[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6ADXERQATWFVGSQDDG0QN6W":3,"skill-same:01M6ADXERQATWFVGSQDDG0QN6W":111},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":45,"linkOnly":46,"markdownContent":47,"markdownBody":48,"frontMatter":49,"refName":61,"commitSha":62,"skillDirectory":63,"forksCount":64,"repositoryHtmlUrl":65,"repositoryHomepageUrl":66,"ownerHtmlUrl":67,"ownerBlogUrl":68,"ownerDescription":69,"updatedAt":70,"install":71,"resources":84},"01M6ADXERQATWFVGSQDDG0QN6W","post-training-verl",null,"Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.","davila7\u002Fclaude-code-templates","davila7","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F6216945?v=4","01M6ADEWWFPS191NFM0Y1EX1JH","Daniel Avila","01M6ADJHGM78PHQRD3XZTFDVKK",33481,"MIT","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fai-research\u002Fpost-training-verl\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44],"training","verl","post","post training verl","when","llms","for","implementing","orchestra","at","volcano","algorithms","use","learning","Transformers","vLLM","Ray","PPO","RLHF","Post-Training","GRPO","Distributed Training","reinforcement learning","infrastructure","distributed","research","2026-11-10 10:12:30",false,"---\nname: post-training-verl\ndescription: Provides guidance for training LLMs with reinforcement learning using verl (Volcano Engine RL). Use when implementing RLHF, GRPO, PPO, or other RL algorithms for LLM post-training at scale with flexible infrastructure backends.\nlicense: MIT\nmetadata:\n  version: 1.0.0\n  author: Orchestra Research\n  tags: [Reinforcement Learning, RLHF, GRPO, PPO, Post-Training, Distributed Training]\n  dependencies: [verl>=0.3.0, torch>=2.0.0, ray>=2.41.0, vllm>=0.8.2, transformers>=4.40.0]\n---\n\n# verl: Volcano Engine Reinforcement Learning for LLMs\n\nverl is a flexible, efficient, and production-ready RL training library for large language models from ByteDance's Seed team. It implements the HybridFlow framework (EuroSys 2025) and powers models like Doubao-1.5-pro achieving O1-level performance on math benchmarks.\n\n## When to Use verl\n\n**Choose verl when you need:**\n- Production-ready RL training at scale (tested up to 671B parameters)\n- Flexibility to swap backends (FSDP ↔ Megatron-LM ↔ vLLM ↔ SGLang)\n- Support for multiple RL algorithms (PPO, GRPO, RLOO, REINFORCE++, DAPO)\n- Multi-turn rollout with tool calling for agentic workflows\n- Vision-language model RL training\n\n**Consider alternatives when:**\n- You need Megatron-native training → use **slime** or **miles**\n- You want PyTorch-native abstractions with Monarch → use **torchforge**\n- You only need simple SFT\u002FDPO → use **TRL** or **Axolotl**\n\n## Key Features\n\n- **Training backends**: FSDP, FSDP2, Megatron-LM\n- **Rollout engines**: vLLM, SGLang, HuggingFace Transformers\n- **Algorithms**: PPO, GRPO, DAPO, RLOO, ReMax, REINFORCE++, SPIN, SPPO\n- **Models**: Qwen-3, Llama-3.1, DeepSeek, Gemma-2 (0.5B to 671B)\n- **Advanced**: LoRA RL, sequence parallelism, expert parallelism, multi-turn tools\n\n## Installation\n\n```bash\n# Option 1: pip install\npip install verl[vllm]  # or verl[sglang] for SGLang backend\n\n# Option 2: Docker (recommended for production)\ndocker pull verlai\u002Fverl:vllm011.latest\n\n# Option 3: From source\ngit clone https:\u002F\u002Fgithub.com\u002Fvolcengine\u002Fverl.git\ncd verl && pip install -e .[vllm,math]\n```\n\n## Quick Start: GRPO Training\n\n```bash\npython3 -m verl.trainer.main_ppo \\\n    algorithm.adv_estimator=grpo \\\n    data.train_files=~\u002Fdata\u002Fgsm8k\u002Ftrain.parquet \\\n    actor_rollout_ref.model.path=Qwen\u002FQwen2.5-7B \\\n    actor_rollout_ref.rollout.n=8 \\\n    actor_rollout_ref.actor.use_kl_loss=True \\\n    trainer.n_gpus_per_node=8\n```\n\n## Core Architecture\n\nverl uses a **HybridFlow** programming model separating control flow from computation:\n\n```\n┌─────────────────────────────────────────────────────────┐\n│ Single-Process Controller (Ray)                         │\n│ - Orchestrates: rollout → reward → train → sync        │\n└─────────────────────┬───────────────────────────────────┘\n                      │\n┌─────────────────────▼───────────────────────────────────┐\n│ Multi-Process Workers                                   │\n│ ├── ActorRolloutRefWorker (policy + generation)        │\n│ ├── CriticWorker (value estimation, PPO only)          │\n│ └── RewardManager (model-based or rule-based rewards)  │\n└─────────────────────────────────────────────────────────┘\n```\n\n---\n\n## Workflow 1: Math Reasoning with GRPO\n\nUse this workflow for training reasoning models on math tasks like GSM8K or MATH.\n\n### Prerequisites Checklist\n- [ ] GPU cluster with 8+ GPUs (H100 recommended)\n- [ ] Dataset in parquet format with `prompt` and `reward_model` columns\n- [ ] Base model from HuggingFace Hub\n\n### Step 1: Prepare Dataset\n\n```python\nimport pandas as pd\n\ndata = [\n    {\n        \"prompt\": [{\"role\": \"user\", \"content\": \"What is 15 + 27?\"}],\n        \"reward_model\": {\"ground_truth\": \"42\"}\n    },\n    # ... more examples\n]\ndf = pd.DataFrame(data)\ndf.to_parquet(\"train.parquet\")\n```\n\n### Step 2: Define Reward Function\n\n```python\n# reward_function.py\nimport re\n\ndef compute_reward(responses, ground_truths):\n    rewards = []\n    for response, gt in zip(responses, ground_truths):\n        # Extract answer from response\n        match = re.search(r'\\\\boxed{([^}]+)}', response)\n        if match and match.group(1).strip() == gt.strip():\n            rewards.append(1.0)\n        else:\n            rewards.append(0.0)\n    return rewards\n```\n\n### Step 3: Create Training Config\n\n```yaml\n# config\u002Fgrpo_math.yaml\nalgorithm:\n  adv_estimator: grpo\n  gamma: 1.0\n  lam: 1.0\n\ndata:\n  train_files: \u002Fpath\u002Fto\u002Ftrain.parquet\n  val_files: \u002Fpath\u002Fto\u002Fval.parquet\n  train_batch_size: 256\n  max_prompt_length: 512\n  max_response_length: 2048\n\nactor_rollout_ref:\n  model:\n    path: Qwen\u002FQwen2.5-7B-Instruct\n  actor:\n    use_kl_loss: true\n    kl_loss_coef: 0.001\n    ppo_mini_batch_size: 64\n  rollout:\n    name: vllm\n    n: 8  # samples per prompt\n    temperature: 0.7\n    top_p: 0.95\n\ntrainer:\n  total_epochs: 3\n  n_gpus_per_node: 8\n  save_freq: 100\n```\n\n### Step 4: Launch Training\n\n```bash\npython3 -m verl.trainer.main_ppo \\\n    --config-path config \\\n    --config-name grpo_math \\\n    trainer.experiment_name=grpo_math_qwen7b\n```\n\n### Step 5: Monitor and Validate\n- [ ] Check WandB\u002FTensorBoard for loss curves\n- [ ] Verify reward is increasing over steps\n- [ ] Run evaluation on held-out test set\n\n---\n\n## Workflow 2: PPO with Critic Model\n\nUse this workflow when you need value-based advantage estimation (GAE).\n\n### Key Differences from GRPO\n- Requires separate critic model\n- Uses Generalized Advantage Estimation (GAE)\n- Better for tasks with dense rewards\n\n### Configuration\n\n```yaml\nalgorithm:\n  adv_estimator: gae  # Use GAE instead of GRPO\n  gamma: 0.99\n  lam: 0.95\n\ncritic:\n  model:\n    path: Qwen\u002FQwen2.5-7B-Instruct  # Can be same or different from actor\n  ppo_mini_batch_size: 64\n\nactor_rollout_ref:\n  actor:\n    use_kl_loss: true\n    kl_loss_coef: 0.02\n    clip_ratio: 0.2  # PPO clipping\n```\n\n### Launch with Critic\n\n```bash\npython3 -m verl.trainer.main_ppo \\\n    algorithm.adv_estimator=gae \\\n    critic.model.path=Qwen\u002FQwen2.5-7B-Instruct \\\n    trainer.n_gpus_per_node=8\n```\n\n---\n\n## Workflow 3: Large-Scale Training with Megatron\n\nUse this workflow for models >70B parameters or when you need expert parallelism.\n\n### Prerequisites\n- [ ] Install Megatron-LM bridge: `pip install mbridge`\n- [ ] Convert model to Megatron format\n- [ ] Multi-node cluster with NVLink\u002FInfiniBand\n\n### Configuration for 70B+ Models\n\n```yaml\nactor_rollout_ref:\n  model:\n    path: \u002Fpath\u002Fto\u002Fmegatron\u002Fcheckpoint\n    backend: megatron\n  actor:\n    strategy: megatron\n    tensor_model_parallel_size: 8\n    pipeline_model_parallel_size: 2\n  rollout:\n    name: vllm\n    tensor_parallel_size: 8\n```\n\n### Launch Multi-Node\n\n```bash\n# On head node\nray start --head --port=6379\n\n# On worker nodes\nray start --address='head_ip:6379'\n\n# Launch training\npython3 -m verl.trainer.main_ppo \\\n    trainer.nnodes=4 \\\n    trainer.n_gpus_per_node=8\n```\n\n---\n\n## Configuration Reference\n\n### Algorithm Selection\n\n| Algorithm | `adv_estimator` | Use Case |\n|-----------|-----------------|----------|\n| GRPO | `grpo` | Critic-free, math\u002Freasoning |\n| PPO\u002FGAE | `gae` | Dense rewards, value estimation |\n| REINFORCE++ | `reinforce_plus_plus` | Variance reduction |\n| RLOO | `rloo` | Leave-one-out baseline |\n| ReMax | `remax` | Maximum reward baseline |\n| OPO | `opo` | Optimal policy optimization |\n\n### Key Parameters\n\n```yaml\n# Rollout parameters\nactor_rollout_ref.rollout.n: 8              # Samples per prompt\nactor_rollout_ref.rollout.temperature: 0.7  # Sampling temperature\nactor_rollout_ref.rollout.top_p: 0.95       # Nucleus sampling\n\n# Training parameters\nactor_rollout_ref.actor.lr: 1e-6            # Learning rate\nactor_rollout_ref.actor.ppo_mini_batch_size: 64\nactor_rollout_ref.actor.clip_ratio: 0.2     # PPO clip range\n\n# KL control\nactor_rollout_ref.actor.use_kl_loss: true\nactor_rollout_ref.actor.kl_loss_coef: 0.001\nalgorithm.kl_ctrl.target_kl: 0.1            # For adaptive KL control\n```\n\n---\n\n## Common Issues and Solutions\n\n### Issue: OOM During Rollout\n\n**Symptoms**: CUDA out of memory during generation phase\n\n**Solutions**:\n```yaml\n# Reduce batch size\nactor_rollout_ref.rollout.log_prob_micro_batch_size: 4\n\n# Enable gradient checkpointing\nactor_rollout_ref.model.enable_gradient_checkpointing: true\n\n# Use FSDP2 with CPU offloading\nactor_rollout_ref.actor.strategy: fsdp2\nactor_rollout_ref.actor.fsdp_config.offload_policy: true\n```\n\n### Issue: Training Instability\n\n**Symptoms**: Loss spikes, reward collapse\n\n**Solutions**:\n```yaml\n# Reduce learning rate\nactor_rollout_ref.actor.lr: 5e-7\n\n# Increase KL penalty\nactor_rollout_ref.actor.kl_loss_coef: 0.01\n\n# Enable gradient clipping\nactor_rollout_ref.actor.max_grad_norm: 1.0\n```\n\n### Issue: Slow Weight Sync\n\n**Symptoms**: Long pauses between rollout and training\n\n**Solutions**:\n```bash\n# Use FSDP2 for faster resharding\nactor_rollout_ref.actor.strategy=fsdp2\n\n# Enable async weight transfer\ntrainer.async_weight_update=true\n```\n\n### Issue: vLLM Version Mismatch\n\n**Symptoms**: Import errors or generation failures\n\n**Solution**: Use compatible versions:\n```bash\npip install vllm>=0.8.5,\u003C=0.12.0\n# Avoid vLLM 0.7.x (known bugs)\n```\n\n---\n\n## Advanced Topics\n\n### Multi-Turn Tool Calling\n\nSee [references\u002Fmulti-turn.md](references\u002Fmulti-turn.md) for agentic workflows with tool use.\n\n### Vision-Language Models\n\n```yaml\nactor_rollout_ref:\n  model:\n    path: Qwen\u002FQwen2.5-VL-7B-Instruct\n  rollout:\n    name: vllm\n    enable_vision: true\n```\n\n### LoRA Training\n\n```yaml\nactor_rollout_ref:\n  actor:\n    lora:\n      enabled: true\n      r: 16\n      alpha: 32\n      target_modules: [\"q_proj\", \"v_proj\"]\n```\n\n---\n\n## Resources\n\n- **Documentation**: https:\u002F\u002Fverl.readthedocs.io\u002F\n- **Paper**: https:\u002F\u002Farxiv.org\u002Fabs\u002F2409.19256\n- **GitHub**: https:\u002F\u002Fgithub.com\u002Fvolcengine\u002Fverl\n- **Recipes**: https:\u002F\u002Fgithub.com\u002Fverl-project\u002Fverl-recipe (DAPO, GSPO, etc.)\n- **Community**: Slack at verl-project\n\n","# verl: Volcano Engine Reinforcement Learning for LLMs\n\nverl is a flexible, efficient, and production-ready RL training library for large language models from ByteDance's Seed team. It implements the HybridFlow framework (EuroSys 2025) and powers models like Doubao-1.5-pro achieving O1-level performance on math benchmarks.\n\n## When to Use verl\n\n**Choose verl when you need:**\n- Production-ready RL training at scale (tested up to 671B parameters)\n- Flexibility to swap backends (FSDP ↔ Megatron-LM ↔ vLLM ↔ SGLang)\n- Support for multiple RL algorithms (PPO, GRPO, RLOO, REINFORCE++, DAPO)\n- Multi-turn rollout with tool calling for agentic workflows\n- Vision-language model RL training\n\n**Consider alternatives when:**\n- You need Megatron-native training → use **slime** or **miles**\n- You want PyTorch-native abstractions with Monarch → use **torchforge**\n- You only need simple SFT\u002FDPO → use **TRL** or **Axolotl**\n\n## Key Features\n\n- **Training backends**: FSDP, FSDP2, Megatron-LM\n- **Rollout engines**: vLLM, SGLang, HuggingFace Transformers\n- **Algorithms**: PPO, GRPO, DAPO, RLOO, ReMax, REINFORCE++, SPIN, SPPO\n- **Models**: Qwen-3, Llama-3.1, DeepSeek, Gemma-2 (0.5B to 671B)\n- **Advanced**: LoRA RL, sequence parallelism, expert parallelism, multi-turn tools\n\n## Installation\n\n```bash\n# Option 1: pip install\npip install verl[vllm]  # or verl[sglang] for SGLang backend\n\n# Option 2: Docker (recommended for production)\ndocker pull verlai\u002Fverl:vllm011.latest\n\n# Option 3: From source\ngit clone https:\u002F\u002Fgithub.com\u002Fvolcengine\u002Fverl.git\ncd verl && pip install -e .[vllm,math]\n```\n\n## Quick Start: GRPO Training\n\n```bash\npython3 -m verl.trainer.main_ppo \\\n    algorithm.adv_estimator=grpo \\\n    data.train_files=~\u002Fdata\u002Fgsm8k\u002Ftrain.parquet \\\n    actor_rollout_ref.model.path=Qwen\u002FQwen2.5-7B \\\n    actor_rollout_ref.rollout.n=8 \\\n    actor_rollout_ref.actor.use_kl_loss=True \\\n    trainer.n_gpus_per_node=8\n```\n\n## Core Architecture\n\nverl uses a **HybridFlow** programming model separating control flow from computation:\n\n```\n┌─────────────────────────────────────────────────────────┐\n│ Single-Process Controller (Ray)                         │\n│ - Orchestrates: rollout → reward → train → sync        │\n└─────────────────────┬───────────────────────────────────┘\n                      │\n┌─────────────────────▼───────────────────────────────────┐\n│ Multi-Process Workers                                   │\n│ ├── ActorRolloutRefWorker (policy + generation)        │\n│ ├── CriticWorker (value estimation, PPO only)          │\n│ └── RewardManager (model-based or rule-based rewards)  │\n└─────────────────────────────────────────────────────────┘\n```\n\n---\n\n## Workflow 1: Math Reasoning with GRPO\n\nUse this workflow for training reasoning models on math tasks like GSM8K or MATH.\n\n### Prerequisites Checklist\n- [ ] GPU cluster with 8+ GPUs (H100 recommended)\n- [ ] Dataset in parquet format with `prompt` and `reward_model` columns\n- [ ] Base model from HuggingFace Hub\n\n### Step 1: Prepare Dataset\n\n```python\nimport pandas as pd\n\ndata = [\n    {\n        \"prompt\": [{\"role\": \"user\", \"content\": \"What is 15 + 27?\"}],\n        \"reward_model\": {\"ground_truth\": \"42\"}\n    },\n    # ... more examples\n]\ndf = pd.DataFrame(data)\ndf.to_parquet(\"train.parquet\")\n```\n\n### Step 2: Define Reward Function\n\n```python\n# reward_function.py\nimport re\n\ndef compute_reward(responses, ground_truths):\n    rewards = []\n    for response, gt in zip(responses, ground_truths):\n        # Extract answer from response\n        match = re.search(r'\\\\boxed{([^}]+)}', response)\n        if match and match.group(1).strip() == gt.strip():\n            rewards.append(1.0)\n        else:\n            rewards.append(0.0)\n    return rewards\n```\n\n### Step 3: Create Training Config\n\n```yaml\n# config\u002Fgrpo_math.yaml\nalgorithm:\n  adv_estimator: grpo\n  gamma: 1.0\n  lam: 1.0\n\ndata:\n  train_files: \u002Fpath\u002Fto\u002Ftrain.parquet\n  val_files: \u002Fpath\u002Fto\u002Fval.parquet\n  train_batch_size: 256\n  max_prompt_length: 512\n  max_response_length: 2048\n\nactor_rollout_ref:\n  model:\n    path: Qwen\u002FQwen2.5-7B-Instruct\n  actor:\n    use_kl_loss: true\n    kl_loss_coef: 0.001\n    ppo_mini_batch_size: 64\n  rollout:\n    name: vllm\n    n: 8  # samples per prompt\n    temperature: 0.7\n    top_p: 0.95\n\ntrainer:\n  total_epochs: 3\n  n_gpus_per_node: 8\n  save_freq: 100\n```\n\n### Step 4: Launch Training\n\n```bash\npython3 -m verl.trainer.main_ppo \\\n    --config-path config \\\n    --config-name grpo_math \\\n    trainer.experiment_name=grpo_math_qwen7b\n```\n\n### Step 5: Monitor and Validate\n- [ ] Check WandB\u002FTensorBoard for loss curves\n- [ ] Verify reward is increasing over steps\n- [ ] Run evaluation on held-out test set\n\n---\n\n## Workflow 2: PPO with Critic Model\n\nUse this workflow when you need value-based advantage estimation (GAE).\n\n### Key Differences from GRPO\n- Requires separate critic model\n- Uses Generalized Advantage Estimation (GAE)\n- Better for tasks with dense rewards\n\n### Configuration\n\n```yaml\nalgorithm:\n  adv_estimator: gae  # Use GAE instead of GRPO\n  gamma: 0.99\n  lam: 0.95\n\ncritic:\n  model:\n    path: Qwen\u002FQwen2.5-7B-Instruct  # Can be same or different from actor\n  ppo_mini_batch_size: 64\n\nactor_rollout_ref:\n  actor:\n    use_kl_loss: true\n    kl_loss_coef: 0.02\n    clip_ratio: 0.2  # PPO clipping\n```\n\n### Launch with Critic\n\n```bash\npython3 -m verl.trainer.main_ppo \\\n    algorithm.adv_estimator=gae \\\n    critic.model.path=Qwen\u002FQwen2.5-7B-Instruct \\\n    trainer.n_gpus_per_node=8\n```\n\n---\n\n## Workflow 3: Large-Scale Training with Megatron\n\nUse this workflow for models >70B parameters or when you need expert parallelism.\n\n### Prerequisites\n- [ ] Install Megatron-LM bridge: `pip install mbridge`\n- [ ] Convert model to Megatron format\n- [ ] Multi-node cluster with NVLink\u002FInfiniBand\n\n### Configuration for 70B+ Models\n\n```yaml\nactor_rollout_ref:\n  model:\n    path: \u002Fpath\u002Fto\u002Fmegatron\u002Fcheckpoint\n    backend: megatron\n  actor:\n    strategy: megatron\n    tensor_model_parallel_size: 8\n    pipeline_model_parallel_size: 2\n  rollout:\n    name: vllm\n    tensor_parallel_size: 8\n```\n\n### Launch Multi-Node\n\n```bash\n# On head node\nray start --head --port=6379\n\n# On worker nodes\nray start --address='head_ip:6379'\n\n# Launch training\npython3 -m verl.trainer.main_ppo \\\n    trainer.nnodes=4 \\\n    trainer.n_gpus_per_node=8\n```\n\n---\n\n## Configuration Reference\n\n### Algorithm Selection\n\n| Algorithm | `adv_estimator` | Use Case |\n|-----------|-----------------|----------|\n| GRPO | `grpo` | Critic-free, math\u002Freasoning |\n| PPO\u002FGAE | `gae` | Dense rewards, value estimation |\n| REINFORCE++ | `reinforce_plus_plus` | Variance reduction |\n| RLOO | `rloo` | Leave-one-out baseline |\n| ReMax | `remax` | Maximum reward baseline |\n| OPO | `opo` | Optimal policy optimization |\n\n### Key Parameters\n\n```yaml\n# Rollout parameters\nactor_rollout_ref.rollout.n: 8              # Samples per prompt\nactor_rollout_ref.rollout.temperature: 0.7  # Sampling temperature\nactor_rollout_ref.rollout.top_p: 0.95       # Nucleus sampling\n\n# Training parameters\nactor_rollout_ref.actor.lr: 1e-6            # Learning rate\nactor_rollout_ref.actor.ppo_mini_batch_size: 64\nactor_rollout_ref.actor.clip_ratio: 0.2     # PPO clip range\n\n# KL control\nactor_rollout_ref.actor.use_kl_loss: true\nactor_rollout_ref.actor.kl_loss_coef: 0.001\nalgorithm.kl_ctrl.target_kl: 0.1            # For adaptive KL control\n```\n\n---\n\n## Common Issues and Solutions\n\n### Issue: OOM During Rollout\n\n**Symptoms**: CUDA out of memory during generation phase\n\n**Solutions**:\n```yaml\n# Reduce batch size\nactor_rollout_ref.rollout.log_prob_micro_batch_size: 4\n\n# Enable gradient checkpointing\nactor_rollout_ref.model.enable_gradient_checkpointing: true\n\n# Use FSDP2 with CPU offloading\nactor_rollout_ref.actor.strategy: fsdp2\nactor_rollout_ref.actor.fsdp_config.offload_policy: true\n```\n\n### Issue: Training Instability\n\n**Symptoms**: Loss spikes, reward collapse\n\n**Solutions**:\n```yaml\n# Reduce learning rate\nactor_rollout_ref.actor.lr: 5e-7\n\n# Increase KL penalty\nactor_rollout_ref.actor.kl_loss_coef: 0.01\n\n# Enable gradient clipping\nactor_rollout_ref.actor.max_grad_norm: 1.0\n```\n\n### Issue: Slow Weight Sync\n\n**Symptoms**: Long pauses between rollout and training\n\n**Solutions**:\n```bash\n# Use FSDP2 for faster resharding\nactor_rollout_ref.actor.strategy=fsdp2\n\n# Enable async weight transfer\ntrainer.async_weight_update=true\n```\n\n### Issue: vLLM Version Mismatch\n\n**Symptoms**: Import errors or generation failures\n\n**Solution**: Use compatible versions:\n```bash\npip install vllm>=0.8.5,\u003C=0.12.0\n# Avoid vLLM 0.7.x (known bugs)\n```\n\n---\n\n## Advanced Topics\n\n### Multi-Turn Tool Calling\n\nSee [references\u002Fmulti-turn.md](references\u002Fmulti-turn.md) for agentic workflows with tool use.\n\n### Vision-Language Models\n\n```yaml\nactor_rollout_ref:\n  model:\n    path: Qwen\u002FQwen2.5-VL-7B-Instruct\n  rollout:\n    name: vllm\n    enable_vision: true\n```\n\n### LoRA Training\n\n```yaml\nactor_rollout_ref:\n  actor:\n    lora:\n      enabled: true\n      r: 16\n      alpha: 32\n      target_modules: [\"q_proj\", \"v_proj\"]\n```\n\n---\n\n## Resources\n\n- **Documentation**: https:\u002F\u002Fverl.readthedocs.io\u002F\n- **Paper**: https:\u002F\u002Farxiv.org\u002Fabs\u002F2409.19256\n- **GitHub**: https:\u002F\u002Fgithub.com\u002Fvolcengine\u002Fverl\n- **Recipes**: https:\u002F\u002Fgithub.com\u002Fverl-project\u002Fverl-recipe (DAPO, GSPO, etc.)\n- **Community**: Slack at verl-project\n\n",{"name":5,"license":15,"metadata":50,"description":7},{"tags":51,"author":53,"version":54,"dependencies":55},[52,37,39,36,38,40],"Reinforcement Learning","Orchestra Research","1.0.0",[56,57,58,59,60],"verl>=0.3.0","torch>=2.0.0","ray>=2.41.0","vllm>=0.8.2","transformers>=4.40.0","main","2b749c82b5d11fac39d9a6e021ce3f5b8a541bd1","cli-tool\u002Fcomponents\u002Fskills\u002Fai-research\u002Fpost-training-verl",3931,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates","https:\u002F\u002Faitmpl.com","https:\u002F\u002Fgithub.com\u002Fdavila7","https:\u002F\u002Fdanielavila.me","Building AI dev tools with LLMs","2026-11-14 06:14:46",{"source":8,"skillName":5,"commands":72,"prompt":82,"zipAvailable":83},[73,76,79],{"tool":74,"command":75},"npx","npx skills add davila7\u002Fclaude-code-templates --skill post-training-verl",{"tool":77,"command":78},"bunx","bunx skills add davila7\u002Fclaude-code-templates --skill post-training-verl",{"tool":80,"command":81},"pnpm","pnpm dlx skills add davila7\u002Fclaude-code-templates --skill post-training-verl","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fai-research\u002Fpost-training-verl\u002FSKILL.md",true,[85,91,97,101,105],{"path":86,"resourceType":87,"resourceTypeDesc":88,"sizeBytes":89,"sourceHtmlUrl":90},"assets\u002Fexample_asset.txt","ASSET","资产文件",865,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fenterprise-communication\u002Fcapa-officer\u002Fassets\u002Fexample_asset.txt",{"path":92,"resourceType":93,"resourceTypeDesc":94,"sizeBytes":95,"sourceHtmlUrl":96},"references\u002Fapi_reference.md","REFERENCE","参考文档",963,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fenterprise-communication\u002Fcapa-officer\u002Freferences\u002Fapi_reference.md",{"path":98,"resourceType":93,"resourceTypeDesc":94,"sizeBytes":99,"sourceHtmlUrl":100},"references\u002Fapi-reference.md",6941,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fai-research\u002Fpost-training-verl\u002Freferences\u002Fapi-reference.md",{"path":102,"resourceType":93,"resourceTypeDesc":94,"sizeBytes":103,"sourceHtmlUrl":104},"references\u002Ftroubleshooting.md",6792,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fai-research\u002Fpost-training-verl\u002Freferences\u002Ftroubleshooting.md",{"path":106,"resourceType":107,"resourceTypeDesc":108,"sizeBytes":109,"sourceHtmlUrl":110},"scripts\u002Fexample.py","SCRIPT","脚本",583,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fenterprise-communication\u002Fcapa-officer\u002Fscripts\u002Fexample.py",[112,141,168,201,236,260],{"skillNo":113,"name":114,"nameZh":6,"description":115,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":116,"categories":117,"keywords":123,"publishedAt":45},"01M70NHBTQ8BV0A2SCAZXEPHCX","wealthville-pool-scoring","Score DeFi liquidity pools before providing liquidity. Returns a 0-100 pool score and an Enter\u002FHold\u002FExit\u002FReduce\u002FAvoid verdict with calibrated confidence for ~68,800 Solana pools and 575 EVM pools, plus a public miss-inclusive track record. Use when asked whether a pool is worth LPing into, which pools look best right now, or how accurate the scores have actually been.","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fweb-data\u002Fwealthville-pool-scoring\u002FSKILL.md",[118],{"categoryNo":119,"categoryName":120,"taxonomyType":121,"taxonomyTypeDesc":122},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[124,125,126,127,128,25,129,130,31,23,131,132,133,134,135,136,137,138,139,140],"scoring","pool","wealthville pool scoring","liquidity","which","exit","now","reduce","with","record","track","verdict","evm","solana","how","public","look",{"skillNo":142,"name":143,"nameZh":6,"description":144,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":145,"categories":146,"keywords":148,"publishedAt":45},"01M70MB6VR8ERJ4EM8D5KHBMEV","technical-debt-auditor","Scans a codebase and produces a quantified technical debt report with a prioritized remediation plan. Analyzes dead code, TODO\u002FFIXME\u002FHACK markers, dependency staleness, complexity hotspots, code duplication signals, test coverage gaps, unused exports, and outdated patterns. Use this skill when the user asks to audit their codebase for tech debt, wants a health check of their code quality, says things like \"where should I focus refactoring\", \"find code smells\", \"audit my codebase\", \"what technical debt do we have\", \"prioritize cleanup work\", or asks about code quality metrics. Also trigger when the user mentions \"code rot\", \"legacy cleanup\", \"refactoring priorities\", or wants to understand the overall health of a project before a major effort. Trigger even if the user doesn't say \"technical debt\" explicitly — any request to systematically assess code quality and identify improvement priorities belongs here.","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fdevelopment\u002Ftechnical-debt-auditor\u002FSKILL.md",[147],{"categoryNo":119,"categoryName":120,"taxonomyType":121,"taxonomyTypeDesc":122},[149,150,151,152,153,154,155,156,157,158,159,23,160,161,162,163,164,165,166,167],"technical debt","debt","technical","technical debt auditor","auditor","coverage","complexity","markers","project","cleanup","unused","code","todo","work","plan","focus","signals","audit","dependency",{"skillNo":169,"name":170,"nameZh":6,"description":171,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":172,"categories":173,"keywords":183,"publishedAt":45},"01M70KPYH9QNHCNME5GJQGE22K","readonly-sql-guard","Make an agent's database access provably read-only. Use when an agent runs caller-supplied SQL against a production\u002FERP\u002Fanalytics database (a \"query\" MCP tool, text-to-SQL, an admin console), or when reviewing DB-access code for write\u002FDDL\u002Fside-effect\u002Ftransaction-escape risks. Checks what a statement CALLS, not just its shape.","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fdatabase\u002Freadonly-sql-guard\u002FSKILL.md",[174,177,178],{"categoryNo":175,"categoryName":176,"taxonomyType":121,"taxonomyTypeDesc":122},"01JBSEEDCAT000000000000010","后端与接口",{"categoryNo":119,"categoryName":120,"taxonomyType":121,"taxonomyTypeDesc":122},{"categoryNo":179,"categoryName":180,"taxonomyType":181,"taxonomyTypeDesc":182},"01JBSEEDCAT000000000000033","数据分析师","PROFESSION","适用职业",[184,185,186,187,188,189,190,25,191,192,193,194,195,196,197,31,198,199,23,200],"readonly","guard","sql guard","sql","readonly sql guard","erp","effect","read-only","db","just","tool","calls","only","MCP","access","database","query",{"skillNo":202,"name":203,"nameZh":6,"description":204,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":205,"licenseFileUrl":206,"sourceHtmlUrl":207,"categories":208,"keywords":212,"publishedAt":45},"01M6AFY6F7CC2BBZP1QD6VTNVC","web-artifacts-builder","Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn\u002Fui). Use for complex artifacts requiring state management, routing, or shadcn\u002Fui components - not for simple single-file HTML\u002FJSX artifacts.","Apache-2.0","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Futilities\u002Fweb-artifacts-builder\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Futilities\u002Fweb-artifacts-builder\u002FSKILL.md",[209],{"categoryNo":210,"categoryName":211,"taxonomyType":121,"taxonomyTypeDesc":122},"01JBSEEDCAT000000000000009","前端开发",[213,214,215,216,217,218,219,220,221,222,223,224,225,226,227,228,229,230,231,232,233,234,25,235],"artifacts builder","web","web artifacts builder","artifacts","builder","state","routing","state management","component","css","frontend","ui","Tools","file","components","shadcn","react","Tailwind","suite","jsx","html","single-file","ui components",{"skillNo":237,"name":238,"nameZh":6,"description":239,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":205,"licenseFileUrl":240,"sourceHtmlUrl":241,"categories":242,"keywords":247,"publishedAt":45},"01M6AFWF33SFV9Z3SF4PEEE4BF","webapp-testing","Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fsecurity\u002Fwebapp-testing\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fsecurity\u002Fwebapp-testing\u002FSKILL.md",[243,246],{"categoryNo":244,"categoryName":245,"taxonomyType":121,"taxonomyTypeDesc":122},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":210,"categoryName":211,"taxonomyType":121,"taxonomyTypeDesc":122},[248,249,250,251,252,214,253,223,224,254,255,256,257,25,258,132,259],"webapp testing","testing","webapp","browser","playwright","applications","debugging","toolkit","logs","screenshots","local","behavior",{"skillNo":261,"name":262,"nameZh":6,"description":263,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":205,"licenseFileUrl":264,"sourceHtmlUrl":265,"categories":266,"keywords":274,"publishedAt":45},"01M6AF67Q67JJJ5V51JR2P52KA","discernment-nudge","After you give a substantive answer or draft that the user may act on — advice or recommendations, drafted artifacts such as goals, plans, pitches, proposals, or emails, estimates or projections, analysis or interpretation of data, factual claims they may rely on, or a multi-step argument — invoke this skill BEFORE finalizing your reply and then, if it applies, append 2-3 short follow-up questions, each tied to something specific in what you just produced, that help the user check key facts, probe the reasoning or assumptions, and notice missing context. Do this at most once per conversation. Skip it when the user asked a trivial how-to or simple lookup, wants a purely educational explanation, asked you only to format, convert, or assemble a file from content they provided, is writing code they will run, is doing creative writing or casual chat, or already asked you to double-check, cite, or review — the skill file explains these boundaries and the exact output format.","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fproductivity\u002Fdiscernment-nudge\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fproductivity\u002Fdiscernment-nudge\u002FSKILL.md",[267,270,271],{"categoryNo":268,"categoryName":269,"taxonomyType":121,"taxonomyTypeDesc":122},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":119,"categoryName":120,"taxonomyType":121,"taxonomyTypeDesc":122},{"categoryNo":272,"categoryName":273,"taxonomyType":181,"taxonomyTypeDesc":182},"01JBSEEDCAT000000000000036","内容创作者",[275,276,277,278,279,280,281,282,216,283,226,160,284,285,286,287,288,196,289,23,290,291,292,293,294,138,295],"discernment nudge","nudge","interpretation","help","run","reasoning","data","writing","lookup","content","review","conversation","short","up","invoke","you","analysis","claims","argument","assemble","recommendations"]