verl

KO | EN

gitlite — search

#react #typescript #golang #android #machine-learning #macos #windows #awesome #cli #awesome-list #pytorch #security #mcp #open-source

(★ 23,065)

verl/HybridFlow: A Flexible and Efficient RL Post-Training Framework

File Explorer

Download Latest Version (.zip)

- - issue
  - pr
- - issue
  - pr
- - bug-report.yml
  - config.yml
  - feature-request.yml
- - check-pr-title.yml
  - cpu_unit_tests.yml
  - doc.yml
  - docker-build-ascend-a2.yml
  - docker-build-ascend-a3.yml
  - docker-build-ascend-sglang-a2.yml
  - docker-build-ascend-sglang-a3.yml
  - e2e_ascend.yml
  - e2e_ppo_grpo_trainer_trtllm.yml
  - e2e_ppo_trainer.yml
  - e2e_ppo_trainer_fsdp_turbo_vllm.yml
  - e2e_ppo_trainer_megatron_sglang.yml
  - e2e_ppo_trainer_megatron_sglang_2.yml
  - e2e_ppo_trainer_megatron_sglang_2_ascend.yml
  - e2e_ppo_trainer_megatron_sglang_ascend.yml
  - e2e_ppo_trainer_megatron_vllm.yml
  - e2e_ppo_trainer_megatron_vllm_2.yml
  - e2e_ppo_trainer_megatron_vllm_2_ascend.yml
  - e2e_ppo_trainer_megatron_vllm_rocm.yml
  - e2e_ppo_trainer_veomni_vllm.yml
  - e2e_ppo_trainer_veomni_vllm_ascend.yml
  - e2e_sft_llm.yml
  - e2e_sft_llm_ascend.yml
  - e2e_sft_vlm.yml
  - e2e_v1_separate_async.yml
  - e2e_v1_separate_async_ascend.yml
  - e2e_v1_separate_async_opd.yml
  - gpu_unit_tests.yml
  - model.yml
  - model_ascend.yml
  - nightly_ascend.yml
  - nightly_ascend_multinode.yml
  - npu_unit_tests.yml
  - pre-commit.yml
  - precommit-autofix.yml
  - README.md
  - reward_model_sglang.yml
  - reward_model_sglang_ascend.yml
  - reward_model_vllm.yml
  - reward_model_vllm_ascend.yml
  - sanity.yml
  - scorecard.yml
  - secrets_scan.yml
  - sgl.yml
  - sgl_ascend.yml
  - type-coverage-check.yml
  - vllm.yml
  - vllm_ascend.yml
- CODEOWNERS
- dependabot.yml
- PULL_REQUEST_TEMPLATE.md
- settings.json
- - Dockerfile.ascend.sglang_8.3.rc1_a2
  - Dockerfile.ascend.sglang_8.3.rc1_a3
  - Dockerfile.ascend.sglang_8.5.0_a2
  - Dockerfile.ascend.sglang_8.5.0_a3
  - Dockerfile.ascend_8.2.rc1_a2
  - Dockerfile.ascend_8.2.rc1_a3
  - Dockerfile.ascend_8.3.rc1_a2
  - Dockerfile.ascend_8.3.rc1_a3
  - Dockerfile.ascend_8.5.0_a2
  - Dockerfile.ascend_8.5.0_a2_v0.7.1
  - Dockerfile.ascend_8.5.0_a3
  - Dockerfile.ascend_8.5.0_a3_v0.7.1
  - Dockerfile.ascend_8.5.2_a2_qwen3-5
  - Dockerfile.ascend_8.5.2_a3_qwen3-5
  - Dockerfile.ascend_9.0.0_a2_v0.8.0
  - Dockerfile.ascend_9.0.0_a3_v0.8.0
  - Dockerfile.ascend_9.1.0_a2
  - Dockerfile.ascend_9.1.0_a3
  - supported_tags.md
- - Dockerfile.extention.awsefa
  - Dockerfile.ngc.vllm0.8.sagemaker
- - Apptainerfile.rocm
  - Dockerfile.rocm
  - Dockerfile.rocm.deepseek-v4
  - Dockerfile.rocm6
  - Dockerfile.rocm702
  - Dockerfile.rocm_verl-0.3.0.post1
  - Dockerfile.rocm_verl-0.4.1
  - README.md
- - Dockerfile.app.sglang.vllm.mcore0.12
  - Dockerfile.app.sglang.vllm.mcore0.12.deepep
  - Dockerfile.app.sglang.vllm.mcore0.13.preview
  - Dockerfile.app.vllm.mcore0.12
  - Dockerfile.app.vllm.mcore0.12.deepep
  - Dockerfile.app.vllm.mcore0.13.preview
  - Dockerfile.base
  - README.md
- - Dockerfile.app.sglang0.4.10.post2.mcore0.13
  - Dockerfile.app.sglang0.4.9.post6.mcore0.13
  - Dockerfile.app.vllm.mcore0.13
  - Dockerfile.app.vllm.mcore0.15
  - Dockerfile.base.torch2.7.1
  - README.md
- - Dockerfile.app.sglang.mcore0.12
  - Dockerfile.app.sglang.mcore0.13.preview
  - Dockerfile.base
  - README.md
- - Dockerfile.app.sglang.megatron
  - Dockerfile.base
  - README.md
- - Dockerfile.app.sglang
  - Dockerfile.base
  - Dockerfile.vllm011.mcore_gpt-oss
- - Dockerfile.sglang056exp
  - Dockerfile.vllm012exp
- Dockerfile.isaaclab230
- Dockerfile.stable.sglang
- Dockerfile.stable.trtllm
- Dockerfile.stable.vllm
- Dockerfile.uv.cu130
- README.md
- - custom.css
  - logo.png
- - agent_loop.rst
  - async-on-policy-distill.md
  - attention_implementation.rst
  - checkpoint.rst
  - deepseek_v4_integration.rst
  - delta_weight_sync.md
  - determinism.md
  - dpo_extension.rst
  - dynamic_context_parallel.rst
  - dynamic_schedule.md
  - fsdp_extension.rst
  - fsdp_turbo_backend.rst
  - fully_async.md
  - grafana_prometheus.md
  - megatron_extension.rst
  - megatron_lite_backend.rst
  - mtp.md
  - one_step_off.md
  - placement.rst
  - ppo_lora.rst
  - reward_loop.rst
  - rl_insight.md
  - rollout_trace.rst
  - rope.rst
  - skip_manager.rst
  - v1_async_trainer.md
- - baseline.md
  - dapo.md
  - dppo.md
  - dro.md
  - entropy.md
  - gpg.md
  - grpo.md
  - opd.md
  - opo.md
  - otb.md
  - ppo.md
  - rollout_corr.md
  - rollout_corr_math.md
  - spin.md
  - sppo.md
- - amd_quick_start.rst
  - amd_vllm_page.rst
  - index.rst
- - data.rst
  - single_controller.rst
  - trainer.rst
  - utils.rst
- - index.rst
  - README.md
- - v0.7.md
- - editing-agent-instructions.md
- - transfer_queue.md
- - config.rst
  - gsm8k_example.rst
  - megatron_fsdp_example.rst
  - multi_modal_example.rst
  - ppo_code_architecture.rst
  - sandbox_fusion_example.rst
  - skypilot_examples.rst
- - faq.rst
- - multi_chip_support.rst
- - fp8.md
  - nvfp4_qat.md
- - best_practices.rst
  - device_tuning.rst
  - dpsk.md
  - nsight_profiling.md
  - perf_tuning.rst
  - rollout_kv_offload.md
  - torch_profiling.md
  - verl_profiler_system.md
- - prepare_data.rst
  - reward_function.rst
- - multiturn.rst
  - sandbox_fusion.rst
  - search_tool_example.rst
- - agentic_rl.rst
  - install.rst
  - more_resources.rst
  - multinode.rst
  - quickstart.rst
  - ray_debug_tutorial.rst
- - automodel_workers.rst
  - engine_workers.rst
  - model_engine.rst
  - ray_trainer.rst
  - sglang_worker.rst
  - torchtitan_workers.rst
  - trtllm_worker.rst
- conf.py
- extend_guide.rst
- hybrid_flow.rst
- index.rst
- Makefile
- README.md
- README_vllm0.7.md
- README_vllm0.8.md
- requirements-docs.txt
- single_controller.rst
- - README.md
  - run_qwen3_8b_fsdp.sh
- - aime2024_multiturn_w_tool.py
  - dapo_multiturn_w_tool.py
  - full_hh_rlhf.py
  - geo3k.py
  - geo3k_multiturn_w_tool.py
  - gsm8k.py
  - gsm8k_multiturn_sft.py
  - gsm8k_multiturn_w_tool.py
  - gsm8k_tool_agent_loop.py
  - hellaswag.py
  - math_dataset.py
  - multiturn.py
  - openr1mm.py
  - pokemon.py
  - preprocess_search_r1_dataset.py
  - tinyllava_video_r1.py
- - README.md
  - run_qwen3_30b_a3b_megatron.sh
- - README.md
  - run_qwen3_8b_fsdp.sh
- - run_deepseek_llm_7b.sh
- - README.md
  - run_qwen3_8b_fsdp.sh
- - README.md
  - run_qwen3_8b_fsdp.sh
  - run_qwen3_8b_megatron.sh
- - README.md
  - run_deepseek_v3_671b_megatron.sh
  - run_deepseek_v4_flash_megatron.sh
  - run_deepseek_v4_megatron_lite.sh
  - run_deepseek_v4_veomni.sh
  - run_glm4_1v_9b_fsdp.sh
  - run_gpt_oss_20b_fsdp.sh
  - run_hy_v3_megatron.sh
  - run_kimi_k2_6_glm5_1_megatron_lite.sh
  - run_minicpm_o_2_6_fsdp.sh
  - run_mistral_nemo_12b_skyworkrm_fsdp.sh
  - run_moonlight_16b_a3b_megatron.sh
  - run_nemotron_nano_v3_30b_a3b_megatron.sh
  - run_qwen2-7b_math_megatron_fsdp.sh
  - run_qwen2_5_32b_fsdp.sh
  - run_qwen2_5_vl_7b_fsdp.sh
  - run_qwen2_5_vl_7b_megatron.sh
  - run_qwen3_235b_a22b_megatron.sh
  - run_qwen3_30b_a3b_fsdp.sh
  - run_qwen3_30b_a3b_megatron.sh
  - run_qwen3_30b_veomni.sh
  - run_qwen3_4b_fsdp.sh
  - run_qwen3_5-122b-a10b_veomni.sh
  - run_qwen3_5_122b_a10b_megatron.sh
  - run_qwen3_5_27b_fsdp.sh
  - run_qwen3_5_27b_fsdp_turbo.sh
  - run_qwen3_5_2b_openr1_fsdp.sh
  - run_qwen3_5_2b_video_fsdp.sh
  - run_qwen3_5_35b_a3b_veomni.sh
  - run_qwen3_5_35b_fsdp.sh
  - run_qwen3_5_35b_fsdp_turbo.sh
  - run_qwen3_5_35b_megatron.sh
  - run_qwen3_5_35b_megatron_lite.sh
  - run_qwen3_5_397b_megatron.sh
  - run_qwen3_8b_fsdp.sh
  - run_qwen3_8b_megatron.sh
  - run_qwen3_next_80b_a3b_fsdp.sh
  - run_qwen3_vl_235b_a22b_megatron.sh
  - run_qwen3_vl_30b_a3b_fsdp.sh
  - run_qwen3_vl_30b_a3b_megatron.sh
  - run_qwen3_vl_30b_moe_veomni.sh
  - run_qwen3_vl_8b_fsdp.sh
  - run_qwen3_vl_8b_megatron.sh
  - run_seed_oss_36b_fsdp.sh
- - README.md
  - run_qwen3_30b_a3b_megatron.sh
  - run_qwen3_8b_fsdp.sh
- - README.md
  - run_mimo_7b_mtp_fully_async_megatron_multinode.sh
  - run_mimo_7b_mtp_megatron.sh
  - run_mimo_7b_mtp_rl_vllm_sgl_megatron.sh
- - README.md
- - README.md
  - run_qwen3_0.6b_opd_veomni.sh
  - run_qwen3_5_4b_fsdp.sh
  - run_qwen3_8b_fsdp.sh
  - run_qwen3_8b_megatron.sh
  - run_qwen3_8b_mopd_fsdp.sh
  - run_qwen3_8b_mopd_veomni.sh
  - run_qwen3_vl_8b_fsdp.sh
- - README.md
  - run_qwen3_8b_fsdp.sh
- - README.md
  - run_qwen3_8b_fsdp.sh
  - run_qwen3_8b_megatron.sh
- - README.md
  - run_qwen3_8b_fsdp.sh
- - README.md
  - run_qwen2_5_7b_torch_profile.sh
  - run_qwen2_5_vl_7b_torch_memory.sh
  - run_qwen3_8b_npu_profile_discrete.sh
  - run_qwen3_8b_npu_profile_e2e.sh
- - README.md
  - run_qwen3_8b_fsdp.sh
- - README.md
  - run_qwen2.5_math_7b_sync_fsdp.sh
  - run_qwen3_8b_fsdp.sh
- - README.md
  - run_qwen3_8b_fsdp.sh
- - run_qwen2_5_7b_fsdp.sh
  - run_qwen2_5_7b_fsdp_multi_rs.sh
- - README.md
  - run_qwen3_30b_a3b_megatron.sh
  - run_qwen3_5_35b_a3b_veomni.sh
- - README.md
  - run_qwen3_30b_a3b_fsdp.sh
  - run_qwen3_8b_fsdp.sh
- - README.md
- - README.md
- README.md
- - moe_merge.py
  - moe_split.py
- __init__.py
- chat_template_checker.py
- chat_template_mock_trajectories.py
- converter_hf_to_mcore.py
- diagnose.py
- generate_trainer_config.sh
- init_random_model.py
- install_sglang_mcore_npu.sh
- install_vllm_mcore_npu.sh
- install_vllm_sglang_mcore.sh
- legacy_model_merger.py
- megatron_merge_lora.py
- migrate_megatron_checkpoint_layout.py
- print_cfg.py
- rollout_viewer.py
- - __init__.py
  - test_block_placement.py
  - test_correctness_on_gpu.py
  - test_correctness_on_npu.py
  - test_fsdp_lora_only_checkpoint_on_cpu.py
  - test_global_steps_on_cpu.py
  - test_sglang_loader.py
  - test_sharded_delta.py
  - test_special_server_adapter.py
  - test_utils.py
- - __init__.py
- - test_output_validation_on_cpu.py
- - test_engine.py
  - test_fsdp_no_padding_on_gpu.py
  - test_fused_kernels_ulysses_sp_on_cpu.py
  - test_liger_vl_compat.py
  - test_model_forward_fused.py
  - test_moe_zloss_per_token_loss.py
  - test_tiled_mlp_accuracy.py
  - test_transformer.py
  - test_transformers_ulysses.py
- - __init__.py
  - test_platform_abstraction.py
- - __init__.py
  - test_auto_padding_on_cpu.py
  - test_colocated_workers.py
  - test_colocated_workers_fused.py
  - test_data_transfer.py
  - test_decorator_on_cpu.py
  - test_device_mesh_register.py
  - test_driverfunc_to_worker.py
  - test_fused_workers_on_cpu.py
  - test_get_set_dispatch_collect_cpu.py
  - test_high_level_scheduling_api.py
  - test_nested_worker.py
  - test_ray_collectives.py
  - test_ray_local_envs_on_cpu.py
  - test_ray_utils_on_cpu.py
  - test_rvdz.py
  - test_split_resource_pool.py
  - test_worker_group_basics.py
  - test_worker_group_torch.py
- - README.md
  - run_all.sh
  - test_fsdp2_cpu_offload_state_dict.py
  - test_fsdp2_pinned_model_transfer.py
  - test_fsdp_ckpt.py
  - test_fused_kernels_ulysses_sp.py
  - test_mcore_config_converter.py
  - test_mcore_probe_differential.py
  - test_megatron_dynamic_cp_features.py
  - test_qwen35_linear_attention_ulysses_sp.py
  - test_sglang_delta_loop.py
  - test_sharded_delta_gather.py
  - test_tensor_dict.py
  - test_torch_functional.py
- - __init__.py
  - check_custom_rwd_fn.py
  - check_results.py
  - profiler_finish_hook_marker.sh
  - README.md
  - run_ppo_trainer_fsdp_turbo.sh
  - run_ppo_trainer_megatron.sh
  - run_ppo_trainer_torchtitan.sh
  - run_ppo_trainer_veomni.sh
  - run_v1_colocate_async_disrm.sh
  - run_v1_separate_async.sh
  - run_v1_separate_async_ascend.sh
  - run_v1_separate_async_opd.sh
- - run_qwen3_06b_grpo_mindspeed.sh
  - run_qwen3_06b_ppo.sh
  - run_qwen3_30b_grpo_mindspeed.sh
  - run_qwen3_8b_grpo_profiling.sh
  - run_qwen3_vl_8b_Instruct_fsdp2_npu.sh
- - check_api_docs.py
  - check_dataproto_usage.py
  - check_device_api_usage.py
  - check_docs_time_info.py
  - check_docstrings.py
  - check_example_naming.py
  - check_license.py
  - check_pr_description.py
  - check_pr_title.py
  - check_uv_gpu_only.py
  - test_check_example_naming.py
  - test_config_docs.py
  - test_import.py
  - type_coverage_check.py
  - validate_imported_docs.py
  - validate_structure.py
- - README.md
  - test_memory_buffers.py
- - __init__.py
  - _stub_search_tools.py
  - test_function_tool_on_cpu.py
  - test_mixed_tools_on_cpu.py
- - __init__.py
  - test_constants_ppo_on_cpu.py
  - test_multi_trajectories_advantage_on_cpu.py
- - _test_module.py
  - test_activation_offload.py
  - test_audio_input_support_on_cpu.py
  - test_bucketed_weight_transfer.py
  - test_check_ipc_version_support_on_npu.py
  - test_check_profiler_output.py
  - test_config_on_cpu.py
  - test_continuous_token_on_cpu.py
  - test_dynamic_cp_scheduler.py
  - test_flops_counter.py
  - test_fs_on_cpu.py
  - test_fsdp2_model_transfer_on_cpu.py
  - test_fsdp2_peft_wrapping.py
  - test_fsdp_lora_merge.py
  - test_fsdp_wrap_policy_on_cpu.py
  - test_groupwise.py
  - test_import_utils_on_cpu.py
  - test_linear_cross_entropy.py
  - test_megatron_bshd_preprocess.py
  - test_megatron_mtp_dcp.py
  - test_megatron_router_replay_dcp.py
  - test_mlflow_key_sanitization.py
  - test_model_on_cpu.py
  - test_normalize_peft_param_name.py
  - test_normalize_peft_param_name_on_cpu.py
  - test_nvtx_profile.py
  - test_padding_on_cpu.py
  - test_precision_debugger_profile.py
  - test_prepare_micro_batches_with_group_size.py
  - test_rl_insight_logger_on_cpu.py
  - test_rollout_trace_on_cpu.py
  - test_seqlen_balancing.py
  - test_server_profiler.py
  - test_sglang_fp8_utils.py
  - test_shared_memory.py
  - test_special_adapter_path_integration.py
  - test_special_linear_cross_entropy_tp.py
  - test_special_megatron_kl_loss_tp.py
  - test_special_mstx_profile.py
  - test_temp_env_on_cpu.py
  - test_timeout_decorator_cpu.py
  - test_tokenizer_normalize_on_cpu.py
  - test_torch_functional.py
  - test_torch_profile.py
  - test_tracking_on_cpu.py
  - test_transferqueue_utils_on_cpu.py
  - test_turn_separator_on_cpu.py
  - test_utils_skip.py
  - test_vllm_quant_utils_moe_on_cpu.py
  - test_vllm_weight_name_normalization_on_cpu.py
- - test_chunked_topk_log_probs_on_cpu.py
  - test_distillation_topk_symmetry_on_cpu.py
  - test_engine_forward_step_detach_on_cpu.py
  - test_engine_return_model_output_on_cpu.py
  - test_engine_workers_lora_sync.py
  - test_engine_workers_mini_batch_trace_on_cpu.py
  - test_fsdp_gradient_accumulation_sync_on_cpu.py
  - test_megatron_ddp_config_on_cpu.py
  - test_megatron_distillation_only_on_cpu.py
  - test_megatron_value_head_cp_layout_on_cpu.py
  - test_packed_pad_to_length_on_cpu.py
  - test_router_replay_engine_helpers_on_cpu.py
- __init__.py
- kill_github_tests.sh
- README.md
- test_base_config_on_cpu.py
- test_protocol_on_cpu.py
- test_protocol_v2_on_cpu.py
- - __init__.py
  - base.py
  - delta_checkpoint_engine.py
  - hccl_checkpoint_engine.py
  - kimi_checkpoint_engine.py
  - mooncake_checkpoint_engine.py
  - nccl_checkpoint_engine.py
  - nixl_checkpoint_engine.py
  - README.md
- - __init__.py
- - __init__.py
  - __main__.py
  - base_model_merger.py
  - fsdp_model_merger.py
  - megatron_model_merger.py
  - output_validation.py
- - __init__.py
  - README.md
  - registry.py
  - weight_loader_registry.py
- - __init__.py
- - __init__.py
- - __init__.py
- - __init__.py
  - base_tool.py
  - function_tool.py
  - schemas.py
  - tool_registry.py
- - __init__.py
  - constants_ppo.py
  - main_eval.py
  - main_generation_server.py
  - main_ppo.py
  - main_ppo_v0.py
  - README.md
  - runtime_env.yaml
  - sft_trainer.py
  - sft_trainer_ray.py
- - __init__.py
  - activation_offload.py
  - attention_utils.py
  - config.py
  - device.py
  - distributed.py
  - dynamic_cp_scheduler.py
  - flops_counter.py
  - fp8_utils.py
  - fs.py
  - fsdp_utils.py
  - groupwise.py
  - hdfs_io.py
  - import_utils.py
  - logging_utils.py
  - megatron_peft_utils.py
  - megatron_utils.py
  - memory_utils.py
  - model.py
  - net_utils.py
  - npu_flash_attn_utils.py
  - py_functional.py
  - ray_utils.py
  - rollout_trace.py
  - seqlen_balancing.py
  - tensordict_utils.py
  - torch_dtypes.py
  - torch_functional.py
  - tracking.py
  - transferqueue_utils.py
  - transformers_compat.py
  - ulysses.py
- - version
- - __init__.py
  - engine_workers.py
  - engine_workers_tinker.py
- __init__.py
- base_config.py
- protocol.py
- py.typed
.dockerignore
.git-blame-ignore-revs
.gitignore
.gitmodules
.pre-commit-config.yaml
.readthedocs.yaml
AGENTS.md
CLAUDE.md
CONTRIBUTING.md
LICENSE
manage_envs.py
Notice.txt
pyproject.toml
README.md
recipe
requirements-npu.txt
requirements-test.txt
requirements.txt
setup.py
uv.lock

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

<div align="center">
 👋 Hi, everyone!
    verl is a RL training library initiated by <b>ByteDance Seed team</b> and maintained by the verl community.
    <br>
    <br>
</div>

<a href="https://deepwiki.com/verl-project/verl"><img src="https://devin.ai/assets/deepwiki-badge.png" alt="Ask DeepWiki.com" style="height:20px;"></a>
[![GitHub Repo stars](https://img.shields.io/github/stars/verl-project/verl)](https://github.com/verl-project/verl/stargazers)
[![Twitter](https://img.shields.io/twitter/follow/verl_project)](https://twitter.com/verl_project)
<a href="https://join.slack.com/t/verl-project/shared_invite/zt-3c6mc2khw-v0lo6NfDPuFP6OnkrZwfqw"><img src="https://img.shields.io/badge/Slack-verl-blueviolet?logo=slack&amp"></a>
<a href="https://arxiv.org/pdf/2409.19256"><img src="https://img.shields.io/static/v1?label=EuroSys&message=Paper&color=red"></a>
[![Documentation](https://img.shields.io/badge/documentation-blue)](https://verl.readthedocs.io/en/latest/)
<a href="https://raw.githubusercontent.com/eric-haibin-lin/verl-community/refs/heads/main/WeChat.JPG"><img src="https://img.shields.io/badge/微信-green?logo=wechat&amp"></a>

</div>

![seed logo](https://github.com/user-attachments/assets/c42e675e-497c-4508-8bb9-093ad4d1f216)

<h1 style="text-align: center;">verl: Volcano Engine Reinforcement Learning for LLMs</h1>

verl is a flexible, efficient and production-ready RL training library for large language models (LLMs).

verl is the open-source version of **[HybridFlow: A Flexible and Efficient RLHF Framework](https://arxiv.org/abs/2409.19256v2)** paper.

verl is flexible and easy to use with:

- **Easy extension of diverse RL algorithms**: The hybrid-controller programming model enables flexible representation and efficient execution of complex post-training dataflows. Build RL dataflows such as GRPO, PPO in a few lines of code.

- **Seamless integration of existing LLM infra with modular APIs**: Decouples computation and data dependencies, enabling seamless integration with existing LLM frameworks, such as FSDP, Megatron-LM, vLLM, SGLang, etc

- **Flexible device mapping**: Supports various placement of models onto different sets of GPUs for efficient resource utilization and scalability across different cluster sizes.

- Ready integration with popular HuggingFace models

verl is fast with:

- **State-of-the-art throughput**: SOTA LLM training and inference engine integrations and SOTA RL throughput.

- **Efficient actor model resharding with 3D-HybridEngine**: Eliminates memory redundancy and significantly reduces communication overhead during transitions between training and generation phases.

</p>

## News
- [2026/08] [verl-vla](https://github.com/verl-project/verl-vla) `v0.1.0` is released: a unified VLA post-training framework for human-in-the-loop data collection, supervised fine-tuning, and reinforcement learning across simulators, real robots, and distributed cloud-edge resources, built on top of verl.
- [2026/08] [VeRL-Tinker](https://github.com/verl-project/verl-recipe/tree/main/verl_tinker) is released: keep the Tinker Cookbook loop you know, and run SFT, RL, and distillation on verl-managed GPU workers you control; read the blog [here](https://verl-project.github.io/posts/2026-08-11-verl-tinker/).
- [2026/08] [VeRL-Omni](https://github.com/verl-project/verl-omni) `v0.2.0` is released: faster diffusion RL, rebuilt Qwen3-Omni multimodal training (DPO & GSPO), plus LTX-2.3, Qwen-Image-Edit support and more.
- [2026/07] [RL-Insight](https://github.com/verl-project/rl-insight) is released: online observability for reinforcement learning training. RL-Insight connects training-side metrics, RL state traces, and service dashboards across distributed rollout and optimization workloads.
- [2026/06] [verl-SpeCo](https://github.com/verl-project/verl-SpeCo) is pre-released: a co-training framework for speculative decoding across RL training and inference, keeping draft models aligned during training and reusable for accelerated serving, built on top of verl.
- [2026/05] [uni-agent](https://github.com/verl-project/uni-agent) is released: a unified agent framework to build, run, and train LLM agents at scale, built on top of verl.
- [2026/05] [VeRL-Omni](https://github.com/verl-project/verl-omni) is pre-released: a unified RL stack for diffusion and omni-modal model post-training built on top of verl. Read the [blog post](https://vllm.ai/blog/2026-05-14-verl-omni) for details.
- [2026/05] verl's zero-mismatch HuggingFace rollout [vexact](https://github.com/verl-project/vexact) is released: with batch-invariant kernels, shared model definition with FSDP, and out-of-box examples compatible with VeOmni.
- [2026/04] verl's Megatron backend LoRA and router replay support is showcased at [PyTorch Conference Europe 2026](https://pytorchconferenceeu2026.sched.com/event/2Juce/optimizing-reinforcement-learning-at-trillion-parameter-scale-songlin-jiang-aalto-university-mind-lab).
- [2026/03] verl is presented at NVIDIA GTC26: [session#1](https://www.nvidia.com/en-us/on-demand/session/gtc26-S81829/), [session#2](https://www.nvidia.com/en-us/on-demand/session/gtc26-S81620/)
- [2026/01] verl has been migrated to the [verl-project](https://github.com/verl-project)
- [2026/01] verl first meetup was successfully held in Shanghai on 01/10, hosted by Volcengine and NVIDIA, the slides has been uploaded to [verl-data](https://github.com/verl-project/verl-data).
- [2026/01] The `recipe` directory has been migrated to a dedicated repository: [verl-recipe](https://github.com/verl-project/verl-recipe) and added as a submodule. See https://github.com/verl-project/verl/pull/4795. It can be used as it was after `git submodule update --init --recursive recipe`. Note that [`transfer_queue`](verl/experimental/transfer_queue), [`fully_async_policy`](verl/experimental/fully_async_policy), [`one_step_off_policy`](verl/experimental/one_step_off_policy) and [`vla`](verl/experimental/vla) are kept under [`verl/experimental`](verl/experimental) since they are planned to be merged into the main library. Use them through `verl.experimental.{module}`.
- [2025/12] [Mind Lab](https://macaron.im/mindlab) successfully used [verl](https://github.com/verl-project/verl) and [Megatron-bridge](https://github.com/NVIDIA-NeMo/Megatron-Bridge) to train GRPO Lora for Trillion-parameter model on 64 H800 - See their [techblog](https://macaron.im/mindlab/research/building-trillion-parameter-reasoning-rl-with-10-gpus).
- [2025/10] verl is presented in the [PyTorch Conference 2025](https://pytorch.org/event/pytorch-conference-2025/).
- [2025/08] verl is presented in the [PyTorch Expert Exchange Webinar](https://www.youtube.com/watch?v=Vd79NmmqY3Q&t=2s). [Slides](https://github.com/eric-haibin-lin/verl-community/blob/main/slides/verl_talk_pytorch_2025_08.pdf) available.
- [2025/07] The [ReTool](https://arxiv.org/pdf/2504.11536) recipe is fully open sourced. [Blog](https://www.notion.so/verl-reTool-recipe-Using-multi-round-conversations-and-code-sandboxing-to-improve-the-math-of-large-23a8b5b7feba80b386b2e5b5e3c1cde0)
- [2025/07] The first verl meetup will be held at ICML Vancouver on July 16th! Please [join us](https://lu.ma/0ek2nyao) if you are at ICML! (onsite only)
- [2025/06] verl with Megatron backend enables large MoE models such as [DeepSeek-671B and Qwen3-235B](https://verl.readthedocs.io/en/latest/perf/dpsk.html).
- [2025/03] [DAPO](https://dapo-sia.github.io/) is the open-sourced SOTA RL algorithm that achieves 50 points on AIME 2024 based on the Qwen2.5-32B pre-trained model, surpassing the previous SOTA achieved by DeepSeek's GRPO (DeepSeek-R1-Zero-Qwen-32B). DAPO's training is fully powered by verl and the reproduction code is available in `recipe/dapo` now.
<details><summary> more... </summary>
<ul>
  <li>[2025/04] [Seed-Thinking-v1.5](https://github.com/ByteDance-Seed/Seed-Thinking-v1.5/blob/main/seed-thinking-v1.5.pdf) tech report is released! Trained with verl, Seed-Thinking-v1.5 achieves 86.7 on AIME 2024, 55.0 on Codeforces and 77.3 on GPQA, demonstrating excellent reasoning abilities in STEM and coding. Beyond reasoning tasks, the method demonstrates notable generalization across diverse domains.</li>
  <li>[2025/07] verl keynote at [AWS AI Hours Singapore](https://pages.awscloud.com/aws-ai-hours-sg.html#agenda) on 7/8, verl & verl-agent project updates at [Agent for SWE meetup](https://lu.ma/e498qhsi) by LF AI & Data Singapore on 7/11.</li>
  <li>[2025/06] verl team will provide latest project updates at [PyTorch Day China](https://www.lfasiallc.com/pytorch-day-china/) on June 7th. Meet our dev team in Beijing!</li>
  <li> [2025/04] [VAPO](https://arxiv.org/pdf/2504.05118) (value-based augmented PPO) paper covers our latest RL method for reasoning models. Trained from Qwen-32B-base model, VAPO achieves 60.4 on AIME 2024, outperforming DAPO-32B.</li>
  <li>[2025/05] [PF-PPO](https://arxiv.org/abs/2409.06957), accepted to ICML 2025, is now supported in verl! PF-PPO enhances policy learning efficiency and robustness by filtering potentially noisy reward signals and reusing high-quality experiences via a replay buffer.</li>
  <li>[2025/04] We will give a tutorial about latest post-training techniques and programming guide for verl at [ICLR 2025 Expo](https://iclr.cc/virtual/2025/calendar?filter_events=Expo+Talk+Panel&filter_rooms=), [SCI-FM workshop](https://open-foundation-model.github.io/) and [LMSys afterparty](https://lu.ma/d23nyynm). Talk materials available [here](https://github.com/eric-haibin-lin/verl-community/tree/main/iclr25). </li>
  <li>[2025/03] verl v0.3.0.post1 is released! See [release note](https://github.com/verl-project/verl/releases/) for details. It achieves [~1.4x speedup](https://tongyx361.github.io/blogs/posts/verl-intro/#/verl-flexible-and-efficient-rl-for-llms) compared to prev versions.</li>
  <li>[2025/05] verl will be presented at [A2M Shanghai](https://a2m.msup.com.cn/home/?aid=4488&city=shanghai) on 5/16 - 5/17.</li>
  <li>[2025/05] verl will be presented at [GOSIM x PyTorch Day 2025](https://paris2025.gosim.org/). See you in Paris! </li>
  <li>[2025/03] We introduced the programming model of verl at the [vLLM Beijing Meetup](https://mp.weixin.qq.com/s/n77GibL2corAtQHtVEAzfg) and [verl intro and updates](https://github.com/eric-haibin-lin/verl-community/blob/main/slides/verl-lmsys-meetup.pdf) at the [SGLang-LMSYS Org Meetup](https://lu.ma/ntjrr7ig) in Sunnyvale mid-March.</li>
  <li>[2025/03] We will present verl(HybridFlow) at EuroSys 2025. See you in Rotterdam!</li>
  <li>[2025/02] verl v0.2.0.post2 is released!</li>
  <li>[2025/02] We presented verl in the <a href="https://lu.ma/ji7atxux">Bytedance/NVIDIA/Anyscale Ray Meetup</a>. See you in San Jose!</li>
  <li>[2025/01] [Doubao-1.5-pro](https://team.doubao.com/zh/special/doubao_1_5_pro) is released with SOTA-level performance on LLM & VLM. The RL scaling preview model is trained using verl, reaching OpenAI O1-level performance on math benchmarks (70.0 pass@1 on AIME).</li>
  <li>[2024/12] verl is presented at Ray Forward 2024. Slides available <a href="https://github.com/eric-haibin-lin/verl-community/blob/main/slides/Ray_Forward_2024_%E5%B7%AB%E9%94%A1%E6%96%8C.pdf">here</a></li>
  <li>[2024/12] The team presented <a href="https://neurips.cc/Expo/Conferences/2024/workshop/100677">Post-training LLMs: From Algorithms to Infrastructure</a> at NeurIPS 2024. <a href="https://github.com/eric-haibin-lin/verl-data/tree/neurips">Slides</a> and <a href="https://neurips.cc/Expo/Conferences/2024/workshop/100677">video</a> available.</li>
  <li>[2024/10] verl is presented at Ray Summit. <a href="https://www.youtube.com/watch?v=MrhMcXkXvJU&list=PLzTswPQNepXntmT8jr9WaNfqQ60QwW7-U&index=37">Youtube video</a> available.</li>
  <li>[2024/08] HybridFlow (verl) is accepted to EuroSys 2025.</li>
</ul>
</details>

## Key Features

- **FSDP**, **FSDP2** and **Megatron-LM** for training.
- **vLLM**, **SGLang** and **HF Transformers** for rollout generation.
- Compatible with Hugging Face Transformers and Modelscope Hub: Qwen3.5, Qwen3, Qwen-2.5, Llama3.1, Gemma2, DeepSeek-LLM, etc
- Supervised fine-tuning.
- Reinforcement learning with [PPO](examples/ppo_trainer/), [GRPO](examples/grpo_trainer/), [GSPO](https://github.com/verl-project/verl-recipe/tree/main/gspo/), [ReMax](examples/remax_trainer/), [REINFORCE++](https://verl.readthedocs.io/en/latest/examples/config.html#algorithm), [RLOO](examples/rloo_trainer/), [PRIME](https://github.com/verl-project/verl-recipe/tree/main/prime/), [DAPO](https://github.com/verl-project/verl-recipe/tree/main/dapo/), [DrGRPO](https://github.com/verl-project/verl-recipe/tree/main/drgrpo), [KL_Cov & Clip_Cov](https://github.com/verl-project/verl-recipe/tree/main/entropy) etc.
  - Support model-based reward and function-based reward (verifiable reward) for math, [coding](https://github.com/verl-project/verl-recipe/tree/main/dapo), etc
  - Support vision-language models (VLMs) and [multi-modal RL](examples/grpo_trainer/run_qwen2_5_vl_7b_fsdp.sh) with Qwen2.5-vl, Kimi-VL
  - [Multi-turn with tool calling](examples/tutorial/agent_loop_get_started/)
- LLM alignment recipes such as [Self-play preference optimization (SPPO)](https://github.com/verl-project/verl-recipe/tree/main/sppo)
- Flash attention 2, sequence packing, sequence parallelism via DeepSpeed Ulysses, [LoRA](examples/tuning/lora/run_qwen3_8b_fsdp.sh), [Liger-kernel](examples/sft/gsm8k/run_qwen2_5_0_5b_fsdp.sh) (`USE_LIGER=1`).
- Scales up to 671B models and hundreds of GPUs with [expert parallelism](https://github.com/verl-project/verl/pull/1467)
- Multi-gpu [LoRA RL](https://verl.readthedocs.io/en/latest/advance/ppo_lora.html) support to save memory.
- Experiment tracking with wandb, swanlab, mlflow and tensorboard.
- Hardware Support: Supports NVIDIA, AMD, [Ascend](https://github.com/verl-project/verl/blob/main/docs/ascend_tutorial/README.md)

## Getting Started

<a href="https://verl.readthedocs.io/en/latest/index.html"><b>Documentation</b></a>

**Quickstart:**

- [Installation](https://verl.readthedocs.io/en/latest/start/install.html)
- [Quickstart](https://verl.readthedocs.io/en/latest/start/quickstart.html)
- [Programming Guide](https://verl.readthedocs.io/en/latest/hybrid_flow.html) & [Tech Talk](https://hcqnc.xetlk.com/sl/3vACOK) (in Chinese)
- [PPO in verl](https://verl.readthedocs.io/en/latest/algo/ppo.html)
- [GRPO in verl](https://verl.readthedocs.io/en/latest/algo/grpo.html)

**Running a PPO example step-by-step:**

- [Prepare Data for Post-Training](https://verl.readthedocs.io/en/latest/preparation/prepare_data.html)
- [Implement Reward Function for Dataset](https://verl.readthedocs.io/en/latest/preparation/reward_function.html)
- [PPO Example Architecture](https://verl.readthedocs.io/en/latest/examples/ppo_code_architecture.html)
- [Config Explanation](https://verl.readthedocs.io/en/latest/examples/config.html)

**Reproducible algorithm baselines:**

- [RL performance on coding, math](https://verl.readthedocs.io/en/latest/algo/baseline.html)

**Algorithm recipes (`recipe/`):**

- Optional workflows and baselines live under [`recipe/`](recipe/). Each recipe subdirectory includes a small **`REQUIRED_VERL.txt`** file describing the intended `verl` install: pinned recipes use a **tag or fixed git SHA**; rolling recipes record an explicit **`VERL_COMMIT`** (and related submodule / recipe-folder SHAs) so you can `pip install verl@git+…@<sha>` without guessing. See [`recipe/README.md`](recipe/README.md) for the full index and links.

**For code explanation and advance usage (extension):**

- PPO Trainer and Workers

- [PPO Ray Trainer](https://verl.readthedocs.io/en/latest/workers/ray_trainer.html)
  - [Model Engine](https://verl.readthedocs.io/en/latest/workers/model_engine.html)
  - [Engine Workers (FSDP / Megatron-LM / Automodel / VeOmni / TorchTitan)](https://verl.readthedocs.io/en/latest/workers/engine_workers.html)

- Advanced Usage and Extension
  - [Add Models with the FSDP Backend](https://verl.readthedocs.io/en/latest/advance/fsdp_extension.html)
  - [Add Models with the Megatron-LM Backend](https://verl.readthedocs.io/en/latest/advance/megatron_extension.html)
  - [Multi-turn Rollout Support](https://verl.readthedocs.io/en/latest/sglang_multiturn/multiturn.html)
  - [Search Tool Integration](https://verl.readthedocs.io/en/latest/sglang_multiturn/search_tool_example.html)
  - [Sandbox Fusion Integration](https://verl.readthedocs.io/en/latest/examples/sandbox_fusion_example.html)
  - [Extend to Other RL(HF) algorithms](https://verl.readthedocs.io/en/latest/advance/dpo_extension.html)
  - [Ray API design tutorial](https://verl.readthedocs.io/en/latest/advance/placement.html)

**Blogs from the community**

- [When Reasoning Models Break Tokenization: The Hidden Complexity of Multiturn Training](https://github.com/zhaochenyang20/Awesome-ML-SYS-Tutorial/blob/main/rlhf/verl/multi-turn/fast_tokenization/multiturn_tokenization_and_masking.md)
- [verl deployment on AWS SageMaker](https://medium.com/@kaige.yang0110/run-verl-on-sagemaker-using-4x8-l40s-gpus-8e6d5c3c61d3)
- [verl x SGLang Multi-turn Code Walkthrough](https://github.com/zhaochenyang20/Awesome-ML-SYS-Tutorial/blob/main/rlhf/verl/multi-turn/code-walk-through/readme_EN.md)
- [Optimizing SGLang Memory Usage in verl](https://hebiao064.github.io/rl-memory-management)
- [SGLang, verl, OpenBMB and Tsinghua University: Pioneering End-to-End Multi-Turn RLHF](https://github.com/zhaochenyang20/Awesome-ML-SYS-Tutorial/blob/main/rlhf/verl/multi-turn/verl-multiturn-rollout-Release.md)
- [Reinforcement Learning from Human Feedback on AMD GPUs with verl and ROCm Integration](https://rocm.blogs.amd.com/artificial-intelligence/verl-large-scale/README.html)
- [veMLP x verl ：玩转强化学习训练](https://mp.weixin.qq.com/s/7nbqxk4knMGd-hQE9ls2tA)
- [使用 verl 进行 GRPO 分布式强化学习训练最佳实践](https://www.volcengine.com/docs/6459/1463942)
- [HybridFlow verl 原文浅析](https://github.com/zhaochenyang20/Awesome-ML-SYS-Tutorial/blob/main/rlhf/verl/readme.md)
- [最高提升 20 倍吞吐量！豆包大模型团队发布全新 RLHF 框架，现已开源！](https://team.doubao.com/en/blog/%E6%9C%80%E9%AB%98%E6%8F%90%E5%8D%8720%E5%80%8D%E5%90%9E%E5%90%90%E9%87%8F-%E8%B1%86%E5%8C%85%E5%A4%A7%E6%A8%A1%E5%9E%8B%E5%9B%A2%E9%98%9F%E5%8F%91%E5%B8%83%E5%85%A8%E6%96%B0-rlhf-%E6%A1%86%E6%9E%B6-%E7%8E%B0%E5%B7%B2%E5%BC%80%E6%BA%90)

## Performance Tuning Guide

The performance is essential for on-policy RL algorithm. We have written a detailed [performance tuning guide](https://verl.readthedocs.io/en/latest/perf/perf_tuning.html) to help you optimize performance.

## Upgrade to vLLM >= v0.8.2

verl now supports vLLM>=0.8.2 when using FSDP as the training backend. Please refer to [this document](https://github.com/verl-project/verl/blob/main/docs/README_vllm0.8.md) for the installation guide and more information. Please avoid vllm 0.7.x, which contains bugs that may lead to OOMs and unexpected errors.

## Use Latest SGLang

SGLang is fully supported with verl, and SGLang RL Group is working extensively on building unique features, including multi-turn agentic RL, VLM RLHF, server-based RL, and partial rollout. Please refer to [this document](https://verl.readthedocs.io/en/latest/workers/sglang_worker.html) for the installation guide and more information.

## Upgrade to FSDP2

verl is fully embracing FSDP2! FSDP2 is recommended by torch distributed team, providing better throughput and memory usage, and is composible with other features (e.g. torch.compile). To enable FSDP2, simply use verl main and set the following options:

```
actor_rollout_ref.ref.strategy=fsdp2
actor_rollout_ref.actor.strategy=fsdp2
critic.strategy=fsdp2
```

Furthermore, FSDP2 cpu offloading is compatible with gradient accumulation. You can turn it on to save memory with `actor_rollout_ref.actor.fsdp_config.offload_policy=True`. For more details, see https://github.com/verl-project/verl/pull/1026

## AMD Support (ROCm Kernel)

verl runs on AMD ROCm GPUs (MI300X / MI325X / MI355X) with FSDP, FSDP2, and Megatron trainer backends, and vLLM as the validated inference engine (SGLang support is in progress). See the [AMD ROCm quick-start guide](https://github.com/verl-project/verl/blob/main/docs/amd_tutorial/amd_quick_start.rst) for container bring-up, environment verification, and training examples.

## Citation and acknowledgement

If you find the project helpful, please cite:

- [HybridFlow: A Flexible and Efficient RLHF Framework](https://arxiv.org/abs/2409.19256v2)
- [A Framework for Training Large Language Models for Code Generation via Proximal Policy Optimization](https://i.cs.hku.hk/~cwu/papers/gmsheng-NL2Code24.pdf)

```bibtex
@article{sheng2024hybridflow,
  title   = {HybridFlow: A Flexible and Efficient RLHF Framework},
  author  = {Guangming Sheng and Chi Zhang and Zilingfeng Ye and Xibin Wu and Wang Zhang and Ru Zhang and Yanghua Peng and Haibin Lin and Chuan Wu},
  year    = {2024},
  journal = {arXiv preprint arXiv: 2409.19256}
}
```

verl is inspired by the design of Nemo-Aligner, Deepspeed-chat and OpenRLHF. The project is adopted and contributed by Bytedance, Anyscale, LMSys.org, [Alibaba Qwen team](https://github.com/QwenLM/), Shanghai AI Lab, Tsinghua University, UC Berkeley, UCLA, UIUC, University of Hong Kong, ke.com, [All Hands AI](https://www.all-hands.dev/), [ModelBest](http://modelbest.cn/), JD AI Lab, Microsoft Research, [StepFun](https://www.stepfun.com/), Amazon, LinkedIn, Meituan, [Camel-AI](https://www.camel-ai.org/), [OpenManus](https://github.com/OpenManus), Xiaomi, NVIDIA research, [Baichuan](https://www.baichuan-ai.com/home), [RedNote](https://www.xiaohongshu.com/), [SwissAI](https://www.swiss-ai.org/), [Moonshot AI (Kimi)](https://www.moonshot-ai.com/), Baidu, Snowflake, Skywork.ai, JetBrains, [IceSword Lab](https://www.iceswordlab.com), and many more.

## Awesome Projects Built with `verl`

Welcome to register your awesome project build with `verl` for other developers' reference!

- [TinyZero](https://github.com/Jiayi-Pan/TinyZero): a reproduction of **DeepSeek R1 Zero** recipe for reasoning tasks ![GitHub Repo stars](https://img.shields.io/github/stars/Jiayi-Pan/TinyZero)
- [SkyThought](https://github.com/NovaSky-AI/SkyThought): RL training for Sky-T1-7B by NovaSky AI team. ![GitHub Repo stars](https://img.shields.io/github/stars/NovaSky-AI/SkyThought)
- [simpleRL-reason](https://github.com/hkust-nlp/simpleRL-reason): SimpleRL-Zoo: Investigating and Taming Zero Reinforcement Learning for Open Base Models in the Wild ![GitHub Repo stars](https://img.shields.io/github/stars/hkust-nlp/simpleRL-reason)
- [Easy-R1](https://github.com/hiyouga/EasyR1): **Multi-modal** RL training framework ![GitHub Repo stars](https://img.shields.io/github/stars/hiyouga/EasyR1)
- [RandOpt](https://github.com/sunrainyg/RandOpt): Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights ![GitHub Repo stars](https://img.shields.io/github/stars/sunrainyg/RandOpt)
- [OpenManus-RL](https://github.com/OpenManus/OpenManus-RL): LLM Agents RL tuning framework for multiple agent environments. ![GitHub Repo stars](https://img.shields.io/github/stars/OpenManus/OpenManus-RL)
- [rllm](https://github.com/agentica-project/rllm): async RL training with [verl-pipeline](https://github.com/agentica-project/verl-pipeline) ![GitHub Repo stars](https://img.shields.io/github/stars/agentica-project/rllm)
- [RAGEN](https://github.com/ZihanWang314/ragen): a general-purpose reasoning **agent** training framework ![GitHub Repo stars](https://img.shields.io/github/stars/ZihanWang314/ragen)
- [Search-R1](https://github.com/PeterGriffinJin/Search-R1): RL with reasoning and **searching (tool-call)** interleaved LLMs ![GitHub Repo stars](https://img.shields.io/github/stars/PeterGriffinJin/Search-R1)
- [ReSearch](https://github.com/Agent-RL/ReSearch): Learning to **Re**ason with **Search** for LLMs via Reinforcement Learning ![GitHub Repo stars](https://img.shields.io/github/stars/Agent-RL/ReSearch)
- [Skywork-OR1](https://github.com/SkyworkAI/Skywork-OR1): Skywork open reaonser series ![GitHub Repo stars](https://img.shields.io/github/stars/SkyworkAI/Skywork-OR1)
- [ToRL](https://github.com/GAIR-NLP/ToRL): Scaling tool-integrated RL ![GitHub Repo stars](https://img.shields.io/github/stars/GAIR-NLP/ToRL)
- [Absolute Zero Reasoner](https://github.com/LeapLabTHU/Absolute-Zero-Reasoner): [A no human curated data self-play framework for reasoning](https://arxiv.org/abs/2505.03335) ![GitHub Repo stars](https://img.shields.io/github/stars/LeapLabTHU/Absolute-Zero-Reasoner)
- [verl-agent](https://github.com/langfengQ/verl-agent): A scalable training framework for **long-horizon LLM/VLM agents**, along with a new algorithm **GiGPO** ![GitHub Repo stars](https://img.shields.io/github/stars/langfengQ/verl-agent)
- [RL-Factory](https://github.com/Simple-Efficient/RL-Factory): An easy and efficient RL post-training framework for Agentic Learning ![GitHub Repo stars](https://img.shields.io/github/stars/Simple-Efficient/RL-Factory)
- [ReTool](https://retool-rl.github.io/): ReTool: reinforcement learning for strategic tool use in LLMs. Code release is in progress...
- [verl-tool](https://github.com/TIGER-AI-Lab/verl-tool): An unified and easy-to-extend tool-agent training framework based on verl![GitHub Repo stars](https://img.shields.io/github/stars/TIGER-AI-Lab/verl-tool)
- [PRIME](https://github.com/PRIME-RL/PRIME): Process reinforcement through implicit rewards ![GitHub Repo stars](https://img.shields.io/github/stars/PRIME-RL/PRIME)
- [MemAgent](https://github.com/BytedTsinghua-SIA/MemAgent): MemAgent: Reshaping Long-Context LLM with Multi-Conv RL based Memory Agent ![GitHub Repo stars](https://img.shields.io/github/stars/BytedTsinghua-SIA/MemAgent)
- [POLARIS](https://github.com/ChenxinAn-fdu/POLARIS): A Post-training recipe for scaling RL on Advanced Reasoning models ![GitHub Repo stars](https://img.shields.io/github/stars/ChenxinAn-fdu/POLARIS)
- [GUI-R1](https://github.com/ritzz-ai/GUI-R1): **GUI-R1**: A Generalist R1-style Vision-Language Action Model For **GUI Agents** ![GitHub Repo stars](https://img.shields.io/github/stars/ritzz-ai/GUI-R1)
- [DeepRetrieval](https://github.com/pat-jj/DeepRetrieval): RL Training of **Search Agent** with **Search/Retrieval Outcome** ![GitHub Repo stars](https://img.shields.io/github/stars/pat-jj/DeepRetrieval)
- [Code-R1](https://github.com/ganler/code-r1): Reproducing R1 for **Code** with Reliable Rewards ![GitHub Repo stars](https://img.shields.io/github/stars/ganler/code-r1)
- [DeepResearcher](https://github.com/GAIR-NLP/DeepResearcher): Scaling deep research via reinforcement learning in real-world environments ![GitHub Repo stars](https://img.shields.io/github/stars/GAIR-NLP/DeepResearcher)
- [VAGEN](https://github.com/RAGEN-AI/VAGEN): Training VLM agents with multi-turn reinforcement learning ![GitHub Repo stars](https://img.shields.io/github/stars/RAGEN-AI/VAGEN)
- [RM-R1](https://arxiv.org/abs/2505.02387): RL training of reasoning reward models ![GitHub Repo stars](https://img.shields.io/github/stars/RM-R1-UIUC/RM-R1)
- [Dr. MAS](https://arxiv.org/pdf/2602.08847): Stable **end-to-end RL** post-training for **multi-agent LLM systems** ![GitHub Repo stars](https://img.shields.io/github/stars/langfengQ/DrMAS)
- [LUFFY](https://arxiv.org/pdf/2504.14945): Learning to Reason under Off-Policy Guidance![GitHub Repo stars](https://img.shields.io/github/stars/ElliottYan/LUFFY)
- [DeepMath](https://github.com/zwhe99/DeepMath): DeepMath-103K data and series models for math reasoning![GitHub Repo stars](https://img.shields.io/github/stars/zwhe99/DeepMath)
- [PACS](https://github.com/ritzz-ai/PACS): Implicit Actor Critic Coupling via a Supervised Learning Framework for RLVR ![GitHub Repo stars](https://img.shields.io/github/stars/ritzz-ai/PACS)
- [Entropy Mechanism of RL](https://github.com/PRIME-RL/Entropy-Mechanism-of-RL): The Entropy Mechanism of Reinforcement Learning for Large Language Model Reasoning![GitHub Repo stars](https://img.shields.io/github/stars/PRIME-RL/Entropy-Mechanism-of-RL)
- [LLaSA-TTS-GRPO](https://github.com/channel-io/ch-tts-llasa-rl-grpo): TTS fine-tuning with GRPO optimization based on LLASA models ![GitHub Repo stars](https://img.shields.io/github/stars/channel-io/ch-tts-llasa-rl-grpo)
- [PF-PPO](https://arxiv.org/abs/2409.06957): Policy Filtration for PPO based on the reliability of reward signals for more efficient and robust RLHF.
- [RACRO](https://github.com/gyhdog99/RACRO2): Build multi-modal reasoning models via decoupling it into query-conditioned captioning and text-only reasoning ![GitHub Repo stars](https://img.shields.io/github/stars/gyhdog99/RACRO2)
- [Agent Lightning](https://github.com/microsoft/agent-lightning): A flexible and extensible framework that enables seamless agent optimization for any existing agent framework. ![GitHub Repo stars](https://img.shields.io/github/stars/microsoft/agent-lightning)
- [VTool-R1](https://github.com/VTOOL-R1/vtool-r1): VLMs Learn to Think with Images via Reinforcement Learning on Multimodal Tool Use. ![GitHub Repo stars](https://img.shields.io/github/stars/VTOOL-R1/vtool-r1)
- [Kimina-Prover-RL](https://github.com/project-numina/kimina-prover-rl/tree/main/recipe/kimina_prover_rl): Training pipeline for formal theorem proving, based on a paradigm inspired by DeepSeek-R1.
- [RL-PLUS](https://github.com/YihongDong/RL-PLUS): Countering Capability Boundary Collapse of LLMs in Reinforcement Learning with Hybrid-policy Optimization.
- [rStar2-Agent](https://github.com/microsoft/rStar): Using reinforcement learning with multi-step tool-calling for math tasks, rStar2-Agent-14B reaches frontier-level math reasoning in just 510 RL training steps ![GitHub Repo stars](https://img.shields.io/github/stars/microsoft/rStar)
- [Vision-SR1](https://github.com/zli12321/Vision-SR1): Self-Rewarding Vision-Language Model via Reasoning Decomposition ![GitHub Repo stars](https://img.shields.io/github/stars/zli12321/Vision-SR1)
- [SimpleVLA-RL](https://github.com/PRIME-RL/SimpleVLA-RL): SimpleVLA-RL: A Simple yet Effective Vision-Language Action Model for Reinforcement Learning ![GitHub Repo stars](https://img.shields.io/github/stars/PRIME-RL/SimpleVLA-RL)
- [Table-R1](https://github.com/Table-R1/Table-R1): Table-R1: Inference-Time Scaling for Table Reasoning ![GitHub Repo stars](https://img.shields.io/github/stars/Table-R1/Table-R1)
- [Revisual-R1](https://github.com/CSfufu/Revisual-R1): Revisual-R1: Advancing Multimodal Reasoning From Optimized Cold Start to Staged Reinforcement Learning ![GitHub Repo stars](https://img.shields.io/github/stars/CSfufu/Revisual-R1)
- [ARES](https://github.com/shawn0728/ARES): ARES: Multimodal Adaptive Reasoning via Difficulty-Aware Token-Level Entropy Shaping ![GitHub Repo stars](https://img.shields.io/github/stars/shawn0728/ARES)
- [Meta-Bandit-LLM](https://github.com/sanxing-chen/meta-bandit-llm): Meta-Bandit-LLM: Long-horizon multiturn interactive training for meta-bandit agents ![GitHub Repo stars](https://img.shields.io/github/stars/sanxing-chen/meta-bandit-llm)
- [PokeeResearch](https://github.com/Pokee-AI/PokeeResearchOSS): PokeeResearch: State-of-the-art 7B DeepResearch Agent that leverages web search and content reading capabilities to answer complex questions using the most up-to-date information available online. ![Github Repo Stars](https://img.shields.io/github/stars/Pokee-AI/PokeeResearchOSS)
- [Search Self-play](https://github.com/Alibaba-Quark/SSP): Pushing the Frontier of Agent Capability without Supervision ![GitHub Repo stars](https://img.shields.io/github/stars/Alibaba-Quark/SSP)
- [OneThinker](https://github.com/tulerfeng/OneThinker): All-in-one Reasoning Model for Image and Video ![GitHub Repo stars](https://img.shields.io/github/stars/tulerfeng/OneThinker)
- [OpenTinker](https://github.com/open-tinker/OpenTinker): Democratizing Agentic Reinforcement Learning as a Service ![GitHub Repo stars](https://img.shields.io/github/stars/open-tinker/OpenTinker)
- [FlowRL](https://github.com/Xuekai-Zhu/FlowRL): Matching reward distributions via **flow balance** for diverse exploration and generalizable reasoning ![GitHub Repo stars](https://img.shields.io/github/stars/Xuekai-Zhu/FlowRL)
- [Logic-RL](https://github.com/Unakar/Logic-RL): a reproduction of DeepSeek R1 Zero on 2K Tiny Logic Puzzle Dataset. ![GitHub Repo stars](https://img.shields.io/github/stars/Unakar/Logic-RL)
- [Seed-Coder](https://github.com/ByteDance-Seed/Seed-Coder): RL training of Seed-Coder boosts performance on competitive programming ![GitHub Repo stars](https://img.shields.io/github/stars/ByteDance-Seed/Seed-Coder)
- [all-hands/openhands-lm-32b-v0.1](https://www.all-hands.dev/blog/introducing-openhands-lm-32b----a-strong-open-coding-agent-model): A strong, open coding agent model, trained with [multi-turn fine-tuning](https://github.com/verl-project/verl/pull/195)
- [s3](https://github.com/pat-jj/s3) **Efficient Yet Effective** Search Agent Training via RL ![GitHub Repo stars](https://img.shields.io/github/stars/pat-jj/s3)
- [Rec-R1](https://arxiv.org/pdf/2503.24289): Bridging Generative Large Language Models and Recommendation Systems via Reinforcement Learning
- [Explore RL Data Scaling](https://arxiv.org/abs/2503.22230): Exploring Data Scaling Trends and Effects in Reinforcement Learning from Human Feedback
- [FIRE](https://arxiv.org/abs/2410.21236): Flaming-hot initiation with regular execution sampling for large language models
- [DQO](https://arxiv.org/abs/2410.09302): Enhancing multi-Step reasoning abilities of language models through direct Q-function optimization
- [ProRL](https://arxiv.org/abs/2505.24864): Prolonged Reinforcement Learning Expands Reasoning Boundaries in Large Language Models
- [cognition-engineering](https://github.com/gair-nlp/cognition-engineering): Test time scaling drives cognition engineering. ![GitHub Repo stars](https://img.shields.io/github/stars/gair-nlp/cognition-engineering)
- [Trust Region Preference Approximation](https://github.com/XueruiSu/Trust-Region-Preference-Approximation): A simple and stable **reinforcement learning algorithm** for LLM reasoning. ![GitHub Repo stars](https://img.shields.io/github/stars/XueruiSu/Trust-Region-Preference-Approximation)
- [AdaRFT](https://github.com/uscnlp-lime/verl): Efficient Reinforcement Finetuning via **Adaptive Curriculum Learning** ![GitHub Repo stars](https://img.shields.io/github/stars/uscnlp-lime/verl)
- [critic-rl](https://github.com/HKUNLP/critic-rl): LLM critics for code generation ![GitHub Repo stars](https://img.shields.io/github/stars/HKUNLP/critic-rl)
- [self-rewarding-reasoning-LLM](https://arxiv.org/pdf/2502.19613): self-rewarding and correction with **generative reward models** ![GitHub Repo stars](https://img.shields.io/github/stars/RLHFlow/Self-rewarding-reasoning-LLM)
- [DeepEnlighten](https://github.com/DolbyUUU/DeepEnlighten): Reproduce R1 with **social reasoning** tasks and analyze key findings ![GitHub Repo stars](https://img.shields.io/github/stars/DolbyUUU/DeepEnlighten)
- [MetaSpatial](https://github.com/PzySeere/MetaSpatial): Reinforcing **3D Spatial Reasoning** in **VLMs** for the **Metaverse** ![GitHub Repo stars](https://img.shields.io/github/stars/PzySeere/MetaSpatial)
- [PURE](https://github.com/CJReinforce/PURE): **Credit assignment** is the key to successful reinforcement fine-tuning using **process reward model** ![GitHub Repo stars](https://img.shields.io/github/stars/CJReinforce/PURE)
- [cognitive-behaviors](https://github.com/kanishkg/cognitive-behaviors): Cognitive Behaviors that Enable Self-Improving Reasoners, or, Four Habits of Highly Effective STaRs ![GitHub Repo stars](https://img.shields.io/github/stars/kanishkg/cognitive-behaviors)
- [deepscaler](https://github.com/agentica-project/rllm/tree/deepscaler): iterative context scaling with GRPO ![GitHub Repo stars](https://img.shields.io/github/stars/agentica-project/deepscaler)
- [DAPO](https://dapo-sia.github.io/): the fully open source SOTA RL algorithm that beats DeepSeek-R1-zero-32B ![GitHub Repo stars](https://img.shields.io/github/stars/verl-project/verl)
- [NoisyRollout](https://github.com/NUS-TRAIL/NoisyRollout): Reinforcing Visual Reasoning with Data Augmentation ![GitHub Repo stars](https://img.shields.io/github/stars/NUS-TRAIL/NoisyRollout)
- [SPEAR](https://github.com/TencentYoutuResearch/SPEAR): **Self-imitation** with **Progressive Exploration** for Agentic Reinforcement Learning (ICLR 2026) ![GitHub Repo stars](https://img.shields.io/github/stars/TencentYoutuResearch/SPEAR)
- [RuleReasoner](https://github.com/bigai-nlco/RuleReasoner): **RuleReasoner:** Reinforced Rule-based Reasoning via **Domain-aware Dynamic Sampling** (ICLR 2026) ![GitHub Repo stars](https://img.shields.io/github/stars/bigai-nlco/RuleReasoner)
- [MetaphorStar](https://metaphorstar.github.io/): **Image Metaphor** Understanding and Reasoning with End-to-End **Visual Reinforcement Learning** ![GitHub Repo stars](https://img.shields.io/github/stars/MING-ZCH/MetaphorStar)
- [DART-GUI](https://github.com/Computer-use-agents/dart-gui): a decoupled agentic RL framework for Computer Use Agents, achieving ~2× training speedup and ~5× environment utilization! ![GitHub Repo stars](https://img.shields.io/github/stars/Computer-use-agents/dart-gui)
- [Rethinking OPD](https://github.com/thunlp/OPD): Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe ![GitHub Repo stars](https://img.shields.io/github/stars/thunlp/OPD)

## Contribution Guide

See [contributions guide](CONTRIBUTING.md)

## About [ByteDance Seed Team](https://team.doubao.com/)

Founded in 2023, ByteDance Seed Team is dedicated to crafting the industry's most advanced AI foundation models. The team aspires to become a world-class research team and make significant contributions to the advancement of science and society. You can get to know Bytedance Seed better through the following channels👇

We are HIRING! Send us an [email](mailto:the.verl.project@gmail.com) if you are interested in internship/FTE opportunities in RL for agents.

verl

File Explorer

Today's picks

Was this content helpful?

Recent Feedback

verl

File Explorer

Today's picks

Was this content helpful?

Recent Feedback

Download README