openrlhf-training
High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.
- License 许可
- MIT
- 最近更新
- 2026-11-15
- Stars
- ★ 2.0k
- Forks
- 233
01 前言 Front Matter
- name
- openrlhf-training
- tags
- [ "Post-Training", "OpenRLHF", "RLHF", "PPO", "GRPO", "RLOO", "DPO", "Ray", "vLLM", "Distributed Training", "Large Models", "ZeRO-3" ]
- author
- Orchestra Research
- license
- MIT
- version
- 1.0.0
- description
- High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.
- dependencies
- [ "openrlhf", "ray", "vllm", "torch", "transformers", "deepspeed" ]
02 资源文件 4
- 参考文档references/algorithm-comparison.md9.6 KB
- 参考文档references/custom-rewards.md15.5 KB
- 参考文档references/hybrid-engine.md7.1 KB
- 参考文档references/multi-node-training.md10.8 KB
03 SKILL.md 正文
OpenRLHF - High-Performance RLHF Training
Quick start
OpenRLHF is a Ray-based RLHF framework optimized for distributed training with vLLM inference acceleration.
Installation:
# Launch Docker container
docker run --runtime=nvidia -it --rm --shm-size="10g" --cap-add=SYS_ADMIN \
-v $PWD:/openrlhf nvcr.io/nvidia/pytorch:25.02-py3 bash
# Uninstall conflicts
sudo pip uninstall xgboost transformer_engine flash_attn pynvml -y
# Install OpenRLHF with vLLM
pip install openrlhf[vllm]
PPO Training (Hybrid Engine):
ray start --head --node-ip-address 0.0.0.0 --num-gpus 8
ray job submit --address="http://127.0.0.1:8265" \
--runtime-env-json='{"working_dir": "/openrlhf"}' \
-- python3 -m openrlhf.cli.train_ppo_ray \
--ref_num_nodes 1 --ref_num_gpus_per_node 8 \
--reward_num_nodes 1 --reward_num_gpus_per_node 8 \
--critic_num_nodes 1 --critic_num_gpus_per_node 8 \
--actor_num_nodes 1 --actor_num_gpus_per_node 8 \
--vllm_num_engines 4 --vllm_tensor_parallel_size 2 \
--colocate_all_models \
--vllm_gpu_memory_utilization 0.5 \
--pretrain OpenRLHF/Llama-3-8b-sft-mixture \
--reward_pretrain OpenRLHF/Llama-3-8b-rm-700k \
--save_path ./output/llama3-8b-rlhf \
--micro_train_batch_size 8 --train_batch_size 128 \
--micro_rollout_batch_size 16 --rollout_batch_size 1024 \
--max_epochs 1 --prompt_max_len 1024 --generate_max_len 1024 \
--zero_stage 3 --bf16 \
--actor_learning_rate 5e-7 --critic_learning_rate 9e-6 \
--init_kl_coef 0.01 --normalize_reward \
--gradient_checkpointing --packing_samples \
--vllm_enable_sleep --deepspeed_enable_sleep
GRPO Training (Group Normalized Policy Optimization):
# Same command as PPO, but add:
--advantage_estimator group_norm
Common workflows
Workflow 1: Full RLHF pipeline (SFT → Reward Model → PPO)
Step 1: Train reward model (DPO):
deepspeed --module openrlhf.cli.train_rm \
--save_path ./output/llama3-8b-rm \
--save_steps -1 --logging_steps 1 \
--eval_steps -1 --train_batch_size 256 \
--micro_train_batch_size 1 --pretrain meta-llama/Meta-Llama-3-8B \
--bf16 --max_epochs 1 --max_len 8192 \
--zero_stage 3 --learning_rate 9e-6 \
--dataset OpenRLHF/preference_dataset_mixture2_and_safe_pku \
--apply_chat_template --chosen_key chosen \
--rejected_key rejected --flash_attn --gradient_checkpointing
Step 2: PPO training:
ray start --head --node-ip-address 0.0.0.0 --num-gpus 8
ray job submit --address="http://127.0.0.1:8265" \
-- python3 -m openrlhf.cli.train_ppo_ray \
--ref_num_nodes 1 --ref_num_gpus_per_node 8 \
--reward_num_nodes 1 --reward_num_gpus_per_node 8 \
--critic_num_nodes 1 --critic_num_gpus_per_node 8 \
--actor_num_nodes 1 --actor_num_gpus_per_node 8 \
--vllm_num_engines 4 --vllm_tensor_parallel_size 2 \
--colocate_all_models \
--pretrain OpenRLHF/Llama-3-8b-sft-mixture \
--reward_pretrain ./output/llama3-8b-rm \
--save_path ./output/llama3-8b-ppo \
--micro_train_batch_size 8 --train_batch_size 128 \
--micro_rollout_batch_size 16 --rollout_batch_size 1024 \
--max_epochs 1 --prompt_max_len 1024 --generate_max_len 1024 \
--zero_stage 3 --bf16 \
--actor_learning_rate 5e-7 --critic_learning_rate 9e-6 \
--init_kl_coef 0.01 --normalize_reward \
--vllm_enable_sleep --deepspeed_enable_sleep
Workflow 2: GRPO training (no critic model needed)
Memory-efficient alternative to PPO:
ray job submit --address="http://127.0.0.1:8265" \
-- python3 -m openrlhf.cli.train_ppo_ray \
--advantage_estimator group_norm \
--ref_num_nodes 1 --ref_num_gpus_per_node 8 \
--reward_num_nodes 1 --reward_num_gpus_per_node 8 \
--actor_num_nodes 1 --actor_num_gpus_per_node 8 \
--vllm_num_engines 4 --vllm_tensor_parallel_size 2 \
--colocate_all_models \
--pretrain OpenRLHF/Llama-3-8b-sft-mixture \
--reward_pretrain OpenRLHF/Llama-3-8b-rm-700k \
--save_path ./output/llama3-8b-grpo \
--micro_train_batch_size 8 --train_batch_size 128 \
--micro_rollout_batch_size 16 --rollout_batch_size 1024 \
--max_epochs 1 --bf16 \
--actor_learning_rate 5e-7 \
--init_kl_coef 0.01 --use_kl_loss --kl_estimator k3 \
--normalize_reward --no_advantage_std_norm
Key GRPO parameters:
--advantage_estimator group_norm- Enables GRPO--use_kl_loss- KL loss from GRPO paper--kl_estimator k3- Loss function (k2 ≈ k1)--no_advantage_std_norm- Disables std normalization
Workflow 3: DPO training (preference optimization)
Simpler alternative without reward model:
deepspeed --module openrlhf.cli.train_dpo \
--save_path ./output/llama3-8b-dpo \
--save_steps -1 --logging_steps 1 \
--eval_steps -1 --train_batch_size 256 \
--micro_train_batch_size 2 --pretrain meta-llama/Meta-Llama-3-8B \
--bf16 --max_epochs 1 --max_len 8192 \
--zero_stage 3 --learning_rate 5e-7 --beta 0.1 \
--dataset OpenRLHF/preference_dataset_mixture2_and_safe_pku \
--apply_chat_template --chosen_key chosen \
--rejected_key rejected --flash_attn --gradient_checkpointing
When to use vs alternatives
Use OpenRLHF when:
- Training large models (7B-70B+) with RL
- Need vLLM inference acceleration
- Want distributed architecture with Ray
- Have multi-node GPU cluster
- Need PPO/GRPO/RLOO/DPO in one framework
Algorithm selection:
- PPO: Maximum control, best for complex rewards
- GRPO: Memory-efficient, no critic needed
- RLOO: Modified PPO with per-token KL
- REINFORCE++: More stable than GRPO, faster than PPO
- DPO: Simplest, no reward model needed
Use alternatives instead:
- TRL: Single-node training, simpler API
- veRL: ByteDance's framework for 671B models
- DeepSpeedChat: Integrated with DeepSpeed ecosystem
Common issues
Issue: GPU OOM with large models
Disable model colocation:
# Remove --colocate_all_models flag
# Allocate separate GPUs for each model
--actor_num_gpus_per_node 8 \
--critic_num_gpus_per_node 8 \
--reward_num_gpus_per_node 8 \
--ref_num_gpus_per_node 8
Issue: DeepSpeed GPU index out of range
Set environment variable:
export RAY_EXPERIMENTAL_NOSET_CUDA_VISIBLE_DEVICES=1
Issue: Training instability
Use Hybrid Engine instead of async:
--colocate_all_models \
--vllm_enable_sleep \
--deepspeed_enable_sleep
Adjust KL coefficient:
--init_kl_coef 0.05 # Increase from 0.01
Issue: Slow generation during PPO
Enable vLLM acceleration:
--vllm_num_engines 4 \
--vllm_tensor_parallel_size 2 \
--vllm_gpu_memory_utilization 0.5
Advanced topics
Hybrid Engine GPU sharing: See references/hybrid-engine.md for vLLM sleep mode, DeepSpeed sleep mode, and optimal node allocation.
Algorithm comparison: See references/algorithm-comparison.md for PPO vs GRPO vs RLOO vs REINFORCE++ benchmarks and hyperparameters.
Multi-node setup: See references/multi-node-training.md for Ray cluster configuration and fault tolerance.
Custom reward functions: See references/custom-rewards.md for reinforced fine-tuning and agent RLHF.
Hardware requirements
- GPU: NVIDIA A100/H100 recommended
- VRAM:
- 7B model: 8× A100 40GB (Hybrid Engine)
- 70B model: 48× A100 80GB (vLLM:Actor:Critic = 1:1:1)
- Multi-node: Ray cluster with InfiniBand recommended
- Docker: NVIDIA PyTorch container 25.02+
Performance:
- 2× faster than DeepSpeedChat
- vLLM inference acceleration
- Hybrid Engine minimizes GPU idle time
Resources
- Docs: https://github.com/OpenRLHF/OpenRLHF
- Paper: https://arxiv.org/abs/2405.11143
- Examples: https://github.com/OpenRLHF/OpenRLHF/tree/main/examples
- Discord: Community support
04 同仓库更多 Skills
theme-factory
Toolkit for styling artifacts with a theme. These artifacts can be slides, docs, reportings, HTML landing pages, etc. There are 10 pre-set themes with colors/fonts that you can apply to any artifact that has been creating, or can generate a new theme on-the-fly.
web-artifacts-builder
Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn/ui). Use for complex artifacts requiring state management, routing, or shadcn/ui components - not for simple single-file HTML/JSX artifacts.
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.
neuropixels-analysis
Neuropixels neural recording analysis. Load SpikeGLX/OpenEphys data, preprocess, motion correction, Kilosort4 spike sorting, quality metrics, Allen/IBL curation, AI-assisted visual analysis, for Neuropixels 1.0/2.0 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.
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.
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.