[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6FCB989YT362W8HEQ6CWCG0":3,"skill-same:01M6FCB989YT362W8HEQ6CWCG0":93},{"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":44,"linkOnly":45,"markdownContent":46,"markdownBody":47,"frontMatter":48,"refName":60,"commitSha":61,"skillDirectory":62,"forksCount":63,"repositoryHtmlUrl":64,"repositoryHomepageUrl":6,"ownerHtmlUrl":65,"ownerBlogUrl":66,"ownerDescription":67,"updatedAt":68,"install":69,"resources":82},"01M6FCB989YT362W8HEQ6CWCG0","verl-rl-training",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.","Dicklesworthstone\u002Fpi_agent_rust","Dicklesworthstone","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F35050222?v=4","01M6CG26AR942CSVMN1M5E3DSE","Jeff Emanuel","01M6CG26EPP71PSHW1VDS71121",1979,"MIT","https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\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],"training","verl rl training","verl","rl","engine","when","for","implementing","orchestra","at","with","volcano","algorithms","use","llms","learning","Transformers","vLLM","Ray","PPO","RLHF","Post-Training","Distributed Training","infrastructure","distributed","2026-11-10 10:12:30",false,"---\nname: verl-rl-training\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.\nversion: 1.0.0\nauthor: Orchestra Research\nlicense: MIT\ntags: [Reinforcement Learning, RLHF, GRPO, PPO, Post-Training, Distributed Training]\ndependencies: [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,"tags":49,"author":52,"license":15,"version":53,"description":7,"dependencies":54},[50,39,51,38,40,41],"Reinforcement Learning","GRPO","Orchestra Research","1.0.0",[55,56,57,58,59],"verl>=0.3.0","torch>=2.0.0","ray>=2.41.0","vllm>=0.8.2","transformers>=4.40.0","main","90b50ed0778745a80fd5bfb13b1843d45b9e82f7","tests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fai-research\u002Fpost-training-verl",231,"https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust","https:\u002F\u002Fgithub.com\u002FDicklesworthstone","https:\u002F\u002Fwww.jeffreyemanuel.com\u002F","Building in NY","2026-11-11 01:23:40",{"source":8,"skillName":5,"commands":70,"prompt":80,"zipAvailable":81},[71,74,77],{"tool":72,"command":73},"npx","npx skills add Dicklesworthstone\u002Fpi_agent_rust --skill verl-rl-training",{"tool":75,"command":76},"bunx","bunx skills add Dicklesworthstone\u002Fpi_agent_rust --skill verl-rl-training",{"tool":78,"command":79},"pnpm","pnpm dlx skills add Dicklesworthstone\u002Fpi_agent_rust --skill verl-rl-training","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fai-research\u002Fpost-training-verl\u002FSKILL.md",true,[83,89],{"path":84,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":87,"sourceHtmlUrl":88},"references\u002Fapi-reference.md","REFERENCE","参考文档",6941,"https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fai-research\u002Fpost-training-verl\u002Freferences\u002Fapi-reference.md",{"path":90,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":91,"sourceHtmlUrl":92},"references\u002Ftroubleshooting.md",6792,"https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fai-research\u002Fpost-training-verl\u002Freferences\u002Ftroubleshooting.md",[94,133,157,183,210,232],{"skillNo":95,"name":96,"nameZh":6,"description":97,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":98,"licenseFileUrl":99,"sourceHtmlUrl":100,"categories":101,"keywords":107,"publishedAt":44},"01M6FEGPPK0ZTKCCM0R3E3KS33","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\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Futilities\u002Fweb-artifacts-builder\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Futilities\u002Fweb-artifacts-builder\u002FSKILL.md",[102],{"categoryNo":103,"categoryName":104,"taxonomyType":105,"taxonomyTypeDesc":106},"01JBSEEDCAT000000000000009","前端开发","SCENARIO","使用场景",[108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,32,126,127,128,129,130,25,131,132],"web artifacts builder","web","artifacts","artifacts builder","builder","frontend","css","components","react","component","routing","state management","state","shadcn","file","ui","Tools","Tailwind","html","single-file","simple","ui components","multi","jsx","suite",{"skillNo":134,"name":135,"nameZh":6,"description":136,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":98,"licenseFileUrl":137,"sourceHtmlUrl":138,"categories":139,"keywords":144,"publishedAt":44},"01M6FEE9VTK1W4GY21AKREJXRG","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\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fsecurity\u002Fwebapp-testing\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fsecurity\u002Fwebapp-testing\u002FSKILL.md",[140,143],{"categoryNo":141,"categoryName":142,"taxonomyType":105,"taxonomyTypeDesc":106},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":103,"categoryName":104,"taxonomyType":105,"taxonomyTypeDesc":106},[145,146,147,113,148,149,109,150,151,152,123,153,154,155,156,29,25],"testing","webapp testing","webapp","screenshots","playwright","debugging","toolkit","browser","logs","applications","behavior","local",{"skillNo":158,"name":159,"nameZh":6,"description":160,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":161,"sourceHtmlUrl":162,"categories":163,"keywords":164,"publishedAt":44},"01M6FE20F7ZTQBFWKF4MPJ2SBX","neuropixels-analysis","Neuropixels neural recording analysis. Load SpikeGLX\u002FOpenEphys data, preprocess, motion correction, Kilosort4 spike sorting, quality metrics, Allen\u002FIBL curation, AI-assisted visual analysis, for Neuropixels 1.0\u002F2.0 extracellular electrophysiology. Use when working with neural recordings, spike sorting, extracellular electrophysiology, or when the user mentions Neuropixels, SpikeGLX, Open Ephys, Kilosort, quality metrics, or unit curation.","https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fscientific\u002Fneuropixels-analysis\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fscientific\u002Fneuropixels-analysis\u002FSKILL.md",[],[165,166,167,168,169,170,171,172,32,173,174,175,176,177,178,179,29,180,181,182,25,24],"analysis","neuropixels analysis","spike","metrics","data","open","quality","user","recordings","Neural","motion","visual","quality metrics","unit","sorting","curation","recording","load",{"skillNo":184,"name":185,"nameZh":6,"description":186,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":98,"licenseFileUrl":187,"sourceHtmlUrl":188,"categories":189,"keywords":193,"publishedAt":44},"01M6FDP0PDK6F1GGX395C7625Z","skill-creator","Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.","https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fproductivity\u002Fskill-creator\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fproductivity\u002Fskill-creator\u002FSKILL.md",[190],{"categoryNo":191,"categoryName":192,"taxonomyType":105,"taxonomyTypeDesc":106},"01JBSEEDCAT000000000000023","Skill与智能体开发",[194,195,196,197,198,199,200,201,202,203,204,205,206,207,29,25,208,24,209],"skill creator","creator","workflows","create","update","capabilities","knowledge","tool","integrations","specialized","guide","new skill","existing skill","users","Existing","new",{"skillNo":211,"name":212,"nameZh":6,"description":213,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":214,"sourceHtmlUrl":215,"categories":216,"keywords":218,"publishedAt":44},"01M6FDM8Y7VK7NYG9SM7QPA3PW","notebooklm","Use this skill to query your Google NotebookLM notebooks directly from Claude Code for source-grounded, citation-backed answers from Gemini. Browser automation, library management, persistent auth. Drastically reduced hallucinations through document-only responses.","https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fproductivity\u002Fnotebooklm\u002FLICENSE","https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fproductivity\u002Fnotebooklm\u002FSKILL.md",[217],{"categoryNo":191,"categoryName":192,"taxonomyType":105,"taxonomyTypeDesc":106},[212,219,220,221,222,223,224,152,225,226,32,227,228,229,230,25,231],"google","browser automation","gemini","automation","auth","code","library","management","only","document","query","Citation","source",{"skillNo":233,"name":234,"nameZh":6,"description":235,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":98,"licenseFileUrl":236,"sourceHtmlUrl":237,"categories":238,"keywords":245,"publishedAt":44},"01M6FDHPXE10J3XGTCH5HBHS1M","slack-gif-creator","Knowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like \"make me a GIF of X doing Y for Slack.\"","https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fenterprise-communication\u002Fslack-gif-creator\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002FDicklesworthstone\u002Fpi_agent_rust\u002Fblob\u002Fmain\u002Ftests\u002Fext_conformance\u002Fartifacts\u002Ftemplates-davila7\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fenterprise-communication\u002Fslack-gif-creator\u002FSKILL.md",[239,242],{"categoryNo":240,"categoryName":241,"taxonomyType":105,"taxonomyTypeDesc":106},"01JBSEEDCAT000000000000005","邮件与沟通",{"categoryNo":243,"categoryName":244,"taxonomyType":105,"taxonomyTypeDesc":106},"01JBSEEDCAT000000000000016","图像与多媒体",[246,247,248,195,249,124,250,251,200,32,252,207,253,254,255,25,256,24],"slack gif creator","slack","gif","validation","constraints","Animation","me","make","Utilities","concepts","request"]