RAGEN
RAGEN leverages reinforcement learning to train LLM reasoning agents in interactive, stochastic environments.
파일 탐색기
최종 버전 다운로드 (.zip)- reward_hacking.txt
- suck_moment.txt
- _10_deepcoder.yaml
- _10_rubikscube.yaml
- _1_bandit.yaml
- _2_sokoban.yaml
- _3_frozen_lake.yaml
- _4_countdown.yaml
- _5_metamathqa.yaml
- _6_webshop.yaml
- _7_lean.yaml
- _8_sudoku.yaml
- _9_2048.yaml
- _9_search.yaml
- base.yaml
- envs.yaml
- eval.yaml
- eval_webshop.yaml
- evaluate_api_llm.yaml
- stream.yaml
- webshop_full.yaml
- eval.md
- experiment_deepcoder.md
- experiment_frozen_lake_slipper_sweep.md
- experiment_intervention_sweep.md
- experiment_main_table.md
- experiment_search.md
- experiment_sokoban_gradient_analysis.md
- experiment_webshop_release.md
- guide_filtering_and_loss_scaling.md
- guide_gradient_analysis.md
- guide_rollout_filtering.md
- readme_v1.md
- reference_mutual_information_metrics.md
- kimina-lean-server
- webshop-minimal
- plot_gradient_analysis.py
- plot_icml_steps.py
- README.md
- fsdp_checkpoint_manager.py
- README.md
- collapse_types.png
- exp1.png
- exp2.png
- exp3.png
- exp4.png
- exp5.png
- exp6.png
- framework.png
- loss_curve.png
- main_results.png
- ragen.png
- ragen_logo.jpeg
- rico.png
- star-history-202556.png
- starpo_logo.png
- step_1.png
- step_2.png
- teaser.png
- top_p.png
- top_p_top_k_no_filter.png
- __init__.py
- run.py
- alfworld_config.yaml
- config.py
- env.py
- utils.py
- __init__.py
- config.py
- env.py
- __init__.py
- config.py
- env.py
- __init__.py
- config.py
- env.py
- README.md
- run.sh
- utils.py
- __init__.py
- config.py
- env.py
- utils.py
- __init__.py
- config.py
- env.py
- __init__.py
- config.py
- env.py
- __init__.py
- config.py
- env.py
- __init__.py
- config.py
- env.py
- __init__.py
- config.py
- env.py
- README.md
- retrieval_client.py
- reward.py
- __init__.py
- config.py
- env.py
- utils.py
- Base
- config.py
- env.py
- env_old.py
- prompter.py
- prompts.py
- config.py
- env.py
- utils.py
- __init__.py
- config.py
- env.py
- utils.py
- __init__.py
- config.py
- env.py
- __init__.py
- base.py
- __init__.py
- alfworld_memory.py
- base.py
- factory.py
- simple_memory.py
- __init__.py
- agent_proxy.py
- ap_webshop.py
- base_llm.py
- ctx_manager.py
- es_manager.py
- __init__.py
- omega_conf_patch.py
- agent_trainer.py
- collapse_metrics.py
- core_algos.py
- gradient_reporter.py
- rollout_filter.py
- __init__.py
- dp_actor.py
- __init__.py
- dp_critic.py
- base.py
- fsdp_vllm.py
- __init__.py
- fsdp_workers.py
- __init__.py
- eval.py
- eval_api.py
- utils.py
- SimpleSokoban.txt
- test.ipynb
- launch_server.sh
- server.py
- plot_reward_matrix.py
- run_inference.py
- README_webshop_small_combos.md
- run_entropy_sweep.sh
- run_filtering_final.sh
- run_frozen_lake_slipper_rate_sweep.sh
- run_kl_sweep.sh
- run_main_table_diff_algo.sh
- run_main_table_diff_model.sh
- run_main_table_diff_size.sh
- run_search_benchmark.sh
- run_sokoban_ppo_filter_grad_analysis.sh
- run_sokoban_ppo_filter_grad_analysis_probe_ckpt.sh
- run_top_p_sweep.sh
- run_webshop_release_combos.sh
- run_webshop_small_combos.sh
- convert_to_jsonl.py
- debug_rollout.py
- download_data.py
- download_search_index.py
- eval_batch.sh
- eval_qwen_7b_sokoban.sh
- eval_qwen_sokoban.sh
- eval_reward_dist.py
- eval_selected_envs.sh
- prepare_search_data.py
- README_EVAL_SCRIPTS.md
- setup_ragen.md
- setup_ragen.sh
- setup_ragen_old.sh
- setup_webshop.sh
- visualize.py
- test_sokoban_render.py
- test_seed_iteration.py
- test_context_window.py
- test_rollout_filter.py
- verify_lower_ratio.py
- .gitignore
- .gitmodules
- LICENSE
- pytest.ini
- RAGEN-2.pdf
- RAGEN.pdf
- README.md
- setup.py
- train.py
- verl
# 설치 가이드
1. 코드 내려받기
git clone https://github.com/mll-lab-nu/RAGEN
깃허브에서 프로젝트 코드 전체를 내 컴퓨터로 내려받습니다.
cd RAGEN
방금 내려받은 프로젝트 폴더 안으로 이동합니다.
// repository documentation
Was this content helpful?
(0 ratings)
