agilerl 2.8.0.dev0__tar.gz → 2.8.0.dev1__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (518) hide show
  1. agilerl-2.8.0.dev1/.gitattributes +3 -0
  2. agilerl-2.8.0.dev1/.github/ISSUE_TEMPLATE/bug_report.md +27 -0
  3. agilerl-2.8.0.dev1/.github/ISSUE_TEMPLATE/feature_request.md +20 -0
  4. agilerl-2.8.0.dev1/.github/PULL_REQUEST_TEMPLATE/pull_request_template.md +34 -0
  5. agilerl-2.8.0.dev1/.github/badges/arena-github-badge.svg +24 -0
  6. agilerl-2.8.0.dev1/.github/codeql/install_codeql.sh +51 -0
  7. agilerl-2.8.0.dev1/.github/codeql/run_codeql.py +139 -0
  8. agilerl-2.8.0.dev1/.github/dependabot.yml +8 -0
  9. agilerl-2.8.0.dev1/.github/workflows/codeql.yml +101 -0
  10. agilerl-2.8.0.dev1/.github/workflows/linux-tests.yml +94 -0
  11. agilerl-2.8.0.dev1/.github/workflows/macos-tests.yml +50 -0
  12. agilerl-2.8.0.dev1/.github/workflows/windows-tests.yml +51 -0
  13. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/.gitignore +0 -11
  14. agilerl-2.8.0.dev1/.pre-commit-config.yaml +65 -0
  15. agilerl-2.8.0.dev1/.readthedocs.yaml +30 -0
  16. agilerl-2.8.0.dev1/CITATION.cff +20 -0
  17. agilerl-2.8.0.dev1/CODE_OF_CONDUCT.md +128 -0
  18. agilerl-2.8.0.dev1/CONTRIBUTING.md +73 -0
  19. agilerl-2.8.0.dev1/DQN_LEARNING_ALGORITHM_ANALYSIS.md +309 -0
  20. agilerl-2.8.0.dev1/DQN_LEARNING_ANALYSIS.md +168 -0
  21. agilerl-2.8.0.dev1/GPU_CLEANUP_ANALYSIS.md +541 -0
  22. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/PKG-INFO +199 -299
  23. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/README.md +197 -290
  24. agilerl-2.8.0.dev1/agilerl/__init__.py +34 -0
  25. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/cispo.py +3 -6
  26. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/core/base.py +161 -429
  27. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/core/registry.py +2 -27
  28. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/cqn.py +19 -37
  29. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/ddpg.py +21 -34
  30. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/dpo.py +19 -61
  31. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/dqn.py +25 -48
  32. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/dqn_rainbow.py +19 -36
  33. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/grpo.py +22 -64
  34. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/gspo.py +3 -6
  35. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/ilql.py +1 -3
  36. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/ippo.py +35 -65
  37. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/maddpg.py +28 -46
  38. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/matd3.py +30 -61
  39. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/neural_ts_bandit.py +20 -37
  40. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/neural_ucb_bandit.py +24 -39
  41. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/ppo.py +251 -109
  42. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/ppo_llm.py +28 -71
  43. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/reinforce_llm.py +27 -55
  44. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/sft.py +16 -32
  45. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/td3.py +20 -42
  46. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/components/__init__.py +2 -0
  47. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/components/data.py +1 -50
  48. agilerl-2.8.0.dev1/agilerl/components/multi_agent_replay_buffer.py +242 -0
  49. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/components/replay_buffer.py +51 -98
  50. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/components/rollout_buffer.py +7 -2
  51. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/components/sampler.py +10 -4
  52. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/components/segment_tree.py +5 -65
  53. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/hpo/mutation.py +17 -6
  54. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/hpo/tournament.py +22 -33
  55. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/llm_envs/preference.py +1 -19
  56. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/llm_envs/reasoning.py +1 -19
  57. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/llm_envs/sft.py +4 -42
  58. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/cnn.py +1 -1
  59. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/dummy.py +5 -1
  60. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/networks/actors.py +15 -14
  61. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/networks/base.py +1 -1
  62. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/networks/q_networks.py +6 -6
  63. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/networks/value_networks.py +2 -2
  64. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/protocols.py +0 -32
  65. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/rollouts/on_policy.py +6 -0
  66. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/training/train_bandits.py +169 -95
  67. agilerl-2.8.0.dev1/agilerl/training/train_llm.py +1908 -0
  68. agilerl-2.8.0.dev1/agilerl/training/train_multi_agent_off_policy.py +612 -0
  69. agilerl-2.8.0.dev1/agilerl/training/train_multi_agent_on_policy.py +623 -0
  70. agilerl-2.8.0.dev1/agilerl/training/train_off_policy.py +617 -0
  71. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/training/train_offline.py +156 -106
  72. agilerl-2.8.0.dev1/agilerl/training/train_on_policy.py +511 -0
  73. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/typing.py +3 -3
  74. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/algo_utils.py +105 -171
  75. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/llm_utils.py +4 -102
  76. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/minari_utils.py +6 -26
  77. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/probe_envs.py +48 -3
  78. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/probe_envs_ma.py +10 -149
  79. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/utils.py +130 -134
  80. agilerl-2.8.0.dev1/agilerl/wrappers/__init__.py +0 -0
  81. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/wrappers/agent.py +114 -110
  82. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/wrappers/learning.py +13 -62
  83. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/wrappers/pettingzoo_wrappers.py +1 -60
  84. agilerl-2.8.0.dev1/benchmarking/benchmarking_bandits.py +146 -0
  85. agilerl-2.8.0.dev1/benchmarking/benchmarking_llm_multiturn.py +156 -0
  86. agilerl-2.8.0.dev1/benchmarking/benchmarking_llm_preference.py +176 -0
  87. agilerl-2.8.0.dev1/benchmarking/benchmarking_llm_reasoning.py +198 -0
  88. agilerl-2.8.0.dev1/benchmarking/benchmarking_multi_agent_off_policy.py +224 -0
  89. agilerl-2.8.0.dev1/benchmarking/benchmarking_multi_agent_on_policy.py +169 -0
  90. agilerl-2.8.0.dev1/benchmarking/benchmarking_off_policy.py +179 -0
  91. agilerl-2.8.0.dev1/benchmarking/benchmarking_off_policy_distributed.py +107 -0
  92. agilerl-2.8.0.dev1/benchmarking/benchmarking_offline.py +148 -0
  93. agilerl-2.8.0.dev1/benchmarking/benchmarking_offline_distributed.py +144 -0
  94. agilerl-2.8.0.dev1/benchmarking/benchmarking_on_policy.py +118 -0
  95. agilerl-2.8.0.dev1/benchmarking/benchmarking_rainbow.py +175 -0
  96. agilerl-2.8.0.dev1/benchmarking/benchmarking_recurrent.py +176 -0
  97. agilerl-2.8.0.dev1/benchmarking/benchmarking_resnet.py +110 -0
  98. agilerl-2.8.0.dev1/benchmarking/benchmarking_sft.py +172 -0
  99. agilerl-2.8.0.dev1/benchmarking/benchmarking_simba.py +131 -0
  100. agilerl-2.8.0.dev1/benchmarking/configs/ds_config.json +16 -0
  101. agilerl-2.8.0.dev1/benchmarking/make_evolvable_benchmarking.py +490 -0
  102. agilerl-2.8.0.dev1/benchmarking/networks.py +569 -0
  103. agilerl-2.8.0.dev1/configs/accelerate/accelerate.yaml +13 -0
  104. agilerl-2.8.0.dev1/configs/accelerate/bench_accelerate_config.yaml +25 -0
  105. agilerl-2.8.0.dev1/configs/accelerate/grpo_accelerate_config.yaml +25 -0
  106. agilerl-2.8.0.dev1/configs/training/bandit/neural_ts.yaml +53 -0
  107. agilerl-2.8.0.dev1/configs/training/bandit/neural_ucb.yaml +53 -0
  108. agilerl-2.8.0.dev1/configs/training/cqn.yaml +55 -0
  109. agilerl-2.8.0.dev1/configs/training/ddpg/ddpg.yaml +70 -0
  110. agilerl-2.8.0.dev1/configs/training/ddpg/ddpg_lstm.yaml +69 -0
  111. agilerl-2.8.0.dev1/configs/training/ddpg/ddpg_simba.yaml +71 -0
  112. agilerl-2.8.0.dev1/configs/training/dqn/dqn.yaml +56 -0
  113. agilerl-2.8.0.dev1/configs/training/dqn/dqn_lstm.yaml +64 -0
  114. agilerl-2.8.0.dev1/configs/training/dqn/dqn_rainbow.yaml +69 -0
  115. agilerl-2.8.0.dev1/configs/training/llm_finetuning/cispo.yaml +47 -0
  116. agilerl-2.8.0.dev1/configs/training/llm_finetuning/cispo_quant_bench.yaml +114 -0
  117. agilerl-2.8.0.dev1/configs/training/llm_finetuning/cispo_quant_bench_qwen.yaml +69 -0
  118. agilerl-2.8.0.dev1/configs/training/llm_finetuning/dpo.yaml +50 -0
  119. agilerl-2.8.0.dev1/configs/training/llm_finetuning/grpo.yaml +42 -0
  120. agilerl-2.8.0.dev1/configs/training/llm_finetuning/grpo_multiturn.yaml +41 -0
  121. agilerl-2.8.0.dev1/configs/training/llm_finetuning/gspo.yaml +45 -0
  122. agilerl-2.8.0.dev1/configs/training/llm_finetuning/ppo_llm.yaml +44 -0
  123. agilerl-2.8.0.dev1/configs/training/llm_finetuning/ppo_llm_quant_bench.yaml +68 -0
  124. agilerl-2.8.0.dev1/configs/training/llm_finetuning/reinforce_llm.yaml +40 -0
  125. agilerl-2.8.0.dev1/configs/training/llm_finetuning/reinforce_quant_bench.yaml +80 -0
  126. agilerl-2.8.0.dev1/configs/training/multi_agent/ippo.yaml +78 -0
  127. agilerl-2.8.0.dev1/configs/training/multi_agent/ippo_pong.yaml +69 -0
  128. agilerl-2.8.0.dev1/configs/training/multi_agent/maddpg.yaml +94 -0
  129. agilerl-2.8.0.dev1/configs/training/multi_agent/matd3.yaml +70 -0
  130. agilerl-2.8.0.dev1/configs/training/multi_input.yaml +64 -0
  131. agilerl-2.8.0.dev1/configs/training/ppo/ppo.yaml +73 -0
  132. agilerl-2.8.0.dev1/configs/training/ppo/ppo_image.yaml +86 -0
  133. agilerl-2.8.0.dev1/configs/training/ppo/ppo_recurrent.yaml +65 -0
  134. agilerl-2.8.0.dev1/configs/training/sft.yaml +43 -0
  135. agilerl-2.8.0.dev1/configs/training/td3.yaml +64 -0
  136. agilerl-2.8.0.dev1/data/cartpole/cartpole_random_v1.1.0.h5 +0 -0
  137. agilerl-2.8.0.dev1/data/cartpole/cartpole_v1.1.0.h5 +0 -0
  138. agilerl-2.8.0.dev1/data/pendulum/pendulum_random_v1.1.0.h5 +0 -0
  139. agilerl-2.8.0.dev1/data/pendulum/pendulum_v1.1.0.h5 +0 -0
  140. agilerl-2.8.0.dev1/debugging/__init__.py +0 -0
  141. agilerl-2.8.0.dev1/demos/bandits/demo_bandit.py +188 -0
  142. agilerl-2.8.0.dev1/demos/llm/debugging/config_load.py +22 -0
  143. agilerl-2.8.0.dev1/demos/llm/debugging/configs/grpo_constant_target.yaml +24 -0
  144. agilerl-2.8.0.dev1/demos/llm/debugging/configs/grpo_grid_navigation.yaml +26 -0
  145. agilerl-2.8.0.dev1/demos/llm/debugging/configs/ppo_conditional_target.yaml +31 -0
  146. agilerl-2.8.0.dev1/demos/llm/debugging/configs/ppo_constant_target.yaml +27 -0
  147. agilerl-2.8.0.dev1/demos/llm/debugging/configs/ppo_grid_navigation.yaml +29 -0
  148. agilerl-2.8.0.dev1/demos/llm/debugging/configs/ppo_multi_input.yaml +28 -0
  149. agilerl-2.8.0.dev1/demos/llm/debugging/configs/ppo_value_head.yaml +30 -0
  150. agilerl-2.8.0.dev1/demos/llm/debugging/debugging_llm.py +183 -0
  151. agilerl-2.8.0.dev1/demos/llm/debugging/debugging_llm_stage_1.py +249 -0
  152. agilerl-2.8.0.dev1/demos/llm/debugging/debugging_llm_stage_2.py +280 -0
  153. agilerl-2.8.0.dev1/demos/llm/debugging/debugging_llm_stage_3.py +395 -0
  154. agilerl-2.8.0.dev1/demos/llm/debugging/debugging_llm_training_matrix.py +696 -0
  155. agilerl-2.8.0.dev1/demos/llm/debugging/debugging_value.py +190 -0
  156. agilerl-2.8.0.dev1/demos/llm/debugging/llm_debug_utils.py +16 -0
  157. agilerl-2.8.0.dev1/demos/llm/debugging/tiny_model.py +229 -0
  158. agilerl-2.8.0.dev1/demos/llm/demo_llm_finetuning.py +342 -0
  159. agilerl-2.8.0.dev1/demos/multi_agent/demo_multi_agent.py +253 -0
  160. agilerl-2.8.0.dev1/demos/single_agent/demo_custom_network.py +222 -0
  161. agilerl-2.8.0.dev1/demos/single_agent/demo_off_policy.py +196 -0
  162. agilerl-2.8.0.dev1/demos/single_agent/demo_off_policy_distributed.py +226 -0
  163. agilerl-2.8.0.dev1/demos/single_agent/demo_offline.py +147 -0
  164. agilerl-2.8.0.dev1/demos/single_agent/demo_offline_distributed.py +188 -0
  165. agilerl-2.8.0.dev1/demos/single_agent/demo_on_policy.py +155 -0
  166. agilerl-2.8.0.dev1/demos/single_agent/demo_on_policy_rnn_cartpole.py +276 -0
  167. agilerl-2.8.0.dev1/demos/single_agent/demo_on_policy_rnn_memory.py +348 -0
  168. agilerl-2.8.0.dev1/demos/single_agent/demo_on_policy_rnn_minigrid.py +315 -0
  169. agilerl-2.8.0.dev1/demos/single_agent/performance_flamegraph_cartpole.py +226 -0
  170. agilerl-2.8.0.dev1/demos/single_agent/performance_flamegraph_lunar_lander.py +230 -0
  171. agilerl-2.8.0.dev1/demos/single_agent/performance_flamegraph_lunar_lander_rnn.py +240 -0
  172. agilerl-2.8.0.dev1/demos/single_agent/performance_flamegraph_rnn_memory.py +280 -0
  173. agilerl-2.8.0.dev1/docker +10159 -0
  174. agilerl-2.8.0.dev1/docs/Makefile +20 -0
  175. agilerl-2.8.0.dev1/docs/__init__.py +1 -0
  176. agilerl-2.8.0.dev1/docs/_static/arena-github-badge.svg +1 -0
  177. agilerl-2.8.0.dev1/docs/_static/css/custom.css +0 -0
  178. agilerl-2.8.0.dev1/docs/_static/favicon.ico +0 -0
  179. agilerl-2.8.0.dev1/docs/_static/js/expand_sidebar.js +6 -0
  180. agilerl-2.8.0.dev1/docs/_static/logo_teal.png +0 -0
  181. agilerl-2.8.0.dev1/docs/_static/logo_white.png +0 -0
  182. agilerl-2.8.0.dev1/docs/_static/module.png +0 -0
  183. agilerl-2.8.0.dev1/docs/_static/multi_turn_llm_benchmarks.png +0 -0
  184. agilerl-2.8.0.dev1/docs/_static/network.png +0 -0
  185. agilerl-2.8.0.dev1/docs/_static/thumbnails/iris-thumbnail.png +0 -0
  186. agilerl-2.8.0.dev1/docs/_static/thumbnails/pendigits-thumbnail.png +0 -0
  187. agilerl-2.8.0.dev1/docs/_static/thumbnails/rainbow_performance.png +0 -0
  188. agilerl-2.8.0.dev1/docs/_static/thumbnails/simba_thumbnail.png +0 -0
  189. agilerl-2.8.0.dev1/docs/api/algorithms/base.rst +292 -0
  190. agilerl-2.8.0.dev1/docs/api/algorithms/cispo.rst +137 -0
  191. agilerl-2.8.0.dev1/docs/api/algorithms/cql.rst +174 -0
  192. agilerl-2.8.0.dev1/docs/api/algorithms/ddpg.rst +196 -0
  193. agilerl-2.8.0.dev1/docs/api/algorithms/dpo.rst +121 -0
  194. agilerl-2.8.0.dev1/docs/api/algorithms/dqn.rst +166 -0
  195. agilerl-2.8.0.dev1/docs/api/algorithms/dqn_rainbow.rst +175 -0
  196. agilerl-2.8.0.dev1/docs/api/algorithms/grpo.rst +103 -0
  197. agilerl-2.8.0.dev1/docs/api/algorithms/gspo.rst +134 -0
  198. agilerl-2.8.0.dev1/docs/api/algorithms/ilql.rst +15 -0
  199. agilerl-2.8.0.dev1/docs/api/algorithms/index.rst +187 -0
  200. agilerl-2.8.0.dev1/docs/api/algorithms/ippo.rst +344 -0
  201. agilerl-2.8.0.dev1/docs/api/algorithms/llmppo.rst +161 -0
  202. agilerl-2.8.0.dev1/docs/api/algorithms/llmreinforce.rst +182 -0
  203. agilerl-2.8.0.dev1/docs/api/algorithms/maddpg.rst +274 -0
  204. agilerl-2.8.0.dev1/docs/api/algorithms/matd3.rst +273 -0
  205. agilerl-2.8.0.dev1/docs/api/algorithms/neural_ts.rst +154 -0
  206. agilerl-2.8.0.dev1/docs/api/algorithms/neural_ucb.rst +155 -0
  207. agilerl-2.8.0.dev1/docs/api/algorithms/ppo.rst +222 -0
  208. agilerl-2.8.0.dev1/docs/api/algorithms/registry.rst +19 -0
  209. agilerl-2.8.0.dev1/docs/api/algorithms/sft.rst +131 -0
  210. agilerl-2.8.0.dev1/docs/api/algorithms/td3.rst +185 -0
  211. agilerl-2.8.0.dev1/docs/api/algorithms/wrappers.rst +10 -0
  212. agilerl-2.8.0.dev1/docs/api/components/data.rst +46 -0
  213. agilerl-2.8.0.dev1/docs/api/components/index.rst +12 -0
  214. agilerl-2.8.0.dev1/docs/api/components/multi_agent_replay_buffer.rst +27 -0
  215. agilerl-2.8.0.dev1/docs/api/components/replay_buffer.rst +35 -0
  216. agilerl-2.8.0.dev1/docs/api/components/rollout_buffer.rst +61 -0
  217. agilerl-2.8.0.dev1/docs/api/components/sampler.rst +37 -0
  218. agilerl-2.8.0.dev1/docs/api/components/segment_tree.rst +37 -0
  219. agilerl-2.8.0.dev1/docs/api/hpo/index.rst +14 -0
  220. agilerl-2.8.0.dev1/docs/api/hpo/mutation.rst +45 -0
  221. agilerl-2.8.0.dev1/docs/api/hpo/tournament.rst +28 -0
  222. agilerl-2.8.0.dev1/docs/api/modules/base.rst +33 -0
  223. agilerl-2.8.0.dev1/docs/api/modules/bert.rst +17 -0
  224. agilerl-2.8.0.dev1/docs/api/modules/cnn.rst +10 -0
  225. agilerl-2.8.0.dev1/docs/api/modules/custom_activation.rst +7 -0
  226. agilerl-2.8.0.dev1/docs/api/modules/dummy.rst +14 -0
  227. agilerl-2.8.0.dev1/docs/api/modules/gpt.rst +26 -0
  228. agilerl-2.8.0.dev1/docs/api/modules/index.rst +26 -0
  229. agilerl-2.8.0.dev1/docs/api/modules/lstm.rst +10 -0
  230. agilerl-2.8.0.dev1/docs/api/modules/mlp.rst +10 -0
  231. agilerl-2.8.0.dev1/docs/api/modules/multi_input.rst +10 -0
  232. agilerl-2.8.0.dev1/docs/api/modules/resnet.rst +12 -0
  233. agilerl-2.8.0.dev1/docs/api/modules/simba.rst +12 -0
  234. agilerl-2.8.0.dev1/docs/api/networks/actors.rst +28 -0
  235. agilerl-2.8.0.dev1/docs/api/networks/base.rst +10 -0
  236. agilerl-2.8.0.dev1/docs/api/networks/index.rst +18 -0
  237. agilerl-2.8.0.dev1/docs/api/networks/q_networks.rst +28 -0
  238. agilerl-2.8.0.dev1/docs/api/networks/value_networks.rst +10 -0
  239. agilerl-2.8.0.dev1/docs/api/rollouts/index.rst +10 -0
  240. agilerl-2.8.0.dev1/docs/api/rollouts/on_policy.rst +42 -0
  241. agilerl-2.8.0.dev1/docs/api/train.rst +25 -0
  242. agilerl-2.8.0.dev1/docs/api/utils/algo_utils.rst +105 -0
  243. agilerl-2.8.0.dev1/docs/api/utils/cache.rst +7 -0
  244. agilerl-2.8.0.dev1/docs/api/utils/evolvable_networks.rst +65 -0
  245. agilerl-2.8.0.dev1/docs/api/utils/ilql_utils.rst +12 -0
  246. agilerl-2.8.0.dev1/docs/api/utils/index.rst +16 -0
  247. agilerl-2.8.0.dev1/docs/api/utils/llm_utils.rst +12 -0
  248. agilerl-2.8.0.dev1/docs/api/utils/log_utils.rst +9 -0
  249. agilerl-2.8.0.dev1/docs/api/utils/minari_utils.rst +8 -0
  250. agilerl-2.8.0.dev1/docs/api/utils/probe_envs.rst +386 -0
  251. agilerl-2.8.0.dev1/docs/api/utils/torch_utils.rst +12 -0
  252. agilerl-2.8.0.dev1/docs/api/utils/utils.rst +34 -0
  253. agilerl-2.8.0.dev1/docs/api/vector/index.rst +8 -0
  254. agilerl-2.8.0.dev1/docs/api/vector/petting_zoo_async_vector_env.rst +62 -0
  255. agilerl-2.8.0.dev1/docs/api/vector/petting_zoo_vector_env.rst +10 -0
  256. agilerl-2.8.0.dev1/docs/api/wrappers/agent.rst +28 -0
  257. agilerl-2.8.0.dev1/docs/api/wrappers/index.rst +11 -0
  258. agilerl-2.8.0.dev1/docs/api/wrappers/learning.rst +17 -0
  259. agilerl-2.8.0.dev1/docs/api/wrappers/llm_envs.rst +12 -0
  260. agilerl-2.8.0.dev1/docs/api/wrappers/make_evolvable.rst +8 -0
  261. agilerl-2.8.0.dev1/docs/api/wrappers/pettingzoo.rst +8 -0
  262. agilerl-2.8.0.dev1/docs/bandits/index.rst +336 -0
  263. agilerl-2.8.0.dev1/docs/conf.py +80 -0
  264. agilerl-2.8.0.dev1/docs/custom_algorithms/index.rst +75 -0
  265. agilerl-2.8.0.dev1/docs/debugging_rl/index.rst +368 -0
  266. agilerl-2.8.0.dev1/docs/distributed_training/index.rst +253 -0
  267. agilerl-2.8.0.dev1/docs/evo_hyperparam_opt/index.rst +242 -0
  268. agilerl-2.8.0.dev1/docs/evolvable_networks/index.rst +593 -0
  269. agilerl-2.8.0.dev1/docs/get_started/agilerl2changes.rst +66 -0
  270. agilerl-2.8.0.dev1/docs/get_started/index.rst +445 -0
  271. agilerl-2.8.0.dev1/docs/index.rst +162 -0
  272. agilerl-2.8.0.dev1/docs/llm_finetuning/fused_logprobs.rst +172 -0
  273. agilerl-2.8.0.dev1/docs/llm_finetuning/index.rst +81 -0
  274. agilerl-2.8.0.dev1/docs/llm_finetuning/llm_checkpoints.rst +146 -0
  275. agilerl-2.8.0.dev1/docs/llm_finetuning/quantization.rst +346 -0
  276. agilerl-2.8.0.dev1/docs/llm_finetuning/vllm_sleep_handoff.md +74 -0
  277. agilerl-2.8.0.dev1/docs/make.bat +35 -0
  278. agilerl-2.8.0.dev1/docs/multi_agent_training/index.rst +700 -0
  279. agilerl-2.8.0.dev1/docs/off_policy/index.rst +340 -0
  280. agilerl-2.8.0.dev1/docs/offline_training/index.rst +300 -0
  281. agilerl-2.8.0.dev1/docs/on_policy/index.rst +503 -0
  282. agilerl-2.8.0.dev1/docs/pomdp/index.rst +35 -0
  283. agilerl-2.8.0.dev1/docs/releases/index.rst +7 -0
  284. agilerl-2.8.0.dev1/docs/requirements.txt +6 -0
  285. agilerl-2.8.0.dev1/find_dqn_commit.sh +82 -0
  286. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/pyproject.toml +8 -36
  287. agilerl-2.8.0.dev1/saved_checkpoints/debug_matrix/multiturn:GRPO:pop1:no_tournament/attributes.pt +0 -0
  288. agilerl-2.8.0.dev1/saved_checkpoints/debug_matrix/multiturn:LLMPPO:pop1:no_tournament/attributes.pt +0 -0
  289. agilerl-2.8.0.dev1/saved_checkpoints/debug_matrix/multiturn:LLMReinforce:pop1:no_tournament/attributes.pt +0 -0
  290. agilerl-2.8.0.dev1/saved_checkpoints/debug_matrix/preference:DPO:pop1:no_tournament/attributes.pt +0 -0
  291. agilerl-2.8.0.dev1/saved_checkpoints/debug_matrix/reasoning:GRPO:pop1:no_tournament/attributes.pt +0 -0
  292. agilerl-2.8.0.dev1/saved_checkpoints/debug_matrix/reasoning:LLMPPO:pop1:no_tournament/attributes.pt +0 -0
  293. agilerl-2.8.0.dev1/saved_checkpoints/debug_matrix/reasoning:LLMReinforce:pop1:no_tournament/attributes.pt +0 -0
  294. agilerl-2.8.0.dev1/sitecustomize.py +12 -0
  295. agilerl-2.8.0.dev1/tests/__init__.py +5 -0
  296. agilerl-2.8.0.dev1/tests/assets/__init__.py +0 -0
  297. agilerl-2.8.0.dev1/tests/assets/build_minari_fixture.py +67 -0
  298. agilerl-2.8.0.dev1/tests/assets/build_tiny_llm_fixture.py +129 -0
  299. agilerl-2.8.0.dev1/tests/assets/minari_cache/D4RL/door/human-v2/data/main_data.hdf5 +0 -0
  300. agilerl-2.8.0.dev1/tests/assets/minari_cache/D4RL/door/human-v2/data/metadata.json +1 -0
  301. agilerl-2.8.0.dev1/tests/assets/minari_cache/D4RL/door/namespace_metadata.json +1 -0
  302. agilerl-2.8.0.dev1/tests/assets/minari_cache/D4RL/namespace_metadata.json +1 -0
  303. agilerl-2.8.0.dev1/tests/assets/tiny_llm/added_tokens.json +24 -0
  304. agilerl-2.8.0.dev1/tests/assets/tiny_llm/chat_template.jinja +54 -0
  305. agilerl-2.8.0.dev1/tests/assets/tiny_llm/config.json +32 -0
  306. agilerl-2.8.0.dev1/tests/assets/tiny_llm/generation_config.json +6 -0
  307. agilerl-2.8.0.dev1/tests/assets/tiny_llm/model.safetensors +0 -0
  308. agilerl-2.8.0.dev1/tests/assets/tiny_llm/special_tokens_map.json +31 -0
  309. agilerl-2.8.0.dev1/tests/assets/tiny_llm/tokenizer.json +1 -0
  310. agilerl-2.8.0.dev1/tests/assets/tiny_llm/tokenizer_config.json +207 -0
  311. agilerl-2.8.0.dev1/tests/conftest.py +539 -0
  312. agilerl-2.8.0.dev1/tests/helper_functions.py +356 -0
  313. agilerl-2.8.0.dev1/tests/helpers/__init__.py +0 -0
  314. agilerl-2.8.0.dev1/tests/helpers/algorithm_coverage.py +31 -0
  315. agilerl-2.8.0.dev1/tests/pz_vector_test_utils.py +488 -0
  316. agilerl-2.8.0.dev1/tests/subprocess_runner.py +240 -0
  317. agilerl-2.8.0.dev1/tests/test_algorithms/__init__.py +0 -0
  318. agilerl-2.8.0.dev1/tests/test_algorithms/conftest.py +39 -0
  319. agilerl-2.8.0.dev1/tests/test_algorithms/test_bandits/__init__.py +0 -0
  320. agilerl-2.8.0.dev1/tests/test_algorithms/test_bandits/test_neural_ts.py +520 -0
  321. agilerl-2.8.0.dev1/tests/test_algorithms/test_bandits/test_neural_ucb.py +531 -0
  322. agilerl-2.8.0.dev1/tests/test_algorithms/test_base.py +1682 -0
  323. agilerl-2.8.0.dev1/tests/test_algorithms/test_bc_lm.py +1235 -0
  324. agilerl-2.8.0.dev1/tests/test_algorithms/test_core_base.py +6423 -0
  325. agilerl-2.8.0.dev1/tests/test_algorithms/test_llm_ops/__init__.py +0 -0
  326. agilerl-2.8.0.dev1/tests/test_algorithms/test_llm_ops/test_fused_logprobs.py +104 -0
  327. agilerl-2.8.0.dev1/tests/test_algorithms/test_llm_ops/test_fused_lora.py +403 -0
  328. agilerl-2.8.0.dev1/tests/test_algorithms/test_llm_ops/test_fused_loss.py +1291 -0
  329. agilerl-2.8.0.dev1/tests/test_algorithms/test_llms/__init__.py +0 -0
  330. agilerl-2.8.0.dev1/tests/test_algorithms/test_llms/conftest.py +115 -0
  331. agilerl-2.8.0.dev1/tests/test_algorithms/test_llms/test_dpo.py +950 -0
  332. agilerl-2.8.0.dev1/tests/test_algorithms/test_llms/test_grpo.py +6151 -0
  333. agilerl-2.8.0.dev1/tests/test_algorithms/test_llms/test_llm_checkpoint.py +0 -0
  334. agilerl-2.8.0.dev1/tests/test_algorithms/test_llms/test_ppo_llm.py +1875 -0
  335. agilerl-2.8.0.dev1/tests/test_algorithms/test_llms/test_quantization.py +969 -0
  336. agilerl-2.8.0.dev1/tests/test_algorithms/test_llms/test_reinforce_llm.py +1469 -0
  337. agilerl-2.8.0.dev1/tests/test_algorithms/test_llms/test_sft.py +935 -0
  338. agilerl-2.8.0.dev1/tests/test_algorithms/test_llms/test_vllm.py +332 -0
  339. agilerl-2.8.0.dev1/tests/test_algorithms/test_llms/test_vllm_colocate.py +314 -0
  340. agilerl-2.8.0.dev1/tests/test_algorithms/test_multi_agent/__init__.py +0 -0
  341. agilerl-2.8.0.dev1/tests/test_algorithms/test_multi_agent/conftest.py +42 -0
  342. agilerl-2.8.0.dev1/tests/test_algorithms/test_multi_agent/test_ippo.py +2199 -0
  343. agilerl-2.8.0.dev1/tests/test_algorithms/test_multi_agent/test_maddpg.py +1938 -0
  344. agilerl-2.8.0.dev1/tests/test_algorithms/test_multi_agent/test_matd3.py +2138 -0
  345. agilerl-2.8.0.dev1/tests/test_algorithms/test_optimizer_wrapper.py +1543 -0
  346. agilerl-2.8.0.dev1/tests/test_algorithms/test_param_docstrings.py +102 -0
  347. agilerl-2.8.0.dev1/tests/test_algorithms/test_registry.py +1151 -0
  348. agilerl-2.8.0.dev1/tests/test_algorithms/test_single_agent/__init__.py +0 -0
  349. agilerl-2.8.0.dev1/tests/test_algorithms/test_single_agent/test_cqn.py +590 -0
  350. agilerl-2.8.0.dev1/tests/test_algorithms/test_single_agent/test_ddpg.py +872 -0
  351. agilerl-2.8.0.dev1/tests/test_algorithms/test_single_agent/test_dqn.py +649 -0
  352. agilerl-2.8.0.dev1/tests/test_algorithms/test_single_agent/test_dqn_rainbow.py +857 -0
  353. agilerl-2.8.0.dev1/tests/test_algorithms/test_single_agent/test_ilql.py +1335 -0
  354. agilerl-2.8.0.dev1/tests/test_algorithms/test_single_agent/test_ppo.py +2245 -0
  355. agilerl-2.8.0.dev1/tests/test_algorithms/test_single_agent/test_td3.py +1119 -0
  356. agilerl-2.8.0.dev1/tests/test_components/__init__.py +0 -0
  357. agilerl-2.8.0.dev1/tests/test_components/test_multi_agent_replay_buffer.py +638 -0
  358. agilerl-2.8.0.dev1/tests/test_components/test_replay_buffer.py +1056 -0
  359. agilerl-2.8.0.dev1/tests/test_components/test_replay_data.py +170 -0
  360. agilerl-2.8.0.dev1/tests/test_components/test_rollout_buffer.py +2129 -0
  361. agilerl-2.8.0.dev1/tests/test_components/test_sampler.py +485 -0
  362. agilerl-2.8.0.dev1/tests/test_components/test_segment_tree.py +129 -0
  363. agilerl-2.8.0.dev1/tests/test_data.py +456 -0
  364. agilerl-2.8.0.dev1/tests/test_hpo/__init__.py +0 -0
  365. agilerl-2.8.0.dev1/tests/test_hpo/test_mutation.py +2284 -0
  366. agilerl-2.8.0.dev1/tests/test_hpo/test_reinit_shared_networks.py +56 -0
  367. agilerl-2.8.0.dev1/tests/test_hpo/test_tournament.py +522 -0
  368. agilerl-2.8.0.dev1/tests/test_init.py +25 -0
  369. agilerl-2.8.0.dev1/tests/test_modules/__init__.py +0 -0
  370. agilerl-2.8.0.dev1/tests/test_modules/test_base.py +978 -0
  371. agilerl-2.8.0.dev1/tests/test_modules/test_bert.py +456 -0
  372. agilerl-2.8.0.dev1/tests/test_modules/test_cnn.py +1064 -0
  373. agilerl-2.8.0.dev1/tests/test_modules/test_cnn_cpu.py +112 -0
  374. agilerl-2.8.0.dev1/tests/test_modules/test_configs.py +92 -0
  375. agilerl-2.8.0.dev1/tests/test_modules/test_custom_activation.py +27 -0
  376. agilerl-2.8.0.dev1/tests/test_modules/test_dummy.py +86 -0
  377. agilerl-2.8.0.dev1/tests/test_modules/test_gpt.py +455 -0
  378. agilerl-2.8.0.dev1/tests/test_modules/test_lstm.py +494 -0
  379. agilerl-2.8.0.dev1/tests/test_modules/test_mlp.py +291 -0
  380. agilerl-2.8.0.dev1/tests/test_modules/test_multi_input.py +1103 -0
  381. agilerl-2.8.0.dev1/tests/test_modules/test_multi_input_cpu.py +77 -0
  382. agilerl-2.8.0.dev1/tests/test_modules/test_resnet.py +443 -0
  383. agilerl-2.8.0.dev1/tests/test_modules/test_simba.py +357 -0
  384. agilerl-2.8.0.dev1/tests/test_networks/__init__.py +0 -0
  385. agilerl-2.8.0.dev1/tests/test_networks/test_actors.py +855 -0
  386. agilerl-2.8.0.dev1/tests/test_networks/test_base.py +550 -0
  387. agilerl-2.8.0.dev1/tests/test_networks/test_distributions.py +280 -0
  388. agilerl-2.8.0.dev1/tests/test_networks/test_q_networks.py +655 -0
  389. agilerl-2.8.0.dev1/tests/test_networks/test_value_functions.py +190 -0
  390. agilerl-2.8.0.dev1/tests/test_protocols.py +586 -0
  391. agilerl-2.8.0.dev1/tests/test_rollouts/__init__.py +0 -0
  392. agilerl-2.8.0.dev1/tests/test_rollouts/test_on_policy.py +396 -0
  393. agilerl-2.8.0.dev1/tests/test_train/__init__.py +0 -0
  394. agilerl-2.8.0.dev1/tests/test_train/test_train.py +5510 -0
  395. agilerl-2.8.0.dev1/tests/test_train/test_train_llm.py +2910 -0
  396. agilerl-2.8.0.dev1/tests/test_utils/__init__.py +0 -0
  397. agilerl-2.8.0.dev1/tests/test_utils/test_algo_utils.py +2020 -0
  398. agilerl-2.8.0.dev1/tests/test_utils/test_cache.py +62 -0
  399. agilerl-2.8.0.dev1/tests/test_utils/test_ilql_utils.py +82 -0
  400. agilerl-2.8.0.dev1/tests/test_utils/test_llm_packing.py +375 -0
  401. agilerl-2.8.0.dev1/tests/test_utils/test_llm_utils.py +2794 -0
  402. agilerl-2.8.0.dev1/tests/test_utils/test_log_utils.py +230 -0
  403. agilerl-2.8.0.dev1/tests/test_utils/test_minari_utils.py +317 -0
  404. agilerl-2.8.0.dev1/tests/test_utils/test_ppo_value_head.py +481 -0
  405. agilerl-2.8.0.dev1/tests/test_utils/test_probe_envs.py +814 -0
  406. agilerl-2.8.0.dev1/tests/test_utils/test_probe_envs_llm.py +268 -0
  407. agilerl-2.8.0.dev1/tests/test_utils/test_probe_envs_ma.py +855 -0
  408. agilerl-2.8.0.dev1/tests/test_utils/test_sampling_utils.py +175 -0
  409. agilerl-2.8.0.dev1/tests/test_utils/test_torch_utils.py +320 -0
  410. agilerl-2.8.0.dev1/tests/test_utils/test_utils.py +1716 -0
  411. agilerl-2.8.0.dev1/tests/test_utils/test_utils_evolvable.py +804 -0
  412. agilerl-2.8.0.dev1/tests/test_vector/__init__.py +0 -0
  413. agilerl-2.8.0.dev1/tests/test_vector/test_vector.py +1821 -0
  414. agilerl-2.8.0.dev1/tests/test_wrappers/__init__.py +0 -0
  415. agilerl-2.8.0.dev1/tests/test_wrappers/test_agent.py +2045 -0
  416. agilerl-2.8.0.dev1/tests/test_wrappers/test_autoreset.py +282 -0
  417. agilerl-2.8.0.dev1/tests/test_wrappers/test_bandit_env.py +52 -0
  418. agilerl-2.8.0.dev1/tests/test_wrappers/test_llm_envs.py +1192 -0
  419. agilerl-2.8.0.dev1/tests/test_wrappers/test_make_evolvable.py +1350 -0
  420. agilerl-2.8.0.dev1/tests/test_wrappers/test_multiturn_wrappers.py +1284 -0
  421. agilerl-2.8.0.dev1/tests/test_wrappers/test_skills.py +62 -0
  422. agilerl-2.8.0.dev1/tests/utils.py +373 -0
  423. agilerl-2.8.0.dev1/uv.lock +7411 -0
  424. agilerl-2.8.0.dev0/agilerl/__init__.py +0 -77
  425. agilerl-2.8.0.dev0/agilerl/logger.py +0 -253
  426. agilerl-2.8.0.dev0/agilerl/metrics.py +0 -374
  427. agilerl-2.8.0.dev0/agilerl/models/__init__.py +0 -82
  428. agilerl-2.8.0.dev0/agilerl/models/algo.py +0 -596
  429. agilerl-2.8.0.dev0/agilerl/models/algorithms/__init__.py +0 -54
  430. agilerl-2.8.0.dev0/agilerl/models/algorithms/cispo.py +0 -35
  431. agilerl-2.8.0.dev0/agilerl/models/algorithms/cqn.py +0 -41
  432. agilerl-2.8.0.dev0/agilerl/models/algorithms/ddpg.py +0 -48
  433. agilerl-2.8.0.dev0/agilerl/models/algorithms/dpo.py +0 -33
  434. agilerl-2.8.0.dev0/agilerl/models/algorithms/dqn.py +0 -40
  435. agilerl-2.8.0.dev0/agilerl/models/algorithms/grpo.py +0 -66
  436. agilerl-2.8.0.dev0/agilerl/models/algorithms/gspo.py +0 -30
  437. agilerl-2.8.0.dev0/agilerl/models/algorithms/ippo.py +0 -50
  438. agilerl-2.8.0.dev0/agilerl/models/algorithms/llmppo.py +0 -68
  439. agilerl-2.8.0.dev0/agilerl/models/algorithms/llmreinforce.py +0 -63
  440. agilerl-2.8.0.dev0/agilerl/models/algorithms/maddpg.py +0 -49
  441. agilerl-2.8.0.dev0/agilerl/models/algorithms/matd3.py +0 -50
  442. agilerl-2.8.0.dev0/agilerl/models/algorithms/neural_ts.py +0 -43
  443. agilerl-2.8.0.dev0/agilerl/models/algorithms/neural_ucb.py +0 -43
  444. agilerl-2.8.0.dev0/agilerl/models/algorithms/ppo.py +0 -59
  445. agilerl-2.8.0.dev0/agilerl/models/algorithms/rainbow_dqn.py +0 -54
  446. agilerl-2.8.0.dev0/agilerl/models/algorithms/sft.py +0 -33
  447. agilerl-2.8.0.dev0/agilerl/models/algorithms/td3.py +0 -48
  448. agilerl-2.8.0.dev0/agilerl/models/env.py +0 -851
  449. agilerl-2.8.0.dev0/agilerl/models/hpo.py +0 -81
  450. agilerl-2.8.0.dev0/agilerl/models/manifest.py +0 -370
  451. agilerl-2.8.0.dev0/agilerl/models/networks.py +0 -456
  452. agilerl-2.8.0.dev0/agilerl/models/training.py +0 -255
  453. agilerl-2.8.0.dev0/agilerl/population.py +0 -738
  454. agilerl-2.8.0.dev0/agilerl/train.py +0 -159
  455. agilerl-2.8.0.dev0/agilerl/training/train_llm.py +0 -1205
  456. agilerl-2.8.0.dev0/agilerl/training/train_multi_agent_off_policy.py +0 -375
  457. agilerl-2.8.0.dev0/agilerl/training/train_multi_agent_on_policy.py +0 -394
  458. agilerl-2.8.0.dev0/agilerl/training/train_off_policy.py +0 -448
  459. agilerl-2.8.0.dev0/agilerl/training/train_on_policy.py +0 -292
  460. agilerl-2.8.0.dev0/agilerl/training/trainer.py +0 -731
  461. agilerl-2.8.0.dev0/agilerl/utils/env_utils.py +0 -260
  462. agilerl-2.8.0.dev0/agilerl/utils/population_utils.py +0 -212
  463. agilerl-2.8.0.dev0/agilerl/utils/trainer_utils.py +0 -324
  464. agilerl-2.8.0.dev0/agilerl/vector/__init__.py +0 -10
  465. agilerl-2.8.0.dev0/agilerl/vector/dummy_vec_env.py +0 -339
  466. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/LICENSE +0 -0
  467. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/__init__.py +0 -0
  468. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/bc_lm.py +0 -0
  469. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/core/__init__.py +0 -0
  470. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/core/llm_ops/__init__.py +0 -0
  471. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/core/llm_ops/fused_logprobs.py +0 -0
  472. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/core/llm_ops/fused_lora.py +0 -0
  473. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/core/llm_ops/fused_loss.py +0 -0
  474. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/core/llm_ops/vllm_colocate.py +0 -0
  475. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/algorithms/core/optimizer_wrapper.py +0 -0
  476. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/data/__init__.py +0 -0
  477. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/data/language_environment.py +0 -0
  478. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/data/rl_data.py +0 -0
  479. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/data/tokenizer.py +0 -0
  480. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/data/torch_datasets.py +0 -0
  481. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/hpo/__init__.py +0 -0
  482. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/llm_envs/__init__.py +0 -0
  483. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/llm_envs/base.py +0 -0
  484. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/llm_envs/search.py +0 -0
  485. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/llm_envs/sync_vec_env.py +0 -0
  486. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/llm_envs/token_observation.py +0 -0
  487. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/__init__.py +0 -0
  488. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/base.py +0 -0
  489. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/bert.py +0 -0
  490. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/configs.py +0 -0
  491. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/custom_components.py +0 -0
  492. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/gpt.py +0 -0
  493. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/lstm.py +0 -0
  494. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/mlp.py +0 -0
  495. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/multi_input.py +0 -0
  496. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/resnet.py +0 -0
  497. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/modules/simba.py +0 -0
  498. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/networks/__init__.py +0 -0
  499. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/networks/custom_modules.py +0 -0
  500. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/networks/distributions.py +0 -0
  501. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/rollouts/__init__.py +0 -0
  502. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/training/__init__.py +0 -0
  503. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/__init__.py +0 -0
  504. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/cache.py +0 -0
  505. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/evolvable_networks.py +0 -0
  506. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/ilql_utils.py +0 -0
  507. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/llm_packing.py +0 -0
  508. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/log_utils.py +0 -0
  509. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/ppo_value_head.py +0 -0
  510. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/probe_envs_llm.py +0 -0
  511. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/sampling_utils.py +0 -0
  512. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/utils/torch_utils.py +0 -0
  513. {agilerl-2.8.0.dev0/agilerl/wrappers → agilerl-2.8.0.dev1/agilerl/vector}/__init__.py +0 -0
  514. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/vector/pz_async_vec_env.py +0 -0
  515. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/vector/pz_vec_env.py +0 -0
  516. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/wrappers/llm_envs.py +0 -0
  517. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/wrappers/make_evolvable.py +0 -0
  518. {agilerl-2.8.0.dev0 → agilerl-2.8.0.dev1}/agilerl/wrappers/utils.py +0 -0
@@ -0,0 +1,3 @@
1
+ # Large test fixtures: treat as generated/binary so they don't bloat diffs,
2
+ # blame, or PR reviews. This is a minified HuggingFace tokenizer used by tests.
3
+ tests/assets/tiny_llm/tokenizer.json linguist-generated=true -diff
@@ -0,0 +1,27 @@
1
+ ---
2
+ name: Bug report
3
+ about: Create a report to help us improve
4
+ title: ''
5
+ labels: bug
6
+ assignees: ''
7
+
8
+ ---
9
+
10
+ **What version of AgileRL are you using?**
11
+ **What operating system and processor architecture are you using?**
12
+
13
+ **What did you do?**
14
+ Steps to reproduce the behaviour:
15
+ 1. Go to '...'
16
+ 2. Click on '....'
17
+ 3. Scroll down to '....'
18
+ 4. See error
19
+
20
+ **What did you expect to see?**
21
+ A clear and concise description of what you expected to happen.
22
+
23
+ **What did you see instead? Describe the bug.**
24
+ A clear and concise description of what the bug is.
25
+
26
+ **Additional context**
27
+ Add any other context about the problem here.
@@ -0,0 +1,20 @@
1
+ ---
2
+ name: Feature request
3
+ about: Suggest an idea for this project
4
+ title: ''
5
+ labels: ''
6
+ assignees: ''
7
+
8
+ ---
9
+
10
+ **Is your feature request related to a problem? Please describe.**
11
+ A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]
12
+
13
+ **Describe the solution you'd like**
14
+ A clear and concise description of what you want to happen.
15
+
16
+ **Describe alternatives you've considered**
17
+ A clear and concise description of any alternative solutions or features you've considered.
18
+
19
+ **Additional context**
20
+ Add any other context or screenshots about the feature request here.
@@ -0,0 +1,34 @@
1
+ # Pull Request
2
+
3
+ ## Description
4
+
5
+ Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context.
6
+
7
+ Fixes # (issue)
8
+
9
+ ## Type of change
10
+
11
+ Please delete options that are not relevant.
12
+
13
+ - [ ] Bug fix (non-breaking change which fixes an issue)
14
+ - [ ] New feature (non-breaking change which adds functionality)
15
+ - [ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)
16
+ - [ ] Documentation update
17
+
18
+ ## How Has This Been Tested?
19
+
20
+ Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce.
21
+
22
+ - [ ] Test A
23
+ - [ ] Test B
24
+
25
+ ## Checklist:
26
+
27
+ - [ ] My code follows the style guidelines of this project
28
+ - [ ] I have performed a self-review of my own code
29
+ - [ ] I have commented my code, particularly in hard-to-understand areas
30
+ - [ ] I have made corresponding changes to the documentation
31
+ - [ ] My changes generate no new warnings
32
+ - [ ] I have added tests that prove my fix is effective or that my feature works
33
+ - [ ] New and existing unit tests pass locally with my changes
34
+ - [ ] Any dependent changes have been merged and published in downstream modules
@@ -0,0 +1,24 @@
1
+ <svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" width="65" height="20" role="img"
2
+ aria-label="Arena">
3
+ <title>Arena</title>
4
+ <linearGradient id="s" x2="0" y2="100%">
5
+ <stop offset="0" stop-color="#bbb" stop-opacity=".1" />
6
+ <stop offset="1" stop-opacity=".1" />
7
+ </linearGradient>
8
+ <clipPath id="r">
9
+ <rect width="65" height="20" rx="3" fill="#fff" />
10
+ </clipPath>
11
+ <g clip-path="url(#r)">
12
+ <rect width="24" height="20" fill="#555" />
13
+ <rect x="24" width="41" height="20" fill="#467f81" />
14
+ <rect width="65" height="20" fill="url(#s)" />
15
+ </g>
16
+ <g fill="#fff" text-anchor="middle" font-family="Verdana,Geneva,DejaVu Sans,sans-serif"
17
+ text-rendering="geometricPrecision" font-size="110">
18
+ <image x="5" y="3" width="14" height="14"
19
+ xlink:href="data:image/jpeg;base64,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" />
20
+ <text aria-hidden="true" x="435" y="150" fill="#010101" fill-opacity=".3" transform="scale(.1)"
21
+ textLength="310">Arena</text><text x="435" y="140" transform="scale(.1)" fill="#fff"
22
+ textLength="310">Arena</text>
23
+ </g>
24
+ </svg>
@@ -0,0 +1,51 @@
1
+ #!/usr/bin/env bash
2
+ set -euo pipefail
3
+
4
+ # Install the CodeQL CLI locally for developer workflows.
5
+ # Environment variables:
6
+ # - CODEQL_VERSION (default: 2.24.2)
7
+ # - CODEQL_INSTALL_DIR (default: /tmp/codeql)
8
+
9
+ CODEQL_VERSION="${CODEQL_VERSION:-2.24.2}"
10
+ CODEQL_INSTALL_DIR="${CODEQL_INSTALL_DIR:-/tmp/codeql}"
11
+
12
+ if [[ -x "${CODEQL_INSTALL_DIR}/codeql" ]]; then
13
+ echo "CodeQL already installed at ${CODEQL_INSTALL_DIR}/codeql"
14
+ exit 0
15
+ fi
16
+
17
+ OS="$(uname -s)"
18
+ case "${OS}" in
19
+ Linux)
20
+ ASSET="codeql-linux64.zip"
21
+ ;;
22
+ Darwin)
23
+ # CodeQL distributes an osx64 bundle that works on current macOS setups.
24
+ ASSET="codeql-osx64.zip"
25
+ ;;
26
+ MINGW*|MSYS*|CYGWIN*)
27
+ ASSET="codeql-win64.zip"
28
+ ;;
29
+ *)
30
+ echo "Unsupported OS: ${OS}" >&2
31
+ exit 1
32
+ ;;
33
+ esac
34
+
35
+ URL="https://github.com/github/codeql-cli-binaries/releases/download/v${CODEQL_VERSION}/${ASSET}"
36
+ TMP_DIR="$(mktemp -d)"
37
+ ARCHIVE_PATH="${TMP_DIR}/${ASSET}"
38
+ EXTRACT_DIR="${TMP_DIR}/extract"
39
+
40
+ echo "Downloading CodeQL ${CODEQL_VERSION} from ${URL}"
41
+ curl -fsSL "${URL}" -o "${ARCHIVE_PATH}"
42
+
43
+ mkdir -p "${EXTRACT_DIR}"
44
+ unzip -q "${ARCHIVE_PATH}" -d "${EXTRACT_DIR}"
45
+
46
+ mkdir -p "${CODEQL_INSTALL_DIR}"
47
+ rm -rf "${CODEQL_INSTALL_DIR:?}"/*
48
+ cp -R "${EXTRACT_DIR}/codeql/." "${CODEQL_INSTALL_DIR}"
49
+
50
+ echo "Installed CodeQL to ${CODEQL_INSTALL_DIR}"
51
+ "${CODEQL_INSTALL_DIR}/codeql" version
@@ -0,0 +1,139 @@
1
+ # noqa: INP001
2
+ """Run local CodeQL and fail on findings.
3
+
4
+ This script is designed for local developer use (including pre-commit).
5
+ """
6
+
7
+ from __future__ import annotations
8
+
9
+ import csv
10
+ import os
11
+ import shutil
12
+ import subprocess
13
+ import sys
14
+ import tempfile
15
+ from pathlib import Path
16
+
17
+ REPO_ROOT = Path(__file__).resolve().parents[2]
18
+ DEFAULT_DB_PATH = Path(tempfile.gettempdir()) / "agilerl-codeql-db"
19
+ DEFAULT_CONFIG_PATH = REPO_ROOT / ".github" / "codeql" / "codeql-config.yml"
20
+ DEFAULT_CODEQL_INSTALL_DIR = Path("/tmp/codeql") # noqa: S108
21
+ INSTALL_SCRIPT_PATH = Path(__file__).with_name("install_codeql.sh")
22
+ CSV_FILE_COL_IDX = 4
23
+ PYTHON_QUERY_SUITE = (
24
+ "codeql/python-queries:codeql-suites/python-security-and-quality.qls"
25
+ )
26
+
27
+
28
+ def resolve_codeql_bin() -> str | None:
29
+ """Resolve CodeQL binary path from env/PATH/common local install."""
30
+ env_bin = os.environ.get("CODEQL_BIN")
31
+ if env_bin:
32
+ return env_bin
33
+ path_bin = shutil.which("codeql")
34
+ if path_bin:
35
+ return path_bin
36
+ install_dir = Path(
37
+ os.environ.get("CODEQL_INSTALL_DIR", str(DEFAULT_CODEQL_INSTALL_DIR))
38
+ )
39
+ candidates = (
40
+ [install_dir]
41
+ if install_dir.is_file()
42
+ else [
43
+ install_dir / "codeql",
44
+ install_dir / "codeql.exe",
45
+ ]
46
+ )
47
+ for fallback in candidates:
48
+ if fallback.exists():
49
+ return str(fallback)
50
+ return None
51
+
52
+
53
+ def ensure_codeql_installed() -> str | None:
54
+ """Install CodeQL via bootstrap script when not present."""
55
+ codeql = resolve_codeql_bin()
56
+ if codeql is not None:
57
+ return codeql
58
+ if not INSTALL_SCRIPT_PATH.exists():
59
+ return None
60
+
61
+ bash_bin = shutil.which("bash")
62
+ if bash_bin is None:
63
+ return None
64
+
65
+ subprocess.run( # noqa: S603
66
+ [bash_bin, str(INSTALL_SCRIPT_PATH)],
67
+ cwd=REPO_ROOT,
68
+ check=True,
69
+ )
70
+ return resolve_codeql_bin()
71
+
72
+
73
+ def run(cmd: list[str], cwd: Path) -> None:
74
+ """Run subprocess command and propagate failure."""
75
+ subprocess.run(cmd, cwd=cwd, check=True) # noqa: S603
76
+
77
+
78
+ def format_clickable_path(path: str) -> str:
79
+ """Normalize finding file paths for IDE click-through."""
80
+ return path.lstrip("/\\")
81
+
82
+
83
+ def main() -> int:
84
+ """Execute local CodeQL scan and return process exit code."""
85
+ codeql = ensure_codeql_installed()
86
+ if codeql is None:
87
+ print(
88
+ "CodeQL binary not found. Set CODEQL_BIN or install CodeQL "
89
+ "(e.g. /tmp/codeql/codeql).",
90
+ file=sys.stderr,
91
+ )
92
+ return 1
93
+
94
+ with tempfile.TemporaryDirectory(prefix="agilerl-codeql-") as td:
95
+ output_csv = Path(td) / "results.csv"
96
+
97
+ run(
98
+ [
99
+ codeql,
100
+ "database",
101
+ "create",
102
+ str(DEFAULT_DB_PATH),
103
+ "--overwrite",
104
+ "--language=python",
105
+ f"--source-root={REPO_ROOT}",
106
+ ],
107
+ cwd=REPO_ROOT,
108
+ )
109
+ run(
110
+ [
111
+ codeql,
112
+ "database",
113
+ "analyze",
114
+ str(DEFAULT_DB_PATH),
115
+ PYTHON_QUERY_SUITE,
116
+ "--download",
117
+ "--format=csv",
118
+ f"--output={output_csv}",
119
+ ],
120
+ cwd=REPO_ROOT,
121
+ )
122
+
123
+ with output_csv.open(newline="") as f:
124
+ rows = list(csv.reader(f))
125
+
126
+ actionable = [r for r in rows if len(r) > CSV_FILE_COL_IDX]
127
+ if actionable:
128
+ print(f"CodeQL found {len(actionable)} actionable issue(s):", file=sys.stderr)
129
+ for row in actionable[:20]:
130
+ file_path = format_clickable_path(row[4])
131
+ print(f"- {row[0]} at {file_path}:{row[5]}", file=sys.stderr)
132
+ return 1
133
+
134
+ print("CodeQL passed (no actionable issues).")
135
+ return 0
136
+
137
+
138
+ if __name__ == "__main__":
139
+ raise SystemExit(main())
@@ -0,0 +1,8 @@
1
+ ---
2
+ version: 2
3
+ updates:
4
+ - package-ecosystem: uv
5
+ directory: /
6
+ schedule:
7
+ interval: weekly
8
+ target-branch: nightly
@@ -0,0 +1,101 @@
1
+ ---
2
+ # For most projects, this workflow file will not need changing; you simply need
3
+ # to commit it to your repository.
4
+ #
5
+ # You may wish to alter this file to override the set of languages analyzed,
6
+ # or to provide custom queries or build logic.
7
+ #
8
+ # ******** NOTE ********
9
+ # We have attempted to detect the languages in your repository. Please check
10
+ # the `language` matrix defined below to confirm you have the correct set of
11
+ # supported CodeQL languages.
12
+ #
13
+ name: CodeQL Advanced
14
+
15
+ on:
16
+ push:
17
+ branches: [main, nightly]
18
+ pull_request:
19
+ branches: [main, nightly]
20
+ schedule:
21
+ - cron: 40 9 * * 4
22
+
23
+ jobs:
24
+ analyze:
25
+ name: Analyze (${{ matrix.language }})
26
+ # Runner size impacts CodeQL analysis time. To learn more, please see:
27
+ # - https://gh.io/recommended-hardware-resources-for-running-codeql
28
+ # - https://gh.io/supported-runners-and-hardware-resources
29
+ # - https://gh.io/using-larger-runners (GitHub.com only)
30
+ # Consider using larger runners or machines with greater resources for possible analysis time improvements.
31
+ runs-on: ${{ (matrix.language == 'swift' && 'macos-latest') || 'ubuntu-latest' }}
32
+ permissions:
33
+ # required for all workflows
34
+ security-events: write
35
+
36
+ # required to fetch internal or private CodeQL packs
37
+ packages: read
38
+
39
+ # only required for workflows in private repositories
40
+ actions: read
41
+ contents: read
42
+
43
+ strategy:
44
+ fail-fast: false
45
+ matrix:
46
+ include:
47
+ - language: actions
48
+ build-mode: none
49
+ - language: python
50
+ build-mode: none
51
+ # CodeQL supports the following values keywords for 'language': 'actions', 'c-cpp', 'csharp', 'go', 'java-kotlin', 'javascript-typescript', 'python', 'ruby', 'rust', 'swift'
52
+ # Use `c-cpp` to analyze code written in C, C++ or both
53
+ # Use 'java-kotlin' to analyze code written in Java, Kotlin or both
54
+ # Use 'javascript-typescript' to analyze code written in JavaScript, TypeScript or both
55
+ # To learn more about changing the languages that are analyzed or customizing the build mode for your analysis,
56
+ # see https://docs.github.com/en/code-security/code-scanning/creating-an-advanced-setup-for-code-scanning/customizing-your-advanced-setup-for-code-scanning.
57
+ # If you are analyzing a compiled language, you can modify the 'build-mode' for that language to customize how
58
+ # your codebase is analyzed, see https://docs.github.com/en/code-security/code-scanning/creating-an-advanced-setup-for-code-scanning/codeql-code-scanning-for-compiled-languages
59
+ steps:
60
+ - name: Checkout repository
61
+ uses: actions/checkout@v4
62
+
63
+ # Add any setup steps before running the `github/codeql-action/init` action.
64
+ # This includes steps like installing compilers or runtimes (`actions/setup-node`
65
+ # or others). This is typically only required for manual builds.
66
+ # - name: Setup runtime (example)
67
+ # uses: actions/setup-example@v1
68
+
69
+ # Initializes the CodeQL tools for scanning.
70
+ - name: Initialize CodeQL
71
+ uses: github/codeql-action/init@v3
72
+ with:
73
+ languages: ${{ matrix.language }}
74
+ build-mode: ${{ matrix.build-mode }}
75
+ # If you wish to specify custom queries, you can do so here or in a config file.
76
+ # By default, queries listed here will override any specified in a config file.
77
+ # Prefix the list here with "+" to use these queries and those in the config file.
78
+
79
+ # For more details on CodeQL's query packs, refer to: https://docs.github.com/en/code-security/code-scanning/automatically-scanning-your-code-for-vulnerabilities-and-errors/configuring-code-scanning#using-queries-in-ql-packs
80
+ # queries: security-extended,security-and-quality
81
+
82
+ # If the analyze step fails for one of the languages you are analyzing with
83
+ # "We were unable to automatically build your code", modify the matrix above
84
+ # to set the build mode to "manual" for that language. Then modify this step
85
+ # to build your code.
86
+ # ℹ️ Command-line programs to run using the OS shell.
87
+ # 📚 See https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions#jobsjob_idstepsrun
88
+ - if: matrix.build-mode == 'manual'
89
+ shell: bash
90
+ run: |
91
+ echo 'If you are using a "manual" build mode for one or more of the' \
92
+ 'languages you are analyzing, replace this with the commands to build' \
93
+ 'your code, for example:'
94
+ echo ' make bootstrap'
95
+ echo ' make release'
96
+ exit 1
97
+
98
+ - name: Perform CodeQL Analysis
99
+ uses: github/codeql-action/analyze@v3
100
+ with:
101
+ category: /language:${{matrix.language}}
@@ -0,0 +1,94 @@
1
+ ---
2
+ name: Linux
3
+
4
+ on:
5
+ push:
6
+ branches: [main, nightly]
7
+ paths:
8
+ - agilerl/**
9
+ - tests/**
10
+ - .github/workflows/**
11
+ - pyproject.toml
12
+ pull_request:
13
+ paths:
14
+ - agilerl/**
15
+ - tests/**
16
+ - .github/workflows/**
17
+ - pyproject.toml
18
+
19
+ concurrency:
20
+ group: ${{ github.workflow }}-${{ github.head_ref || github.ref }}
21
+ cancel-in-progress: true
22
+
23
+ permissions:
24
+ contents: read
25
+
26
+ jobs:
27
+ tests:
28
+ runs-on: gha-runner-scale-set
29
+ strategy:
30
+ fail-fast: false
31
+ max-parallel: 4
32
+ matrix:
33
+ python-version: ['3.10', '3.11', '3.12', '3.13']
34
+
35
+ container:
36
+ image: pytorch/pytorch:2.11.0-cuda13.0-cudnn9-devel
37
+ options: --user root
38
+
39
+ # Workspace (/__w) is ~1GB with little free space; root (/) has plenty. Put cache and venv on /.
40
+ env:
41
+ UV_CACHE_DIR: /tmp/uv-cache
42
+ UV_PROJECT_ENVIRONMENT: /tmp/agilerl-venv
43
+ HF_HOME: /tmp/hf-cache
44
+ TORCHINDUCTOR_CACHE_DIR: /tmp/inductor-cache
45
+
46
+ steps:
47
+ - uses: actions/checkout@v4
48
+ - uses: astral-sh/setup-uv@v7
49
+ with:
50
+ enable-cache: true
51
+ python-version: ${{ matrix.python-version }}
52
+
53
+ - name: Cache HuggingFace models
54
+ uses: actions/cache@v4
55
+ with:
56
+ path: /tmp/hf-cache
57
+ key: hf-${{ matrix.python-version }}-${{ hashFiles('pyproject.toml') }}
58
+ restore-keys: hf-${{ matrix.python-version }}-
59
+
60
+ - name: Cache torch inductor compilations
61
+ uses: actions/cache@v4
62
+ with:
63
+ path: /tmp/inductor-cache
64
+ key: inductor-${{ matrix.python-version }}-${{ hashFiles('pyproject.toml') }}
65
+ restore-keys: inductor-${{ matrix.python-version }}-
66
+
67
+ - name: Install dependencies
68
+ # swig is needed to build box2d-py from source (no pre-built wheels for py3.10+).
69
+ run: |
70
+ uv sync --locked --all-groups --extra all
71
+ echo "$UV_PROJECT_ENVIRONMENT/bin" >> $GITHUB_PATH
72
+
73
+ - name: Reset coverage data
74
+ run: rm -f .coverage .coverage.*
75
+
76
+ # Single phase: `-n auto --dist loadgroup` from pyproject gives 8
77
+ # workers on the gha-runner-scale-set node. `vllm`- and `gpu`-marked
78
+ # tests share the 4 `gputest0..gputest3` xdist groups defined in
79
+ # `tests/conftest.py`, which caps GPU-touching concurrency at 4
80
+ # workers regardless of -n; the remaining ~4 workers fan out across
81
+ # CPU-only tests. See the `pytest_collection_modifyitems` docstring
82
+ # for the GPU-memory and port-race rationale behind the 4-group cap.
83
+ #
84
+ # Single invocation = single coverage run; pytest-cov auto-combines
85
+ # the per-worker `.coverage.*` shards before writing `coverage.xml`,
86
+ # so we don't need a manual `coverage combine` step (which was
87
+ # tripping over corrupted shards under the old two-phase setup).
88
+ - name: Run tests
89
+ run: uv run pytest --exitfirst --cov=agilerl --cov-report=xml --durations=0 --durations-min=1.0
90
+
91
+ - name: Upload coverage reports to Codecov
92
+ uses: codecov/codecov-action@v3
93
+ env:
94
+ CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
@@ -0,0 +1,50 @@
1
+ ---
2
+ name: macOS
3
+
4
+ on:
5
+ push:
6
+ branches: [main, nightly]
7
+ paths:
8
+ - agilerl/**
9
+ - tests/**
10
+ - .github/workflows/**
11
+ - pyproject.toml
12
+ pull_request:
13
+ branches: [main, nightly]
14
+ paths:
15
+ - agilerl/**
16
+ - tests/**
17
+ - .github/workflows/**
18
+ - pyproject.toml
19
+
20
+ concurrency:
21
+ group: ${{ github.workflow }}-${{ github.head_ref || github.ref }}
22
+ cancel-in-progress: true
23
+
24
+ permissions:
25
+ contents: read
26
+
27
+ jobs:
28
+ tests:
29
+ runs-on: ${{ matrix.os }}
30
+ strategy:
31
+ fail-fast: false
32
+ max-parallel: 4
33
+ matrix:
34
+ os: [macos-26]
35
+ python-version: ['3.10', '3.11', '3.12', '3.13']
36
+
37
+ steps:
38
+ - uses: actions/checkout@v4
39
+ - uses: astral-sh/setup-uv@v7
40
+ with:
41
+ enable-cache: true
42
+ python-version: ${{ matrix.python-version }}
43
+
44
+ - name: Install dependencies
45
+ run: |
46
+ uv sync --locked --all-groups --extra all
47
+
48
+ - name: Test with pytest
49
+ run: |
50
+ uv run pytest --exitfirst --cov=agilerl --cov-report=term-missing --durations=0 --durations-min=1.0
@@ -0,0 +1,51 @@
1
+ ---
2
+ name: Windows
3
+
4
+ on:
5
+ push:
6
+ branches: [main, nightly]
7
+ paths:
8
+ - agilerl/**
9
+ - tests/**
10
+ - .github/workflows/**
11
+ - pyproject.toml
12
+ pull_request:
13
+ branches: [main, nightly]
14
+ paths:
15
+ - agilerl/**
16
+ - tests/**
17
+ - .github/workflows/**
18
+ - pyproject.toml
19
+
20
+ concurrency:
21
+ group: ${{ github.workflow }}-${{ github.head_ref || github.ref }}
22
+ cancel-in-progress: true
23
+
24
+ permissions:
25
+ contents: read
26
+
27
+ jobs:
28
+ tests:
29
+ runs-on: ${{ matrix.os }}
30
+ strategy:
31
+ fail-fast: false
32
+ max-parallel: 4
33
+ matrix:
34
+ os: [windows-2025]
35
+ python-version: ['3.10', '3.11', '3.12', '3.13']
36
+
37
+ steps:
38
+ - uses: actions/checkout@v4
39
+ - uses: astral-sh/setup-uv@v7
40
+ with:
41
+ enable-cache: true
42
+ python-version: ${{ matrix.python-version }}
43
+
44
+ - name: Install dependencies
45
+ run: |
46
+ pip install uv
47
+ uv sync --locked --all-groups --extra all
48
+
49
+ - name: Test with pytest
50
+ run: |
51
+ uv run pytest --exitfirst -m "not llm" --cov=agilerl --cov-report=term-missing --durations=0 --durations-min=1.0
@@ -1,6 +1,3 @@
1
- # Symlink for Pylance/Pyright type checking of namespace package
2
- agilerl/arena
3
-
4
1
  # Byte-compiled / optimized / DLL files
5
2
  __pycache__/
6
3
  *.py[cod]
@@ -116,10 +113,8 @@ celerybeat.pid
116
113
  .claude/
117
114
  .venv*/
118
115
  venv/
119
- .envrc
120
116
  env.bak/
121
117
  venv.bak/
122
- .cursor/
123
118
 
124
119
  # Spyder project settings
125
120
  .spyderproject
@@ -141,7 +136,6 @@ dmypy.json
141
136
 
142
137
  # Saved models
143
138
  outputs/
144
- tensorboard_logs/
145
139
 
146
140
  # Ruff
147
141
  .ruff_cache/
@@ -149,8 +143,6 @@ tensorboard_logs/
149
143
  # Large data files
150
144
  data/wordle/
151
145
  data/wordle.zip
152
- *.tar.gz
153
- *metrics.csv
154
146
 
155
147
  # WandB
156
148
  wandb/
@@ -161,6 +153,3 @@ wandb/
161
153
  # AI assistant instructions
162
154
  CLAUDE.md
163
155
  AGENTS.md
164
-
165
- # Simlink for type-checking
166
- agilerl/arena/