openrlhf-training

davila7's avatarfrom davila7

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.

⭐15.4kstars🔀1.3kforks📁View on GitHub🕐Updated Jan 11, 2026
[Developer Tools]

When & Why to Use This Skill

OpenRLHF is a high-performance RLHF (Reinforcement Learning from Human Feedback) framework built on Ray and vLLM, specifically optimized for the distributed training of large language models ranging from 7B to over 70B parameters. By integrating ZeRO-3 and vLLM inference acceleration, it provides a distributed architecture that is 2x faster than DeepSpeedChat, supporting a wide array of alignment algorithms including PPO, GRPO, RLOO, and DPO.

Use Cases

  • Scaling RLHF training for 70B+ parameter models across multi-node GPU clusters using Ray and DeepSpeed ZeRO-3 for maximum efficiency.
  • Implementing memory-efficient model alignment using GRPO (Group Normalized Policy Optimization) to eliminate the need for a separate critic model.
  • Accelerating PPO (Proximal Policy Optimization) workflows by leveraging vLLM-powered inference for high-speed rollout generation and reduced GPU idle time.
  • Executing full-stack post-training pipelines, from Reward Model training to Direct Preference Optimization (DPO), within a single unified framework.
  • Optimizing large-scale model performance on NVIDIA A100/H100 clusters using advanced features like Hybrid Engine and GPU resource sharing.
nameopenrlhf-training
descriptionHigh-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.
version1.0.0
authorOrchestra Research
licenseMIT
tags[Post-Training, OpenRLHF, RLHF, PPO, GRPO, RLOO, DPO, Ray, vLLM, Distributed Training, Large Models, ZeRO-3]
dependencies[openrlhf, ray, vllm, torch, transformers, deepspeed]

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