data workers\n--num-data-workers 4\n\n# Use streaming dataset\n--streaming-data\n```\n\n---\n\n## Supported Models\n\n| Model Family | Configurations |\n|--------------|----------------|\n| GLM | GLM-4.5, GLM-4.6, GLM-4.7, GLM-Z1-9B |\n| Qwen | Qwen3 (4B, 8B, 30B-A3B), Qwen3-MoE, Qwen2.5 |\n| DeepSeek | V3, V3.1, R1 |\n| Llama | Llama 3 (8B, 70B) |\n| Others | Kimi K2, Moonlight-16B |\n\nEach model has pre-configured scripts in `scripts\u002Fmodels\u002F`.\n\n---\n\n## Advanced Topics\n\n### Co-location Mode\n\nShare GPUs between training and inference to reduce memory:\n\n```bash\npython train.py \\\n    --colocate \\\n    --actor-num-gpus-per-node 8 \\\n    --sglang-mem-fraction-static 0.4 \\\n    ${MODEL_ARGS[@]}\n```\n\n### Custom Reward Model\n\n```python\n# custom_rm.py\nclass CustomRewardModel:\n    def __init__(self, model_path):\n        self.model = load_model(model_path)\n\n    def compute_reward(self, prompts, responses):\n        inputs = self.tokenize(prompts, responses)\n        scores = self.model(inputs)\n        return scores.tolist()\n```\n\n```bash\n--custom-rm-path custom_rm.py\n```\n\n### Evaluation Multi-Task\n\n```bash\n--eval-prompt-data aime \u002Fpath\u002Fto\u002Faime.jsonl \\\n--eval-prompt-data gsm8k \u002Fpath\u002Fto\u002Fgsm8k.jsonl \\\n--n-samples-per-eval-prompt 16\n```\n\n---\n\n## Resources\n\n- **Documentation**: https:\u002F\u002Fthudm.github.io\u002Fslime\u002F\n- **GitHub**: https:\u002F\u002Fgithub.com\u002FTHUDM\u002Fslime\n- **Blog**: https:\u002F\u002Flmsys.org\u002Fblog\u002F2025-07-09-slime\u002F\n- **Examples**: See `examples\u002F` directory for 14+ worked examples\n\n","# slime: LLM Post-Training Framework for RL Scaling\n\nslime 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.\n\n## When to Use slime\n\n**Choose slime when you need:**\n- Megatron-LM native training with SGLang inference\n- Custom data generation workflows with flexible data buffers\n- Training GLM, Qwen3, DeepSeek V3, or Llama 3 models\n- Research-grade framework with production backing (Z.ai)\n\n**Consider alternatives when:**\n- You need enterprise-grade stability features → use **miles**\n- You want flexible backend swapping → use **verl**\n- You need PyTorch-native abstractions → use **torchforge**\n\n## Key Features\n\n- **Training**: Megatron-LM with full parallelism support (TP, PP, DP, SP)\n- **Rollout**: SGLang-based high-throughput generation with router\n- **Data Buffer**: Flexible prompt management and sample storage\n- **Models**: GLM-4.x, Qwen3, DeepSeek V3\u002FR1, Llama 3\n\n## Architecture Overview\n\n```\n┌─────────────────────────────────────────────────────────┐\n│                    Data Buffer                          │\n│ - Prompt initialization and management                  │\n│ - Custom data generation and filtering                  │\n│ - Rollout sample storage                                │\n└─────────────┬───────────────────────────┬───────────────┘\n              │                           │\n┌─────────────▼───────────┐ ┌─────────────▼───────────────┐\n│ Training (Megatron-LM)  │ │ Rollout (SGLang + Router)   │\n│ - Actor model training  │ │ - Response generation       │\n│ - Critic (optional)     │ │ - Reward\u002Fverifier output    │\n│ - Weight sync to rollout│ │ - Multi-turn support        │\n└─────────────────────────┘ └─────────────────────────────┘\n```\n\n## Installation\n\n```bash\n# Recommended: Docker\ndocker pull slimerl\u002Fslime:latest\ndocker run --rm --gpus all --ipc=host --shm-size=16g \\\n  -it slimerl\u002Fslime:latest \u002Fbin\u002Fbash\n\n# Inside container\ncd \u002Froot\u002Fslime && pip install -e . --no-deps\n```\n\n### From Source\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002FTHUDM\u002Fslime.git\ncd slime\npip install -r requirements.txt\npip install -e .\n```\n\n## Quick Start: GRPO Training\n\n```bash\n# Source model configuration\nsource scripts\u002Fmodels\u002Fqwen3-4B.sh\n\n# Launch training\npython train.py \\\n    --actor-num-nodes 1 \\\n    --actor-num-gpus-per-node 4 \\\n    --rollout-num-gpus 4 \\\n    --advantage-estimator grpo \\\n    --use-kl-loss --kl-loss-coef 0.001 \\\n    --rollout-batch-size 32 \\\n    --n-samples-per-prompt 8 \\\n    --global-batch-size 256 \\\n    --num-rollout 3000 \\\n    --prompt-data \u002Fpath\u002Fto\u002Fdata.jsonl \\\n    ${MODEL_ARGS[@]} ${CKPT_ARGS[@]}\n```\n\n---\n\n## Workflow 1: Standard GRPO Training\n\nUse this workflow for training reasoning models with group-relative advantages.\n\n### Prerequisites Checklist\n- [ ] Docker environment or Megatron-LM + SGLang installed\n- [ ] Model checkpoint (HuggingFace or Megatron format)\n- [ ] Training data in JSONL format\n\n### Step 1: Prepare Data\n\n```python\n# data.jsonl format\n{\"prompt\": \"What is 2 + 2?\", \"label\": \"4\"}\n{\"prompt\": \"Solve: 3x = 12\", \"label\": \"x = 4\"}\n```\n\nOr with chat format:\n```python\n{\n    \"prompt\": [\n        {\"role\": \"system\", \"content\": \"You are a math tutor.\"},\n        {\"role\": \"user\", \"content\": \"What is 15 + 27?\"}\n    ],\n    \"label\": \"42\"\n}\n```\n\n### Step 2: Configure Model\n\nChoose a pre-configured model script:\n\n```bash\n# List available models\nls scripts\u002Fmodels\u002F\n# glm4-9B.sh, qwen3-4B.sh, qwen3-30B-A3B.sh, deepseek-v3.sh, llama3-8B.sh, ...\n\n# Source your model\nsource scripts\u002Fmodels\u002Fqwen3-4B.sh\n```\n\n### Step 3: Launch Training\n\n```bash\npython train.py \\\n    --actor-num-nodes 1 \\\n    --actor-num-gpus-per-node 8 \\\n    --rollout-num-gpus 8 \\\n    --advantage-estimator grpo \\\n    --use-kl-loss \\\n    --kl-loss-coef 0.001 \\\n    --prompt-data \u002Fpath\u002Fto\u002Ftrain.jsonl \\\n    --input-key prompt \\\n    --label-key label \\\n    --apply-chat-template \\\n    --rollout-batch-size 32 \\\n    --n-samples-per-prompt 8 \\\n    --global-batch-size 256 \\\n    --num-rollout 3000 \\\n    --save-interval 100 \\\n    --eval-interval 50 \\\n    ${MODEL_ARGS[@]}\n```\n\n### Step 4: Monitor Training\n- [ ] Check TensorBoard: `tensorboard --logdir outputs\u002F`\n- [ ] Verify reward curves are increasing\n- [ ] Monitor GPU utilization across nodes\n\n---\n\n## Workflow 2: Asynchronous Training\n\nUse async mode for higher throughput by overlapping rollout and training.\n\n### When to Use Async\n- Large models with long generation times\n- High GPU idle time in synchronous mode\n- Sufficient memory for buffering\n\n### Launch Async Training\n\n```bash\npython train_async.py \\\n    --actor-num-nodes 1 \\\n    --actor-num-gpus-per-node 8 \\\n    --rollout-num-gpus 8 \\\n    --advantage-estimator grpo \\\n    --async-buffer-size 4 \\\n    --prompt-data \u002Fpath\u002Fto\u002Ftrain.jsonl \\\n    ${MODEL_ARGS[@]}\n```\n\n### Async-Specific Parameters\n\n```bash\n--async-buffer-size 4        # Number of rollouts to buffer\n--update-weights-interval 2  # Sync weights every N rollouts\n```\n\n---\n\n## Workflow 3: Multi-Turn Agentic Training\n\nUse this workflow for training agents with tool use or multi-step reasoning.\n\n### Prerequisites\n- [ ] Custom generate function for multi-turn logic\n- [ ] Tool\u002Fenvironment interface\n\n### Step 1: Define Custom Generate Function\n\n```python\n# custom_generate.py\nasync def custom_generate(args, samples, evaluation=False):\n    \"\"\"Multi-turn generation with tool calling.\"\"\"\n    for sample in samples:\n        conversation = sample.prompt\n\n        for turn in range(args.max_turns):\n            # Generate response\n            response = await generate_single(conversation)\n\n            # Check for tool call\n            tool_call = extract_tool_call(response)\n            if tool_call:\n                tool_result = execute_tool(tool_call)\n                conversation.append({\"role\": \"assistant\", \"content\": response})\n                conversation.append({\"role\": \"tool\", \"content\": tool_result})\n            else:\n                break\n\n        sample.response = response\n        sample.reward = compute_reward(sample)\n\n    return samples\n```\n\n### Step 2: Launch with Custom Function\n\n```bash\npython train.py \\\n    --custom-generate-function-path custom_generate.py \\\n    --max-turns 5 \\\n    --prompt-data \u002Fpath\u002Fto\u002Fagent_data.jsonl \\\n    ${MODEL_ARGS[@]}\n```\n\nSee `examples\u002Fsearch-r1\u002F` for a complete multi-turn search example.\n\n---\n\n## Configuration Reference\n\n### Three Argument Categories\n\nslime uses three types of arguments:\n\n**1. Megatron Arguments** (passed directly):\n```bash\n--tensor-model-parallel-size 2\n--pipeline-model-parallel-size 1\n--num-layers 32\n--hidden-size 4096\n```\n\n**2. SGLang Arguments** (prefixed with `--sglang-`):\n```bash\n--sglang-mem-fraction-static 0.8\n--sglang-context-length 8192\n--sglang-log-level INFO\n```\n\n**3. slime Arguments**:\n```bash\n# Resource allocation\n--actor-num-nodes 1\n--actor-num-gpus-per-node 8\n--rollout-num-gpus 8\n--colocate  # Share GPUs between training\u002Finference\n\n# Data\n--prompt-data \u002Fpath\u002Fto\u002Fdata.jsonl\n--input-key prompt\n--label-key label\n\n# Training loop\n--num-rollout 3000\n--rollout-batch-size 32\n--n-samples-per-prompt 8\n--global-batch-size 256\n\n# Algorithm\n--advantage-estimator grpo  # or: gspo, ppo, reinforce_plus_plus\n--use-kl-loss\n--kl-loss-coef 0.001\n```\n\n### Key Constraints\n\n```\nrollout_batch_size × n_samples_per_prompt = global_batch_size × num_steps_per_rollout\n```\n\nExample: 32 × 8 = 256 × 1\n\n---\n\n## Data Buffer System\n\nslime's data buffer enables flexible data management:\n\n### Basic Data Source\n\n```python\nclass RolloutDataSource:\n    def get_samples(self, num_samples):\n        \"\"\"Fetch prompts from dataset.\"\"\"\n        return self.dataset.sample(num_samples)\n\n    def add_samples(self, samples):\n        \"\"\"Called after generation (no-op by default).\"\"\"\n        pass\n```\n\n### Buffered Data Source (Off-Policy)\n\n```python\nclass RolloutDataSourceWithBuffer(RolloutDataSource):\n    def __init__(self):\n        self.buffer = []\n\n    def add_samples(self, samples):\n        \"\"\"Store generated samples for reuse.\"\"\"\n        self.buffer.extend(samples)\n\n    def buffer_filter(self, args, buffer, num_samples):\n        \"\"\"Custom selection logic (prioritized, stratified, etc.).\"\"\"\n        return select_best(buffer, num_samples)\n```\n\n---\n\n## Common Issues and Solutions\n\n### Issue: SGLang Engine Crash\n\n**Symptoms**: Inference engine dies mid-training\n\n**Solutions**:\n```bash\n# Enable fault tolerance\n--use-fault-tolerance\n\n# Increase memory allocation\n--sglang-mem-fraction-static 0.85\n\n# Reduce batch size\n--rollout-batch-size 16\n```\n\n### Issue: Weight Sync Timeout\n\n**Symptoms**: Training hangs after rollout\n\n**Solutions**:\n```bash\n# Increase sync interval\n--update-weights-interval 5\n\n# Use colocated mode (no network transfer)\n--colocate\n```\n\n### Issue: OOM During Training\n\n**Symptoms**: CUDA OOM in backward pass\n\n**Solutions**:\n```bash\n# Enable gradient checkpointing\n--recompute-activations\n\n# Reduce micro-batch size\n--micro-batch-size 1\n\n# Enable sequence parallelism\n--sequence-parallel\n```\n\n### Issue: Slow Data Loading\n\n**Symptoms**: GPU idle during data fetch\n\n**Solutions**:\n```bash\n# Increase data workers\n--num-data-workers 4\n\n# Use streaming dataset\n--streaming-data\n```\n\n---\n\n## Supported Models\n\n| Model Family | Configurations |\n|--------------|----------------|\n| GLM | GLM-4.5, GLM-4.6, GLM-4.7, GLM-Z1-9B |\n| Qwen | Qwen3 (4B, 8B, 30B-A3B), Qwen3-MoE, Qwen2.5 |\n| DeepSeek | V3, V3.1, R1 |\n| Llama | Llama 3 (8B, 70B) |\n| Others | Kimi K2, Moonlight-16B |\n\nEach model has pre-configured scripts in `scripts\u002Fmodels\u002F`.\n\n---\n\n## Advanced Topics\n\n### Co-location Mode\n\nShare GPUs between training and inference to reduce memory:\n\n```bash\npython train.py \\\n    --colocate \\\n    --actor-num-gpus-per-node 8 \\\n    --sglang-mem-fraction-static 0.4 \\\n    ${MODEL_ARGS[@]}\n```\n\n### Custom Reward Model\n\n```python\n# custom_rm.py\nclass CustomRewardModel:\n    def __init__(self, model_path):\n        self.model = load_model(model_path)\n\n    def compute_reward(self, prompts, responses):\n        inputs = self.tokenize(prompts, responses)\n        scores = self.model(inputs)\n        return scores.tolist()\n```\n\n```bash\n--custom-rm-path custom_rm.py\n```\n\n### Evaluation Multi-Task\n\n```bash\n--eval-prompt-data aime \u002Fpath\u002Fto\u002Faime.jsonl \\\n--eval-prompt-data gsm8k \u002Fpath\u002Fto\u002Fgsm8k.jsonl \\\n--n-samples-per-eval-prompt 16\n```\n\n---\n\n## Resources\n\n- **Documentation**: https:\u002F\u002Fthudm.github.io\u002Fslime\u002F\n- **GitHub**: https:\u002F\u002Fgithub.com\u002FTHUDM\u002Fslime\n- **Blog**: https:\u002F\u002Flmsys.org\u002Fblog\u002F2025-07-09-slime\u002F\n- **Examples**: See `examples\u002F` directory for 14+ worked examples\n\n",{"name":5,"author":23,"license":14,"version":24,"metadata":25,"description":7,"dependencies":34},"Orchestra Research","1.0.0",{"hermes":26},{"tags":27},[28,29,30,31,32,33],"Reinforcement Learning","Megatron-LM","SGLang","GRPO","Post-Training","GLM",[35,36,37,38],"sglang-router>=0.2.3","ray","torch>=2.0.0","transformers>=4.40.0","main","6746276ba18cb774853eb61b3b9399658bc9bc5c","core\u002Fhermes-agent\u002Foptional-skills\u002Fmlops\u002Fslime","https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN","https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch","2026-04-28 17:01:03",{"source":8,"skillName":5,"commands":46,"prompt":56,"zipAvailable":57},[47,50,53],{"tool":48,"command":49},"npx","npx skills add gjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN --skill slime-rl-training",{"tool":51,"command":52},"bunx","bunx skills add gjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN --skill slime-rl-training",{"tool":54,"command":55},"pnpm","pnpm dlx skills add gjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN --skill slime-rl-training","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN\u002Fblob\u002Fmain\u002Fcore\u002Fhermes-agent\u002Foptional-skills\u002Fmlops\u002Fslime\u002FSKILL.md",true,[59,65],{"path":60,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":63,"sourceHtmlUrl":64},"references\u002Fapi-reference.md","REFERENCE","参考文档",11907,"https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN\u002Fblob\u002Fmain\u002Fcore\u002Fhermes-agent\u002Foptional-skills\u002Fmlops\u002Fslime\u002Freferences\u002Fapi-reference.md",{"path":66,"resourceType":61,"resourceTypeDesc":62,"sizeBytes":67,"sourceHtmlUrl":68},"references\u002Ftroubleshooting.md",7221,"https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN\u002Fblob\u002Fmain\u002Fcore\u002Fhermes-agent\u002Foptional-skills\u002Fmlops\u002Fslime\u002Freferences\u002Ftroubleshooting.md",[70,114,139,156,175,194],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":92,"publishedAt":113},"01M71QG27G1R2H2ZWEG4S0WTZT","writing-plans","Write implementation plans: bite-sized tasks, paths, code.","https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fsoftware-development\u002Fwriting-plans\u002FSKILL.md",[76,81,84,89],{"categoryNo":77,"categoryName":78,"taxonomyType":79,"taxonomyTypeDesc":80},"01JBSEEDCAT000000000000015","视觉设计","SCENARIO","使用场景",{"categoryNo":82,"categoryName":83,"taxonomyType":79,"taxonomyTypeDesc":80},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":85,"categoryName":86,"taxonomyType":87,"taxonomyTypeDesc":88},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",{"categoryNo":90,"categoryName":91,"taxonomyType":87,"taxonomyTypeDesc":88},"01JBSEEDCAT000000000000036","内容创作者",[93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112],"writing","writing plans","plans","test-driven","test-driven-development","superpowers","design","workflow","code","tasks","subagent","code-review","review","documentation","write","paths","planning","development","test","implementation","2026-11-10 13:02:02",{"skillNo":115,"name":97,"nameZh":6,"description":116,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":117,"categories":118,"keywords":125,"publishedAt":113},"01M71QFZYRTJXMAE551MRAKZ14","TDD: enforce RED-GREEN-REFACTOR, tests before code.","https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fsoftware-development\u002Ftest-driven-development\u002FSKILL.md",[119,122],{"categoryNo":120,"categoryName":121,"taxonomyType":79,"taxonomyTypeDesc":80},"01JBSEEDCAT000000000000007","代码审查",{"categoryNo":123,"categoryName":124,"taxonomyType":79,"taxonomyTypeDesc":80},"01JBSEEDCAT000000000000008","测试与质量",[126,110,111,127,128,129,130,98,93,131,101,103,132,133,134,135,136,95,137,138],"test driven development","driven","green","red-green-refactor","testing","debugging","systematic","refactor","systematic-debugging","enforce","tests","tdd","quality",{"skillNo":140,"name":134,"nameZh":6,"description":141,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":142,"categories":143,"keywords":145,"publishedAt":113},"01M71QFWR5Q8EYEQ6MCS3SM6MC","4-phase root cause debugging: understand bugs before fixing.","https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fsoftware-development\u002Fsystematic-debugging\u002FSKILL.md",[144],{"categoryNo":123,"categoryName":124,"taxonomyType":79,"taxonomyTypeDesc":80},[146,131,132,96,147,148,97,98,93,103,149,150,110,151,95,152,111,153,154,155],"systematic debugging","problem","problem-solving","phase","bugs","root cause","root-cause","investigation","understand","hermes",{"skillNo":157,"name":158,"nameZh":6,"description":159,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":160,"categories":161,"keywords":167,"publishedAt":113},"01M71QFS9S04MD8JP747W5WM0N","subagent-driven-development","Execute plans via delegate_task subagents (2-stage review).","https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fsoftware-development\u002Fsubagent-driven-development\u002FSKILL.md",[162,163,166],{"categoryNo":82,"categoryName":83,"taxonomyType":79,"taxonomyTypeDesc":80},{"categoryNo":164,"categoryName":165,"taxonomyType":79,"taxonomyTypeDesc":80},"01JBSEEDCAT000000000000023","Skill与智能体开发",{"categoryNo":90,"categoryName":91,"taxonomyType":87,"taxonomyTypeDesc":88},[168,103,110,127,96,97,169,170,98,93,100,101,104,105,171,172,95,173,174,111],"subagent driven development","delegate","delegation","subagents","parallel","stage","task",{"skillNo":176,"name":177,"nameZh":6,"description":178,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":179,"categories":180,"keywords":183,"publishedAt":113},"01M71QFQ1DBCVAZ511MXFE6T2V","requesting-code-review","Pre-commit review: security scan, quality gates, auto-fix.","https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fsoftware-development\u002Frequesting-code-review\u002FSKILL.md",[181,182],{"categoryNo":120,"categoryName":121,"taxonomyType":79,"taxonomyTypeDesc":80},{"categoryNo":85,"categoryName":86,"taxonomyType":87,"taxonomyTypeDesc":88},[184,185,101,105,96,97,186,187,98,188,93,189,190,103,104,191,192,110,193,95],"code review","requesting code review","fix","auto-fix","auto","security scan","commit","verification","security","scan",{"skillNo":195,"name":196,"nameZh":6,"description":197,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":198,"categories":199,"keywords":201,"publishedAt":113},"01M71QFM7KWM540YR67FRN807Q","python-debugpy","Debug Python: pdb REPL + debugpy remote (DAP).","https:\u002F\u002Fgithub.com\u002Fgjbvsjjvgm-sketch\u002FWORM_V2_SOVEREIGN\u002Fblob\u002Fmain\u002Fcore\u002Fskills\u002Fsoftware-development\u002Fpython-debugpy\u002FSKILL.md",[200],{"categoryNo":85,"categoryName":86,"taxonomyType":87,"taxonomyTypeDesc":88},[202,203,204,205,206,207,208,209,131,132,210,211,212,134,213,214,215,216,155,217],"debugpy","python debugpy","python","dap","inspect","post-mortem","pdb","remote","debugger","node","commands","debug","post","breakpoints","node-inspect","tui"]