RLinf: Reinforcement Learning Infrastructure for Embodied and Agentic AI
This report presents the forensic synthetic code analysis of RLinf/RLinf, a Python project with 4,484 GitHub stars. SynthScan v2.0 examined 350,758 lines of code across 1471 source files, recording 4140 pattern matches distributed across 20 syntactic categories. The overall adjusted score of 16.1 places this repository in the Moderate AI signal band.
The scanner applied 160+ deterministic lexical heuristics, multi-line block detectors, abstract syntax tree depth profilers, and a cross-file Jaccard similarity matrix to construct a statistically normalised synthetic code estimate. All matches are individually weighted by severity coefficient and contextual multiplier before summation, and the resulting headline score is temporally discounted to account for the repository's development history relative to the commercial emergence of large language model coding tooling (November 2022 onward).
Longitudinal tracking requires multiple scan runs. Once this repository is re-scanned after new commits land, this chart will visualise how the synthetic code signal evolves over time — enabling you to detect whether AI authorship is growing, stabilising, or being actively corrected by human engineers.
Classifies detected patterns by their diagnostic confidence and structural impact. CRITICAL patterns (coefficient 10) represent definitive synthetic signatures — hallucinated imports, explicit LLM attribution metadata — virtually never produced by human authors. HIGH (5) indicates strong structural tells such as cross-file repetition or cross-linguistic idioms. MEDIUM (2) covers recognisable conversational padding and AI-specific vocabulary. LOW (1) captures subtle indicators like tautological comments and generic boilerplate that require density to carry independent signal.
This horizontal bar chart decomposes the repository's raw synthetic code score by top-level directory, allowing you to pinpoint precisely which modules or components carry the highest AI authorship density. Directories with disproportionately high scores relative to their size warrant targeted manual review: concentrated AI signatures often trace back to mass-generated configuration layers, auto-ported test suites, LLM-scaffolded boilerplate classes, or entire subsystems authored under heavy copilot assistance. Use this view to prioritise your human code-review effort.
The scanner identified 4140 distinct pattern matches across 20 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a statistically significant AI authorship indicator. Expand any category row to inspect the individual file paths, line numbers, code snippets, and the lexical context (CODE, COMMENT, or STRING) in which each match was detected.
Reading the findings table: The Severity column indicates the diagnostic confidence level (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — comment-context matches receive a ×1.5 weight because LLMs systematically over-annotate. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier as dense clusters constitute far stronger evidence of synthetic authorship than isolated hits.
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW⚡ | tests/unit_tests/test_comm.py | 301 | def test_send_non_contiguous_tensor(self): | CODE |
| LOW⚡ | tests/unit_tests/test_comm.py | 309 | def test_send_non_contiguous_tensor_list(self): | CODE |
| LOW⚡ | tests/unit_tests/test_comm.py | 317 | def test_send_non_contiguous_tensor_dict(self): | CODE |
| LOW⚡ | tests/unit_tests/test_comm.py | 612 | def test_recv_tensor_dataclass(self, async_op=False): | CODE |
| LOW⚡ | tests/unit_tests/test_comm.py | 615 | def test_recv_tensor_list_dataclass(self, async_op=False): | CODE |
| LOW⚡ | tests/unit_tests/test_comm.py | 618 | def test_recv_tensor_dict_dataclass(self, async_op=False): | CODE |
| LOW⚡ | tests/unit_tests/test_comm.py | 625 | async def test_recv_tensor_dataclass_asyncio(self): | CODE |
| LOW⚡ | tests/unit_tests/test_comm.py | 628 | async def test_recv_tensor_list_dataclass_asyncio(self): | CODE |
| LOW⚡ | tests/unit_tests/test_comm.py | 631 | async def test_recv_tensor_dict_dataclass_asyncio(self): | CODE |
| LOW⚡ | tests/unit_tests/test_comm.py | 859 | def test_cross_group_broadcast_object(self, groups, async_op=False): | CODE |
| LOW⚡ | tests/unit_tests/test_comm.py | 863 | def test_cross_group_broadcast_tensor(self, groups, on_cpu, async_op=False): | CODE |
| LOW⚡ | tests/unit_tests/test_comm.py | 868 | def test_broadcast_object_with_src(self, groups, src, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 126 | def make_non_contiguous_tensor(device): | CODE |
| LOW | tests/unit_tests/test_comm.py | 186 | def test_send_plain_dataclass(self, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 209 | def test_send_mixed_tensor_list(self, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 220 | def test_send_mixed_tensor_dict(self, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 231 | def test_send_mixed_tensor_list_dataclass(self, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 247 | def test_send_mixed_tensor_dict_dataclass(self, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 268 | def test_send_tensor_dataclass(self, on_cpu, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 276 | def test_send_tensor_list_dataclass(self, on_cpu, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 288 | def test_send_tensor_dict_dataclass(self, on_cpu, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 328 | def test_send_non_contiguous_tensor_dataclass(self): | CODE |
| LOW | tests/unit_tests/test_comm.py | 338 | def test_send_non_contiguous_tensor_list_dataclass(self): | CODE |
| LOW | tests/unit_tests/test_comm.py | 350 | def test_send_non_contiguous_tensor_dict_dataclass(self): | CODE |
| LOW | tests/unit_tests/test_comm.py | 365 | def test_send_tensor_non_contiguous_inplace(self): | CODE |
| LOW | tests/unit_tests/test_comm.py | 380 | async def test_send_tensor_dataclass_asyncio(self, on_cpu): | CODE |
| LOW | tests/unit_tests/test_comm.py | 390 | async def test_send_tensor_list_dataclass_asyncio(self, on_cpu): | CODE |
| LOW | tests/unit_tests/test_comm.py | 403 | async def test_send_tensor_dict_dataclass_asyncio(self, on_cpu): | CODE |
| LOW | tests/unit_tests/test_comm.py | 432 | def test_consecutive_send_recv(self, on_cpu): | CODE |
| LOW | tests/unit_tests/test_comm.py | 565 | def test_consecutive_send_recv(self, on_cpu): | CODE |
| LOW | tests/unit_tests/test_comm.py | 597 | def test_recv_plain_dataclass(self, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 661 | async def test_async_wait_yields_control(self): | CODE |
| LOW | tests/unit_tests/test_comm.py | 710 | def test_broadcast_plain_dataclass(self, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 723 | def test_broadcast_tensor_list(self, on_cpu, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 728 | def test_broadcast_mixed_tensor_list(self, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 743 | def test_broadcast_mixed_tensor_list_dataclass(self, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 762 | def test_broadcast_tensor_dict(self, on_cpu, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 767 | def test_broadcast_tensor_dataclass(self, on_cpu, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 780 | def test_broadcast_tensor_list_dataclass(self, on_cpu, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 793 | def test_broadcast_tensor_dict_dataclass(self, on_cpu, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 808 | async def test_broadcast_tensor_asyncio(self, on_cpu): | CODE |
| LOW | tests/unit_tests/test_comm.py | 823 | async def test_cross_group_broadcast_tensor_asyncio(self, groups, on_cpu): | CODE |
| LOW | tests/unit_tests/test_comm.py | 880 | def test_broadcast_tensor_with_src(self, groups, src, on_cpu, async_op=False): | CODE |
| LOW | tests/unit_tests/test_comm.py | 893 | def test_broadcast_tensor_dataclass_with_src( | CODE |
| LOW | tests/unit_tests/test_comm.py | 916 | def test_cross_group_broadcast_tensor_dataclass( | CODE |
| LOW | tests/unit_tests/test_comm.py | 937 | async def test_broadcast_tensor_dataclass_asyncio(self, on_cpu): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1085 | def test_object_communication(self, worker_groups, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1099 | def test_plain_dataclass_communication(self, worker_groups, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1119 | def test_tensor_communication(self, worker_groups, on_cpu, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1139 | def test_tensor_list_communication(self, worker_groups, on_cpu, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1157 | def test_mixed_tensor_list_communication(self, worker_groups, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1178 | def test_mixed_tensor_dict_communication(self, worker_groups, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1204 | def test_mixed_tensor_list_dataclass_communication(self, worker_groups, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1228 | def test_mixed_tensor_dict_dataclass_communication(self, worker_groups, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1260 | def test_tensor_dict_communication(self, worker_groups, on_cpu, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1282 | def test_inplace_tensor_communication(self, worker_groups, on_cpu, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1302 | def test_tensor_dataclass_communication(self, worker_groups, on_cpu, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1324 | def test_tensor_list_dataclass_communication(self, worker_groups, on_cpu, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1349 | def test_tensor_dict_dataclass_communication(self, worker_groups, on_cpu, async_op): | CODE |
| LOW | tests/unit_tests/test_comm.py | 1373 | def test_asyncio_communication(self, worker_groups, on_cpu): | CODE |
| 1093 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | .cursor/skills/test-install/driver.py | 1 | #!/usr/bin/env python3 | COMMENT |
| LOW | .cursor/skills/install-check/check.sh | 1 | #!/usr/bin/env bash | COMMENT |
| LOW | .cursor/skills/create-pr/lint_pr.py | 1 | #!/usr/bin/env python3 | COMMENT |
| LOW | evaluations/eval_embodied_agent.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_comm.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/check_missing_init.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_robot_autoconfig.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/bench_channel.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_history_manager.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_distributed_ray_log_collector.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_robotwin_seed_partition.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/check_import_rlinf_package.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_weight_syncer.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/_robot_autoconfig_cluster.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_patcher_skipimport.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_comm_mapper.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_overlap_env_bootstrap.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_lerobot_compat.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_math_dataset_multithread.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_nvidia_profiling.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_distributed_tracing.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_placement.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_broadcast_sync.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_channel.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_net_emulation.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_worker.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_intra_gpu_comm.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_maniskill_offload_env.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_intel_gpu_platform.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_custom_model_registration.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_metric_utils.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_cluster_config.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_value_model_embedding_scale.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_rot6d.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_auto_placement.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/bench_broadcast.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_delay_sampler.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/unit_tests/test_gloo_fallback_staging.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/parity_tests/check.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/parity_tests/compare_baseline.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/parity_tests/parse_success_once.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/parity_tests/run_all.sh | 41 | # ENV_NAME: Environment name (maniskill_libero, behavior, isaaclab, metaworld, calvin, etc.) | COMMENT |
| LOW | tests/parity_tests/run_all.sh | 61 | # "maniskill_libero openvla openvla maniskill_async_ppo_openvla 1 120 -1" | COMMENT |
| LOW | tests/parity_tests/run_all.sh | 81 | COMMENT | |
| LOW | tests/parity_tests/plot_success_once.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | tests/parity_tests/analyze_logs.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | …ts/e2e_tests/agent/searchr1/qwen2.5-1.5b-megatron.yaml | 241 | overlap_param_gather: False | COMMENT |
| LOW | .claude/skills/test-install/driver.py | 1 | #!/usr/bin/env python3 | COMMENT |
| LOW | .claude/skills/install-check/check.sh | 1 | #!/usr/bin/env bash | COMMENT |
| LOW | .claude/skills/create-pr/lint_pr.py | 1 | #!/usr/bin/env python3 | COMMENT |
| LOW | toolkits/__init__.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | toolkits/lerobot/calculate_norm_stats.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | toolkits/lerobot/generate_dreamzero_metadata.py | 1 | #!/usr/bin/env python3 | COMMENT |
| LOW | toolkits/lerobot/visualize_lerobot_dataset.py | 1 | #!/usr/bin/env python3 | COMMENT |
| LOW | toolkits/lerobot/merge_lerobot_datasets.py | 1 | #!/usr/bin/env python3 | COMMENT |
| LOW | toolkits/lerobot/collect_maniskill_peg_lerobot_joint.py | 1 | #!/usr/bin/env python3 | COMMENT |
| LOW | toolkits/replay_buffer/visualize_headless.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | toolkits/replay_buffer/visualize.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| LOW | toolkits/replay_buffer/merge_or_split_replay_buffer.py | 1 | # Copyright 2025 The RLinf Authors. | COMMENT |
| LOW | toolkits/standalone_eval_scripts/__init__.py | 1 | # Copyright 2026 The RLinf Authors. | COMMENT |
| 807 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | .cursor/skills/install-check/check.sh | 31 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 34 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 46 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 48 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 63 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 67 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 91 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 94 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 127 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 130 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 141 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 145 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/unit_tests/test_lerobot_compat.py | 151 | # -------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/unit_tests/test_lerobot_compat.py | 153 | # -------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/test_broadcast_sync.py | 29 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/test_broadcast_sync.py | 31 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/test_broadcast_sync.py | 55 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/test_broadcast_sync.py | 57 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/test_broadcast_sync.py | 189 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/test_broadcast_sync.py | 191 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/bench_broadcast.py | 102 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/bench_broadcast.py | 104 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/bench_broadcast.py | 185 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/bench_broadcast.py | 187 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/bench_broadcast.py | 480 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/unit_tests/bench_broadcast.py | 482 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 31 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 34 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 46 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 48 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 63 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 67 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 91 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 94 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 127 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 130 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 141 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 145 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | toolkits/lerobot/merge_lerobot_datasets.py | 184 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | toolkits/lerobot/merge_lerobot_datasets.py | 186 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | toolkits/lerobot/merge_lerobot_datasets.py | 199 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | toolkits/lerobot/merge_lerobot_datasets.py | 201 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | toolkits/lerobot/merge_lerobot_datasets.py | 268 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | toolkits/lerobot/merge_lerobot_datasets.py | 270 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | toolkits/lerobot/merge_lerobot_datasets.py | 275 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | toolkits/lerobot/merge_lerobot_datasets.py | 277 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | toolkits/lerobot/merge_lerobot_datasets.py | 344 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | toolkits/lerobot/merge_lerobot_datasets.py | 346 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | toolkits/dual_franka/backfill_tcp_rot6d.py | 62 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | toolkits/dual_franka/backfill_tcp_rot6d.py | 64 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | toolkits/dual_franka/backfill_tcp_rot6d.py | 133 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | toolkits/dual_franka/backfill_tcp_rot6d.py | 135 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | toolkits/dual_franka/backfill_tcp_rot6d.py | 245 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | toolkits/dual_franka/backfill_tcp_rot6d.py | 247 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | toolkits/dual_franka/backfill_tcp_rot6d.py | 298 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | toolkits/dual_franka/backfill_tcp_rot6d.py | 300 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | toolkits/realworld_check/test_gello.py | 57 | # ───────────────────────── shared helpers ────────────────────────────── | COMMENT |
| MEDIUM | toolkits/realworld_check/test_gello.py | 238 | # ───────────────────── subcommand: align-check ──────────────────────── | COMMENT |
| MEDIUM | toolkits/realworld_check/test_gello.py | 324 | # ─────────────────── subcommand: align-sequential ───────────────────── | COMMENT |
| MEDIUM | toolkits/realworld_check/test_gello.py | 511 | # ───────────────────── subcommand: calibrate ────────────────────────── | COMMENT |
| 184 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | .cursor/skills/create-pr/lint_pr.py | 34 | CODE | |
| LOW | tests/unit_tests/test_nvidia_profiling.py | 17 | CODE | |
| LOW | tests/unit_tests/test_intel_gpu_platform.py | 17 | CODE | |
| LOW | tests/unit_tests/bench_broadcast.py | 84 | CODE | |
| LOW | tests/unit_tests/test_delay_sampler.py | 16 | CODE | |
| LOW | .claude/skills/create-pr/lint_pr.py | 34 | CODE | |
| LOW | toolkits/lerobot/generate_dreamzero_metadata.py | 49 | CODE | |
| LOW | toolkits/lerobot/visualize_lerobot_dataset.py | 44 | CODE | |
| LOW | toolkits/lerobot/merge_lerobot_datasets.py | 40 | CODE | |
| LOW | toolkits/lerobot/collect_maniskill_peg_lerobot_joint.py | 34 | CODE | |
| LOW | toolkits/lerobot/collect_maniskill_peg_lerobot_joint.py | 177 | CODE | |
| LOW | toolkits/lerobot/collect_maniskill_peg_lerobot_joint.py | 451 | CODE | |
| LOW | toolkits/dual_franka/utils_lerobot.py | 16 | CODE | |
| LOW | toolkits/dual_franka/backfill_tcp_rot6d.py | 32 | CODE | |
| LOW | toolkits/realworld_check/test_gello.py | 29 | CODE | |
| LOW | toolkits/realworld_check/collect_monitor.py | 27 | CODE | |
| LOW | .codex/skills/create-pr/lint_pr.py | 34 | CODE | |
| LOW | rlinf/algorithms/registry.py | 160 | CODE | |
| LOW | rlinf/algorithms/__init__.py | 14 | CODE | |
| LOW | rlinf/algorithms/__init__.py | 16 | CODE | |
| LOW | rlinf/algorithms/__init__.py | 16 | CODE | |
| LOW | rlinf/algorithms/__init__.py | 16 | CODE | |
| LOW | rlinf/algorithms/rlt/expert.py | 20 | CODE | |
| LOW | rlinf/algorithms/rlt/__init__.py | 15 | CODE | |
| LOW | rlinf/algorithms/rlt/__init__.py | 16 | CODE | |
| LOW | rlinf/algorithms/rlt/__init__.py | 17 | CODE | |
| LOW | rlinf/algorithms/rlt/__init__.py | 17 | CODE | |
| LOW | rlinf/algorithms/rlt/__init__.py | 17 | CODE | |
| LOW | rlinf/algorithms/rlt/__init__.py | 17 | CODE | |
| LOW | rlinf/algorithms/rlt/__init__.py | 17 | CODE | |
| LOW | rlinf/algorithms/rlt/__init__.py | 24 | CODE | |
| LOW | rlinf/algorithms/offline/process/mixture_config.py | 23 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 15 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 15 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 15 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 16 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 16 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 17 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 18 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 18 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 18 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 18 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 18 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 18 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 18 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 18 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 28 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 28 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 29 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 29 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 29 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 29 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 29 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 36 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 36 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 36 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 37 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 37 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 37 | CODE | |
| LOW | rlinf/scheduler/__init__.py | 37 | CODE | |
| 569 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | .cursor/skills/test-install/driver.py | 0 | drive the embodied install + e2e pipeline for one model/env. this is the harness behind the `test-install` skill. it is | STRING |
| HIGH | .claude/skills/test-install/driver.py | 0 | drive the embodied install + e2e pipeline for one model/env. this is the harness behind the `test-install` skill. it is | STRING |
| HIGH | .codex/skills/test-install/driver.py | 0 | drive the embodied install + e2e pipeline for one model/env. this is the harness behind the `test-install` skill. it is | STRING |
| HIGH | .cursor/skills/test-install/driver.py | 0 | walk up from cwd to the dir holding requirements/install.sh. | STRING |
| HIGH | .claude/skills/test-install/driver.py | 0 | walk up from cwd to the dir holding requirements/install.sh. | STRING |
| HIGH | .codex/skills/test-install/driver.py | 0 | walk up from cwd to the dir holding requirements/install.sh. | STRING |
| HIGH | .cursor/skills/test-install/driver.py | 0 | match ci jobs by (model, env). with both args, require an exact match. with only one token (env is none), match it again | STRING |
| HIGH | .claude/skills/test-install/driver.py | 0 | match ci jobs by (model, env). with both args, require an exact match. with only one token (env is none), match it again | STRING |
| HIGH | .codex/skills/test-install/driver.py | 0 | match ci jobs by (model, env). with both args, require an exact match. with only one token (env is none), match it again | STRING |
| HIGH | .cursor/skills/test-install/driver.py | 0 | return list of (key, value, kind, exists) for absolute paths in config. | STRING |
| HIGH | .claude/skills/test-install/driver.py | 0 | return list of (key, value, kind, exists) for absolute paths in config. | STRING |
| HIGH | .codex/skills/test-install/driver.py | 0 | return list of (key, value, kind, exists) for absolute paths in config. | STRING |
| HIGH | .cursor/skills/create-pr/lint_pr.py | 0 | lint an rlinf pull-request title and body against contributing.md. checks the conventional commits title format and the | STRING |
| HIGH | .claude/skills/create-pr/lint_pr.py | 0 | lint an rlinf pull-request title and body against contributing.md. checks the conventional commits title format and the | STRING |
| HIGH | .codex/skills/create-pr/lint_pr.py | 0 | lint an rlinf pull-request title and body against contributing.md. checks the conventional commits title format and the | STRING |
| HIGH | .cursor/skills/create-pr/lint_pr.py | 0 | return, per line, whether it sits inside a fenced code block. | STRING |
| HIGH | .claude/skills/create-pr/lint_pr.py | 0 | return, per line, whether it sits inside a fenced code block. | STRING |
| HIGH | .codex/skills/create-pr/lint_pr.py | 0 | return, per line, whether it sits inside a fenced code block. | STRING |
| HIGH | .cursor/skills/create-pr/lint_pr.py | 0 | split a pr body into `###`-delimited sections (code fences respected). | STRING |
| HIGH | .claude/skills/create-pr/lint_pr.py | 0 | split a pr body into `###`-delimited sections (code fences respected). | STRING |
| HIGH | .codex/skills/create-pr/lint_pr.py | 0 | split a pr body into `###`-delimited sections (code fences respected). | STRING |
| HIGH | .cursor/skills/create-pr/lint_pr.py | 0 | best-effort imperative form of a past-tense/third-person verb. | STRING |
| HIGH | .claude/skills/create-pr/lint_pr.py | 0 | best-effort imperative form of a past-tense/third-person verb. | STRING |
| HIGH | .codex/skills/create-pr/lint_pr.py | 0 | best-effort imperative form of a past-tense/third-person verb. | STRING |
| HIGH | .cursor/skills/create-pr/lint_pr.py | 0 | catch bodies pasted with a uniform leading indent (the #1421 defect). | STRING |
| HIGH | .claude/skills/create-pr/lint_pr.py | 0 | catch bodies pasted with a uniform leading indent (the #1421 defect). | STRING |
| HIGH | .codex/skills/create-pr/lint_pr.py | 0 | catch bodies pasted with a uniform leading indent (the #1421 defect). | STRING |
| HIGH | .cursor/skills/create-pr/lint_pr.py | 0 | true when line `i` is an indented continuation of a markdown list. | STRING |
| HIGH | .claude/skills/create-pr/lint_pr.py | 0 | true when line `i` is an indented continuation of a markdown list. | STRING |
| HIGH | .codex/skills/create-pr/lint_pr.py | 0 | true when line `i` is an indented continuation of a markdown list. | STRING |
| HIGH | .cursor/skills/create-pr/lint_pr.py | 0 | flag markdown tables whose rows disagree on leading indentation. | STRING |
| HIGH | .claude/skills/create-pr/lint_pr.py | 0 | flag markdown tables whose rows disagree on leading indentation. | STRING |
| HIGH | .codex/skills/create-pr/lint_pr.py | 0 | flag markdown tables whose rows disagree on leading indentation. | STRING |
| HIGH | .cursor/skills/create-pr/lint_pr.py | 0 | de-indent a uniformly-indented body and drop the template hints. the de-indent only runs when :func:`check_indentation` | STRING |
| HIGH | .claude/skills/create-pr/lint_pr.py | 0 | de-indent a uniformly-indented body and drop the template hints. the de-indent only runs when :func:`check_indentation` | STRING |
| HIGH | .codex/skills/create-pr/lint_pr.py | 0 | de-indent a uniformly-indented body and drop the template hints. the de-indent only runs when :func:`check_indentation` | STRING |
| HIGH | .cursor/skills/create-pr/lint_pr.py | 0 | join hard-wrapped prose into one line per paragraph. headings, list items, tables, and fenced code are left untouched. | STRING |
| HIGH | .claude/skills/create-pr/lint_pr.py | 0 | join hard-wrapped prose into one line per paragraph. headings, list items, tables, and fenced code are left untouched. | STRING |
| HIGH | .codex/skills/create-pr/lint_pr.py | 0 | join hard-wrapped prose into one line per paragraph. headings, list items, tables, and fenced code are left untouched. | STRING |
| HIGH | .cursor/skills/create-pr/lint_pr.py | 0 | \ ### description <one sentence: what this pr changes and where> - <user-visible change 1> - <user-visible change 2> ### | STRING |
| HIGH | .claude/skills/create-pr/lint_pr.py | 0 | \ ### description <one sentence: what this pr changes and where> - <user-visible change 1> - <user-visible change 2> ### | STRING |
| HIGH | .codex/skills/create-pr/lint_pr.py | 0 | \ ### description <one sentence: what this pr changes and where> - <user-visible change 1> - <user-visible change 2> ### | STRING |
| HIGH | tests/unit_tests/test_comm.py | 0 | return whether the worker accelerator backend is available. | STRING |
| HIGH | tests/unit_tests/bench_channel.py | 0 | return whether the worker accelerator backend is available. | STRING |
| HIGH | tests/unit_tests/test_channel.py | 0 | return whether the worker accelerator backend is available. | STRING |
| HIGH | tests/unit_tests/test_worker.py | 0 | return whether the worker accelerator backend is available. | STRING |
| HIGH | tests/unit_tests/test_intra_gpu_comm.py | 0 | return whether the worker accelerator backend is available. | STRING |
| HIGH | tests/unit_tests/test_gloo_fallback_staging.py | 0 | return whether the worker accelerator backend is available. | STRING |
| HIGH | tests/unit_tests/test_comm.py | 0 | return accelerator count through the worker backend abstraction. | STRING |
| HIGH | tests/unit_tests/test_intra_gpu_comm.py | 0 | return accelerator count through the worker backend abstraction. | STRING |
| HIGH | tests/unit_tests/test_cluster_config.py | 0 | return accelerator count through the worker backend abstraction. | STRING |
| HIGH | rlinf/scheduler/hardware/accelerators/kunlun_xpu.py | 0 | get the environment variables related to the accelerator. args: accelerator_type (acceleratortype): the type of the acce | STRING |
| HIGH | rlinf/scheduler/hardware/accelerators/ascend_npu.py | 0 | get the environment variables related to the accelerator. args: accelerator_type (acceleratortype): the type of the acce | STRING |
| HIGH | rlinf/scheduler/hardware/accelerators/nvidia_gpu.py | 0 | get the environment variables related to the accelerator. args: accelerator_type (acceleratortype): the type of the acce | STRING |
| HIGH | rlinf/scheduler/hardware/accelerators/musa_gpu.py | 0 | get the environment variables related to the accelerator. args: accelerator_type (acceleratortype): the type of the acce | STRING |
| HIGH | rlinf/scheduler/hardware/accelerators/amd_gpu.py | 0 | get the environment variables related to the accelerator. args: accelerator_type (acceleratortype): the type of the acce | STRING |
| HIGH | rlinf/scheduler/hardware/accelerators/intel_gpu.py | 0 | get the environment variables related to the accelerator. args: accelerator_type (acceleratortype): the type of the acce | STRING |
| HIGH | rlinf/scheduler/hardware/accelerators/accelerator.py | 0 | get the environment variables related to the accelerator. args: accelerator_type (acceleratortype): the type of the acce | STRING |
| HIGH | rlinf/scheduler/hardware/accelerators/kunlun_xpu.py | 0 | get the network socket interface name environment variable. returns: str: the network socket interface name environment | STRING |
| HIGH | rlinf/scheduler/hardware/accelerators/ascend_npu.py | 0 | get the network socket interface name environment variable. returns: str: the network socket interface name environment | STRING |
| 43 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | .cursor/skills/test-install/driver.py | 100 | CODE | |
| LOW | .cursor/skills/test-install/driver.py | 154 | CODE | |
| LOW | .cursor/skills/create-pr/lint_pr.py | 238 | CODE | |
| LOW | .cursor/skills/create-pr/lint_pr.py | 573 | CODE | |
| LOW | .cursor/skills/create-pr/lint_pr.py | 933 | CODE | |
| LOW | tests/unit_tests/test_distributed_ray_log_collector.py | 172 | CODE | |
| LOW | tests/unit_tests/test_weight_syncer.py | 629 | CODE | |
| LOW | tests/unit_tests/test_channel.py | 895 | CODE | |
| LOW | tests/parity_tests/compare_baseline.py | 73 | CODE | |
| LOW | tests/parity_tests/compare_baseline.py | 289 | CODE | |
| LOW | tests/parity_tests/compare_baseline.py | 470 | CODE | |
| LOW | tests/parity_tests/plot_success_once.py | 172 | CODE | |
| LOW | tests/parity_tests/analyze_logs.py | 60 | CODE | |
| LOW | .claude/skills/test-install/driver.py | 100 | CODE | |
| LOW | .claude/skills/test-install/driver.py | 154 | CODE | |
| LOW | .claude/skills/create-pr/lint_pr.py | 238 | CODE | |
| LOW | .claude/skills/create-pr/lint_pr.py | 573 | CODE | |
| LOW | .claude/skills/create-pr/lint_pr.py | 933 | CODE | |
| LOW | toolkits/lerobot/visualize_lerobot_dataset.py | 426 | CODE | |
| LOW | toolkits/lerobot/merge_lerobot_datasets.py | 162 | CODE | |
| LOW | toolkits/replay_buffer/visualize_headless.py | 33 | CODE | |
| LOW | toolkits/replay_buffer/visualize.py | 126 | CODE | |
| LOW | toolkits/replay_buffer/visualize.py | 245 | CODE | |
| LOW | …olkits/standalone_eval_scripts/dexbotic/libero_eval.py | 74 | CODE | |
| LOW | toolkits/standalone_eval_scripts/openpi/calvin_eval.py | 48 | CODE | |
| LOW | toolkits/standalone_eval_scripts/openpi/libero_eval.py | 74 | CODE | |
| LOW | …lkits/standalone_eval_scripts/openpi/metaworld_eval.py | 45 | CODE | |
| LOW | toolkits/dual_franka/delete_lerobot.py | 87 | CODE | |
| LOW | toolkits/dual_franka/delete_lerobot.py | 120 | CODE | |
| LOW | toolkits/dual_franka/delete_lerobot.py | 238 | CODE | |
| LOW | toolkits/dual_franka/delete_lerobot.py | 354 | CODE | |
| LOW | toolkits/dual_franka/delete_lerobot.py | 534 | CODE | |
| LOW | toolkits/dual_franka/merge_lerobot.py | 375 | CODE | |
| LOW | toolkits/dual_franka/backfill_tcp_rot6d.py | 261 | CODE | |
| LOW | toolkits/realworld_check/test_gello.py | 98 | CODE | |
| LOW | toolkits/realworld_check/test_gello.py | 167 | CODE | |
| LOW | toolkits/realworld_check/test_gello.py | 391 | CODE | |
| LOW | toolkits/realworld_check/test_gello.py | 522 | CODE | |
| LOW | toolkits/realworld_check/collect_monitor.py | 118 | CODE | |
| LOW | toolkits/realworld_check/test_franky_controller.py | 56 | CODE | |
| LOW | toolkits/realworld_check/test_pico_data.py | 85 | CODE | |
| LOW | toolkits/realworld_check/test_turtle2_controller.py | 22 | CODE | |
| LOW | toolkits/realworld_check/test_franka_controller.py | 60 | CODE | |
| LOW | toolkits/auto_placement/auto_placement_worker.py | 55 | CODE | |
| LOW | toolkits/auto_placement/workflow.py | 88 | CODE | |
| LOW | .codex/skills/test-install/driver.py | 100 | CODE | |
| LOW | .codex/skills/test-install/driver.py | 154 | CODE | |
| LOW | .codex/skills/create-pr/lint_pr.py | 238 | CODE | |
| LOW | .codex/skills/create-pr/lint_pr.py | 573 | CODE | |
| LOW | .codex/skills/create-pr/lint_pr.py | 933 | CODE | |
| LOW | rlinf/config.py | 162 | CODE | |
| LOW | rlinf/config.py | 217 | CODE | |
| LOW | rlinf/config.py | 858 | CODE | |
| LOW | rlinf/config.py | 1233 | CODE | |
| LOW | rlinf/config.py | 1401 | CODE | |
| LOW | rlinf/config.py | 1527 | CODE | |
| LOW | rlinf/config.py | 1101 | CODE | |
| LOW | rlinf/algorithms/utils.py | 177 | CODE | |
| LOW | rlinf/algorithms/utils.py | 280 | CODE | |
| LOW | rlinf/algorithms/rewards/rstar2/__init__.py | 56 | CODE | |
| 346 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | tests/unit_tests/check_import_rlinf_package.py | 91 | except Exception: | CODE |
| LOW | tests/unit_tests/_robot_autoconfig_cluster.py | 323 | except Exception as exc: | CODE |
| LOW | tests/parity_tests/compare_baseline.py | 270 | except Exception as e: | CODE |
| MEDIUM | tests/parity_tests/compare_baseline.py | 584 | print(f"Error: baseline file {args.baseline} does not exist") | CODE |
| MEDIUM | tests/parity_tests/compare_baseline.py | 595 | print(f"Error: {path} does not exist") | CODE |
| LOW | tests/parity_tests/compare_baseline.py | 640 | except Exception as e: | CODE |
| LOW | tests/parity_tests/parse_success_once.py | 176 | except Exception as e: | CODE |
| MEDIUM | tests/parity_tests/parse_success_once.py | 177 | print(f"Error processing {log_file}: {e}") | CODE |
| MEDIUM | tests/parity_tests/parse_success_once.py | 283 | print(f"Error: {path} does not exist") | CODE |
| MEDIUM | tests/parity_tests/plot_success_once.py | 256 | print(f"Error: {path} does not exist") | CODE |
| LOW | tests/parity_tests/plot_success_once.py | 314 | except Exception as e: | CODE |
| LOW | tests/parity_tests/plot_success_once.py | 356 | except Exception as e: | CODE |
| MEDIUM⚡ | tests/parity_tests/analyze_logs.py | 112 | print(f"Error: {path} does not exist") | CODE |
| LOW | tests/parity_tests/analyze_logs.py | 155 | except Exception as e: | CODE |
| LOW | tests/parity_tests/analyze_logs.py | 195 | except Exception as e: | CODE |
| LOW | tests/parity_tests/analyze_logs.py | 242 | except Exception as e: | CODE |
| LOW | toolkits/lerobot/visualize_lerobot_dataset.py | 220 | except Exception: | CODE |
| LOW | toolkits/lerobot/collect_maniskill_peg_lerobot_joint.py | 371 | except Exception as exc: # noqa: BLE001 | CODE |
| LOW | toolkits/lerobot/collect_maniskill_peg_lerobot_joint.py | 724 | except Exception as exc: # noqa: BLE001 | CODE |
| MEDIUM | toolkits/lerobot/collect_maniskill_peg_lerobot_joint.py | 175 | def _register_camera_collection_env() -> None: | CODE |
| LOW | toolkits/replay_buffer/visualize_headless.py | 140 | except Exception as e: | CODE |
| MEDIUM | toolkits/replay_buffer/visualize_headless.py | 141 | print(f"Error: {e}") | CODE |
| LOW | toolkits/realworld_check/test_gello.py | 201 | except Exception: | CODE |
| LOW | toolkits/realworld_check/test_gello.py | 208 | except Exception as e: | CODE |
| LOW | toolkits/realworld_check/test_gello.py | 287 | except Exception as e: | CODE |
| LOW | toolkits/realworld_check/test_gello.py | 430 | except Exception as e: | CODE |
| LOW | toolkits/realworld_check/test_franky_controller.py | 34 | except Exception: | CODE |
| LOW | toolkits/realworld_check/test_franky_controller.py | 188 | except Exception as e: | CODE |
| LOW | toolkits/realworld_check/test_franky_controller.py | 196 | except Exception: | CODE |
| LOW | requirements/install.sh | 673 | except Exception: | CODE |
| LOW | requirements/install.sh | 696 | except Exception: | CODE |
| LOW | requirements/install.sh | 1390 | except Exception: | CODE |
| LOW | rlinf/config.py | 195 | except Exception: | CODE |
| LOW | rlinf/algorithms/rewards/rstar2/__init__.py | 39 | except Exception as e: | CODE |
| LOW⚡ | rlinf/algorithms/rewards/rstar2/__init__.py | 114 | except Exception as e: | CODE |
| MEDIUM⚡ | rlinf/algorithms/rewards/rstar2/__init__.py | 115 | print(f"Error terminating {i}: {e}") | CODE |
| LOW⚡ | rlinf/algorithms/rewards/rstar2/__init__.py | 123 | except Exception as e: | CODE |
| MEDIUM⚡ | rlinf/algorithms/rewards/rstar2/__init__.py | 124 | print(f"Error killing {i}: {e}") | CODE |
| LOW⚡ | rlinf/algorithms/rewards/rstar2/__init__.py | 129 | except Exception: | CODE |
| LOW | rlinf/algorithms/rewards/rstar2/__init__.py | 150 | except Exception as e: | CODE |
| MEDIUM | rlinf/algorithms/rewards/rstar2/__init__.py | 151 | print(f"Error collecting result: {e}") | CODE |
| LOW⚡ | rlinf/algorithms/rewards/rstar2/__init__.py | 166 | except Exception as e: | CODE |
| MEDIUM⚡ | rlinf/algorithms/rewards/rstar2/__init__.py | 167 | print(f"Error terminating process {index}: {e}") | CODE |
| LOW⚡ | rlinf/algorithms/rewards/rstar2/__init__.py | 175 | except Exception: | CODE |
| LOW⚡ | rlinf/algorithms/rewards/rstar2/__init__.py | 181 | except Exception as e: | CODE |
| MEDIUM⚡ | rlinf/algorithms/rewards/rstar2/__init__.py | 182 | print(f"Error closing queue: {e}") | CODE |
| LOW | …thms/rewards/rstar2/fused_compute_score/math_verify.py | 38 | except Exception: | CODE |
| LOW | …ms/rewards/rstar2/fused_compute_score/compute_score.py | 25 | except Exception: | CODE |
| LOW | …ms/rewards/rstar2/fused_compute_score/compute_score.py | 32 | except Exception: | CODE |
| LOW | …ewards/rstar2/fused_compute_score/prime_math/grader.py | 111 | except Exception: | CODE |
| LOW | …ewards/rstar2/fused_compute_score/prime_math/grader.py | 224 | except Exception: | CODE |
| LOW | …ewards/rstar2/fused_compute_score/prime_math/grader.py | 227 | except Exception: | CODE |
| LOW | …ewards/rstar2/fused_compute_score/prime_math/grader.py | 309 | except Exception: | CODE |
| LOW | …ewards/rstar2/fused_compute_score/prime_math/grader.py | 335 | except Exception: | CODE |
| LOW | …ewards/rstar2/fused_compute_score/prime_math/grader.py | 350 | except Exception: | CODE |
| LOW | …ewards/rstar2/fused_compute_score/prime_math/grader.py | 364 | except Exception: | CODE |
| LOW | …ewards/rstar2/fused_compute_score/prime_math/grader.py | 374 | except Exception: | CODE |
| LOW | …ewards/rstar2/fused_compute_score/prime_math/grader.py | 417 | except Exception as e: | CODE |
| MEDIUM | …ewards/rstar2/fused_compute_score/prime_math/grader.py | 115 | def is_digit(s): | CODE |
| LOW | …ards/rstar2/fused_compute_score/prime_math/__init__.py | 30 | except Exception: | CODE |
| 211 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| CRITICAL | rlinf/models/embodiment/lingbotvla/__init__.py | 61 | model.vla_model.model.qwenvl_with_expert.qwenvl.eval() | CODE |
| CRITICAL | rlinf/models/embodiment/lingbotvla/__init__.py | 62 | for param in model.vla_model.model.qwenvl_with_expert.qwenvl.parameters(): | CODE |
| CRITICAL | rlinf/envs/maniskill/maniskill_env.py | 183 | proprioception: torch.Tensor = self.env.unwrapped.agent.robot.get_qpos().to( | CODE |
| CRITICAL⚡ | rlinf/envs/maniskill/maniskill_offload_env.py | 60 | "cuda_articulation_link_data": self.env.unwrapped.scene.px.cuda_articulation_link_data.torch().cpu(), | CODE |
| CRITICAL⚡ | rlinf/envs/maniskill/maniskill_offload_env.py | 61 | "cuda_articulation_qacc": self.env.unwrapped.scene.px.cuda_articulation_qacc.torch().cpu(), | CODE |
| CRITICAL⚡ | rlinf/envs/maniskill/maniskill_offload_env.py | 62 | "cuda_articulation_qf": self.env.unwrapped.scene.px.cuda_articulation_qf.torch().cpu(), | CODE |
| CRITICAL⚡ | rlinf/envs/maniskill/maniskill_offload_env.py | 63 | "cuda_articulation_qpos": self.env.unwrapped.scene.px.cuda_articulation_qpos.torch().cpu(), | CODE |
| CRITICAL⚡ | rlinf/envs/maniskill/maniskill_offload_env.py | 64 | "cuda_articulation_qvel": self.env.unwrapped.scene.px.cuda_articulation_qvel.torch().cpu(), | CODE |
| CRITICAL⚡ | rlinf/envs/maniskill/maniskill_offload_env.py | 65 | "cuda_articulation_target_qpos": self.env.unwrapped.scene.px.cuda_articulation_target_qpos.torch().cpu(), | CODE |
| CRITICAL⚡ | rlinf/envs/maniskill/maniskill_offload_env.py | 66 | "cuda_articulation_target_qvel": self.env.unwrapped.scene.px.cuda_articulation_target_qvel.torch().cpu(), | CODE |
| CRITICAL⚡ | rlinf/envs/maniskill/maniskill_offload_env.py | 67 | "cuda_rigid_body_data": self.env.unwrapped.scene.px.cuda_rigid_body_data.torch().cpu(), | CODE |
| CRITICAL⚡ | rlinf/envs/maniskill/maniskill_offload_env.py | 68 | "cuda_rigid_dynamic_data": self.env.unwrapped.scene.px.cuda_rigid_dynamic_data.torch().cpu(), | CODE |
| CRITICAL | rlinf/envs/maniskill/maniskill_offload_env.py | 112 | self.env.unwrapped.scene.px.gpu_update_articulation_kinematics() | CODE |
| CRITICAL | rlinf/envs/maniskill/peg_insertion_side_variants.py | 261 | tcp_pos = env.agent.tcp.pose.p.to(device, dtype=torch.float32) | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | toolkits/standalone_eval_scripts/dexbotic/__init__.py | 279 | __all__ = [ | CODE |
| LOW | rlinf/config.py | 159 | __all__ = ["build_config"] | CODE |
| LOW | rlinf/algorithms/rlt/expert.py | 22 | __all__ = ["build_expert_model_config", "predict_expert_actions"] | CODE |
| LOW | rlinf/algorithms/rlt/__init__.py | 26 | __all__ = [ | CODE |
| LOW | rlinf/algorithms/offline/process/distributed.py | 37 | logger = logging.getLogger(__name__) | CODE |
| LOW | rlinf/algorithms/offline/process/advantage.py | 69 | __all__ = ["quantile_threshold", "apply_boolean_label"] | CODE |
| LOW | rlinf/algorithms/offline/process/mixture_config.py | 83 | __all__ = [ | CODE |
| LOW | rlinf/scheduler/__init__.py | 52 | __all__ = [ | CODE |
| LOW | rlinf/scheduler/cluster/__init__.py | 25 | __all__ = [ | CODE |
| LOW | rlinf/scheduler/hardware/__init__.py | 37 | __all__ = [ | CODE |
| LOW | rlinf/scheduler/hardware/accelerators/__init__.py | 29 | __all__ = [ | CODE |
| LOW⚡ | rlinf/scheduler/hardware/accelerators/nvidia_gpu.py | 34 | logger = logging.getLogger(__name__) | CODE |
| LOW⚡ | rlinf/scheduler/hardware/accelerators/amd_gpu.py | 32 | logger = logging.getLogger(__name__) | CODE |
| LOW | rlinf/scheduler/hardware/robots/__init__.py | 22 | __all__ = [ | CODE |
| LOW | rlinf/scheduler/collective/__init__.py | 27 | __all__ = [ | CODE |
| LOW | rlinf/scheduler/channel/__init__.py | 18 | __all__ = ["Channel", "ChannelWorker", "WeightedItem"] | CODE |
| LOW | rlinf/scheduler/manager/__init__.py | 23 | __all__ = [ | CODE |
| LOW | rlinf/scheduler/worker/__init__.py | 39 | __all__ = [ | CODE |
| LOW | rlinf/scheduler/placement/__init__.py | 20 | __all__ = [ | CODE |
| LOW | rlinf/agents/tool_call/__init__.py | 28 | __all__ = [ | CODE |
| LOW | rlinf/agents/agentlightning/__init__.py | 17 | __all__ = ["RLinf"] | CODE |
| LOW | rlinf/utils/rot6d.py | 25 | __all__ = [ | CODE |
| LOW | rlinf/utils/utils.py | 670 | def set_rng_state(rng_state: dict) -> None: | CODE |
| LOW | …ls/ckpt_convertor/megatron_convertor/utils/__init__.py | 26 | __all__ = [ | CODE |
| LOW | rlinf/models/tokenization/__init__.py | 19 | __all__ = ["hf_tokenizer", "set_pad_token_id"] | CODE |
| LOW | rlinf/models/embodiment/starvla/starvla_action_model.py | 40 | logger = logging.getLogger(__name__) | CODE |
| LOW | rlinf/models/embodiment/starvla/__init__.py | 117 | __all__ = ["StarVLAForRLActionPrediction", "get_model"] | CODE |
| LOW | rlinf/models/embodiment/evo1/__init__.py | 218 | __all__ = ["Evo1ForRLActionPrediction", "get_model"] | CODE |
| LOW | rlinf/models/embodiment/evo1/utils/__init__.py | 21 | __all__ = ["build_evo1_inputs", "Normalizer", "NormalizationType"] | CODE |
| LOW | …nf/models/embodiment/value_model/recap/value_expert.py | 44 | logger = logging.getLogger(__name__) | CODE |
| LOW | rlinf/models/embodiment/value_model/recap/__init__.py | 30 | logger = logging.getLogger(__name__) | CODE |
| LOW | rlinf/models/embodiment/value_model/recap/__init__.py | 165 | __all__ = [ | CODE |
| LOW | …odels/embodiment/value_model/recap/checkpoint_utils.py | 27 | logger = logging.getLogger(__name__) | CODE |
| LOW | …f/models/embodiment/value_model/recap/data_collator.py | 30 | logger = logging.getLogger(__name__) | CODE |
| LOW | …f/models/embodiment/value_model/recap/data_collator.py | 245 | __all__ = ["ValueDataCollator", "stack_tensors"] | CODE |
| LOW | rlinf/models/embodiment/value_model/recap/processing.py | 63 | __all__ = [ | CODE |
| LOW | …models/embodiment/value_model/recap/modeling_critic.py | 33 | logger = logging.getLogger(__name__) | CODE |
| LOW | …models/embodiment/value_model/recap/modeling_critic.py | 1094 | __all__ = [ | CODE |
| LOW | …odels/embodiment/value_model/steam/checkpoint_merge.py | 42 | logger = logging.getLogger(__name__) | CODE |
| LOW | …odels/embodiment/value_model/steam/checkpoint_merge.py | 365 | __all__ = [ | CODE |
| LOW | rlinf/models/embodiment/value_model/steam/__init__.py | 41 | logger = logging.getLogger(__name__) | CODE |
| LOW | rlinf/models/embodiment/value_model/steam/__init__.py | 389 | __all__ = [ | CODE |
| LOW | …bodiment/value_model/steam/ensemble_modeling_critic.py | 49 | logger = logging.getLogger(__name__) | CODE |
| LOW | …bodiment/value_model/steam/ensemble_modeling_critic.py | 396 | __all__ = [ | CODE |
| LOW | rlinf/models/embodiment/value_model/steam/processing.py | 62 | __all__ = [ | CODE |
| LOW | …/models/embodiment/value_model/steam/modeling_steam.py | 549 | __all__ = [ | CODE |
| LOW | …models/embodiment/value_model/steam/modeling_critic.py | 67 | logger = logging.getLogger(__name__) | CODE |
| LOW | …models/embodiment/value_model/steam/modeling_critic.py | 632 | __all__ = [ | CODE |
| LOW | …ment/value_model/value_common/image_text_processing.py | 49 | logger = logging.getLogger(__name__) | CODE |
| LOW | …ment/value_model/value_common/image_text_processing.py | 690 | __all__ = [ | CODE |
| LOW | …models/embodiment/value_model/value_common/__init__.py | 27 | __all__ = [ | CODE |
| LOW | rlinf/models/embodiment/openpi_rlinf/rl_action_model.py | 100 | def set_global_step(self, global_step: int) -> None: | CODE |
| LOW | …models/embodiment/openpi_rlinf/utils/model_builders.py | 34 | logger = logging.getLogger(__name__) | CODE |
| LOW | rlinf/models/embodiment/dreamzero/patch/__init__.py | 24 | __all__ = [ | CODE |
| LOW | rlinf/models/embodiment/molmoact2/molmoact2_policy.py | 179 | __all__ = ["MolmoAct2ForRLActionPrediction"] | CODE |
| LOW | rlinf/models/embodiment/molmoact2/__init__.py | 94 | __all__ = ["MolmoAct2ForRLActionPrediction", "get_model"] | CODE |
| LOW | rlinf/models/embodiment/gr00t/__init__.py | 45 | __all__ = ["get_model"] | CODE |
| LOW | rlinf/models/embodiment/reward/__init__.py | 24 | __all__ = [ | CODE |
| LOW | …ls/embodiment/reward/vlm_reward_utils/reward_parser.py | 22 | logger = logging.getLogger(__name__) | CODE |
| LOW | rlinf/models/embodiment/openpi/transforms/__init__.py | 20 | __all__ = [ | CODE |
| 109 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | toolkits/dual_franka/delete_lerobot.py | 88 | Parse --delete command-line arguments. Args: delete_args: List of deletion specifications in format "id_nam | STRING |
| HIGH | toolkits/dual_franka/delete_lerobot.py | 123 | Precheck directory structure and validate deletion list. Args: data_dir: Root data directory. delet | STRING |
| HIGH | toolkits/dual_franka/delete_lerobot.py | 241 | Rename remaining parquets and rewrite episode_index and index columns. Args: affected: Dictionary with dele | STRING |
| HIGH | toolkits/dual_franka/merge_lerobot.py | 131 | Validate that every per-id directory is well-formed and consistent. Args: rank_dir: Directory containing on | STRING |
| HIGH | toolkits/dual_franka/merge_lerobot.py | 276 | Rewrite ``episode_index`` and ``index`` columns of every parquet. ``index`` is made globally contiguous by accumula | STRING |
| HIGH | rlinf/scheduler/channel/channel.py | 527 | Get an item from the channel queue without waiting. Raises asyncio.QueueEmpty if the queue is empty. Args: | STRING |
| HIGH | rlinf/scheduler/worker/routing.py | 269 | Return worker group metadata from the manager proxy. Args: manager_proxy: Worker manager proxy used to quer | STRING |
| HIGH | rlinf/scheduler/worker/lock.py | 85 | Lock a network port for the current worker. This is useful for preventing port conflicts when multiple workers | STRING |
| HIGH | rlinf/agents/wideseek_r1/tools.py | 127 | Execute a single Serper API search request (low-level network call). Args: session: aiohtt | STRING |
| HIGH | rlinf/agents/wideseek_r1/tools.py | 250 | Execute a single Jina API access request (low-level network call). Args: session: aiohttp | STRING |
| HIGH | rlinf/utils/data_iter_utils.py | 68 | Split a list into equal sized chunks or fixed size chunks Args: inputs: Input list to split num | STRING |
| HIGH | rlinf/models/embodiment/starvla/starvla_action_model.py | 126 | Dispatch forward passes for RLinf. Args: forward_type: Which forward path to run. **kwa | STRING |
| HIGH | rlinf/models/embodiment/starvla/starvla_action_model.py | 150 | Run training-time forward for PPO terms (logprob/entropy/value). This method delegates to an action-head-specif | STRING |
| HIGH | rlinf/models/embodiment/starvla/__init__.py | 39 | Load a starVLA checkpoint and wrap it into RLinf's embodied policy interface. Args: cfg: Model config. Must | STRING |
| HIGH | rlinf/models/embodiment/reward/resnet_reward_model.py | 103 | Preprocess images for ResNet backbone input. This method accepts image tensors in ``uint8`` or floating-point f | STRING |
| HIGH | rlinf/envs/behavior/instance_loader.py | 45 | Parse definition and instance ids from a cached instance filename. Args: filename: Candidate cached instanc | STRING |
| HIGH | rlinf/envs/behavior/instance_loader.py | 95 | Discover cached instance files for one BEHAVIOR activity. Args: activity_instance_dir: Directory containing | STRING |
| HIGH | rlinf/envs/behavior/instance_loader.py | 257 | Build an instance loader from OmniGibson task config. Args: omni_cfg: Full OmniGibson config used t | STRING |
| HIGH | rlinf/envs/behavior/instance_generator.py | 107 | Load an RLinf BEHAVIOR env yaml. Args: config_path: Path to the yaml file. Returns: The loaded | STRING |
| HIGH | rlinf/envs/behavior/instance_generator.py | 161 | Resolve the directory used to save cached instances. Args: omni_cfg: OmniGibson config used for sampling. | STRING |
| HIGH | rlinf/envs/behavior/instance_generator.py | 208 | Build the output file path for the current task instance. Args: task: Active OmniGibson BEHAVIOR task. | STRING |
| HIGH | rlinf/envs/behavior/instance_generator.py | 325 | Save the current sampled instance in the requested format. Args: env: Active OmniGibson environment. | STRING |
| HIGH | rlinf/envs/calvin/calvin_gym_env.py | 157 | By constructing a 2D seed matrix (self.reset_state_ids_all) and applying a seed_offset, users can evenl | STRING |
| HIGH | rlinf/envs/realworld/franka/end_effectors/__init__.py | 29 | Factory function to create an end-effector instance. Args: end_effector_type: The type of end-effector to c | STRING |
| HIGH | rlinf/hybrid_engines/weight_syncer/patch_syncer.py | 60 | View a tensor as a 2-D COO indexing target. Scalars are viewed as shape ``(1, 1)``, vectors as ``(1, N)``, matrices | STRING |
| HIGH | rlinf/data/storage/lerobot/compat.py | 93 | Return each episode's ``[from, to)`` frame range in the flat frame index. Dataset format v2.1 exposes ``LeRobotData | STRING |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | rlinf/config.py | 734 | f"grouped_gemm type only avail in [null, te]. get value ({cfg.model.moe_grouped_gemm})" | CODE |
| HIGH⚡ | rlinf/algorithms/rewards/math/math_verifier/verify.py | 112 | if a.equals(b) or simplify(a - b) == 0: | CODE |
| HIGH⚡ | rlinf/algorithms/rewards/math/math_verifier/verify.py | 136 | if _a.equals(_b): | CODE |
| HIGH | rlinf/scheduler/dynamic_scheduler/scheduler_worker.py | 109 | resource_info = f"[Release && Allocate Info] After train-iter{train_iter}\n" | CODE |
| HIGH | …dels/embodiment/openpi/dataconfig/libero_dataconfig.py | 87 | data_transforms = data_transforms.push( | CODE |
| HIGH | …dels/embodiment/openpi/dataconfig/calvin_dataconfig.py | 56 | data_transforms = data_transforms.push( | CODE |
| HIGH | …dels/embodiment/openpi/dataconfig/franka_dataconfig.py | 87 | data_transforms = data_transforms.push( | CODE |
| HIGH | …s/embodiment/openpi/dataconfig/metaworld_dataconfig.py | 55 | data_transforms = data_transforms.push( | CODE |
| HIGH | …bodiment/openpi/dataconfig/maniskill_rlt_dataconfig.py | 78 | data_transforms = data_transforms.push( | CODE |
| HIGH | …els/embodiment/openpi/dataconfig/polaris_dataconfig.py | 117 | data_transforms = data_transforms.push( | CODE |
| HIGH | …/openpi/dataconfig/dual_franka_tcp_rot6d_dataconfig.py | 71 | data_transforms = data_transforms.push( | CODE |
| HIGH | …ent/openpi/dataconfig/franka_co_training_dataconfig.py | 88 | data_transforms = data_transforms.push( | CODE |
| HIGH | …ls/embodiment/openpi/dataconfig/behavior_dataconfig.py | 96 | data_transforms = data_transforms.push( | CODE |
| HIGH | …odiment/openpi/dataconfig/robotwin_aloha_dataconfig.py | 89 | data_transforms = data_transforms.push( | CODE |
| HIGH | …s/embodiment/openpi/dataconfig/maniskill_dataconfig.py | 88 | data_transforms = data_transforms.push( | CODE |
| HIGH⚡ | rlinf/workers/rollout/sglang/sglang_agent_worker.py | 46 | """LiteLLM/OpenAI clients may omit `content` when null; older SGLang (e.g. 0.4.x) | STRING |
| HIGH⚡ | rlinf/workers/rollout/sglang/sglang_agent_worker.py | 47 | ChatCompletionMessageGenericParam requires the key (value may be null). 0.5.2+ | STRING |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | pyproject.toml | 35 | # dev8 rejects our version_base="1.1"; prerelease bound keeps agentic's dev1 pin valid | COMMENT |
| MEDIUM | README.md | 51 | - [2026/05] 🔥 RLinf supports AgentLightning for single-agent RL training. Doc: [AgentLightning Calc-X](https://rlinf.rea | CODE |
| MEDIUM | README.md | 73 | - [2026/03] 🔥 RLinf supports reinforcement learning on multiagents. Website: [WideSeek-R1](https://wideseek-r1.github.io | CODE |
| MEDIUM | README.md | 78 | - [2026/02] 🔥 RLinf supports agentic reinforcement learning on [rStar2](https://github.com/volcengine/verl/pull/3397). D | CODE |
| MEDIUM | README.md | 89 | - [2025/12] 🔥 RLinf supports agentic reinforcement learning on [Search-R1](https://github.com/PeterGriffinJin/Search-R1) | CODE |
| MEDIUM | README.md | 100 | - [2025/10] 🔥 RLinf now officially supports online reinforcement learning! Doc: [coding_online_rl](https://rlinf.readthe | CODE |
| MEDIUM⚡ | README.md | 297 | <a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/agentic/searchr1.html"> | CODE |
| MEDIUM⚡ | README.md | 302 | <a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/agentic/rstar2.html"> | CODE |
| MEDIUM⚡ | README.md | 307 | <a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/agentic/agentlightning_calc_x.html"> | CODE |
| MEDIUM⚡ | README.md | 312 | <a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/agentic/coding_online_rl.html"> | CODE |
| MEDIUM⚡ | README.md | 317 | <a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/agentic/reasoning.html"> | CODE |
| MEDIUM⚡ | README.md | 326 | <a href="https://rlinf.readthedocs.io/en/latest/rst_source/examples/agentic/wideseek_r1/index.html"> | CODE |
| MEDIUM | README.zh-CN.md | 50 | - [2026/05] 🔥 RLinf 支持 AgentLightning 单智能体强化学习训练。文档: [AgentLightning Calc-X](https://rlinf.readthedocs.io/zh-cn/latest/r | CODE |
| MEDIUM | README.zh-CN.md | 71 | - [2026/03] 🔥 RLinf 支持多智能体强化学习。网站： [WideSeek-R1](https://wideseek-r1.github.io)， 快速启动： [快速启动](https://rlinf.readthedocs. | CODE |
| MEDIUM | README.zh-CN.md | 76 | - [2026/02] 🔥 RLinf支持[rStar2](https://github.com/volcengine/verl/pull/3397)的强化学习微调。 文档: [rStar2](https://rlinf.readthedo | CODE |
| MEDIUM | README.zh-CN.md | 87 | - [2025/12] 🔥 RLinf支持[Search-R1](https://github.com/PeterGriffinJin/Search-R1)的强化学习微调，相比原版实现加速 55%！ 文档: [Search-R1](http | CODE |
| MEDIUM | README.zh-CN.md | 98 | - [2025/10] 🔥 RLinf 正式支持在线强化学习！文档：[coding_online_rl](https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/agent | CODE |
| MEDIUM⚡ | README.zh-CN.md | 297 | <a href="https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/agentic/searchr1.html"> | CODE |
| MEDIUM⚡ | README.zh-CN.md | 302 | <a href="https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/agentic/rstar2.html"> | CODE |
| MEDIUM⚡ | README.zh-CN.md | 307 | <a href="https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/agentic/agentlightning_calc_x.html"> | CODE |
| MEDIUM⚡ | README.zh-CN.md | 312 | <a href="https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/agentic/coding_online_rl.html"> | CODE |
| MEDIUM⚡ | README.zh-CN.md | 317 | <a href="https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/agentic/reasoning.html"> | CODE |
| MEDIUM⚡ | README.zh-CN.md | 326 | <a href="https://rlinf.readthedocs.io/zh-cn/latest/rst_source/examples/agentic/wideseek_r1/index.html"> | CODE |
| MEDIUM | .cursor/skills/create-pr/lint_pr.py | 744 | # Leftover template scaffolding. | COMMENT |
| MEDIUM | .claude/skills/create-pr/lint_pr.py | 744 | # Leftover template scaffolding. | COMMENT |
| MEDIUM | requirements/agentic/sglang_0.4.6.post5_cu12.txt | 2 | # after the agentic-sglang extra; see install_agentic. | COMMENT |
| MEDIUM | .codex/skills/create-pr/lint_pr.py | 744 | # Leftover template scaffolding. | COMMENT |
| MEDIUM | rlinf/algorithms/rewards/math/math_verifier/parser.py | 410 | # Handle Chinese few-shot multiple choice problem answer extraction | COMMENT |
| MEDIUM | …es/agent/wideseek_r1/config/eval_qwen3_widesearch.yaml | 22 | workflow: mas # or sa, mas refers to multi-agent, sa refers to single-agent | CODE |
| MEDIUM | examples/agent/wideseek_r1/config/eval_qwen3_qa.yaml | 21 | workflow: mas # or sa, mas refers to multi-agent, sa refers to single-agent | CODE |
| MEDIUM | …eseek_r1/config/eval_qwen3_widesearch_local_judge.yaml | 23 | workflow: mas # or sa, mas refers to multi-agent, sa refers to single-agent | CODE |
| MEDIUM | …wideseek_r1/config/train_qwen3_hybrid_local_judge.yaml | 34 | workflow: mas # or sa, mas refers to multi-agent, sa refers to single-agent | CODE |
| MEDIUM | …mples/agent/wideseek_r1/config/train_qwen3_hybrid.yaml | 33 | workflow: mas # or sa, mas refers to multi-agent, sa refers to single-agent | CODE |
| MEDIUM | …mples/agent/wideseek_r1/config/eval_qwen3_qa_2eng.yaml | 21 | workflow: mas # or sa, mas refers to multi-agent, sa refers to single-agent | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | tests/unit_tests/test_math_dataset_multithread.py | 114 | # Create a temporary JSON file with sample data | COMMENT |
| MEDIUM | tests/unit_tests/test_math_dataset_multithread.py | 184 | # Create a temporary JSON file with sample data | COMMENT |
| MEDIUM | tests/unit_tests/test_channel.py | 792 | # Create a new channel with maxsize for this specific test | COMMENT |
| MEDIUM | tests/unit_tests/test_auto_placement.py | 262 | # Create a mock config | COMMENT |
| MEDIUM | tests/unit_tests/test_auto_placement.py | 295 | # Create a mock config | COMMENT |
| MEDIUM⚡ | tests/parity_tests/analyze_logs.py | 135 | # Create a directory for individual plots | COMMENT |
| MEDIUM | toolkits/auto_placement/workflow.py | 172 | # Create a frozenset of (node, tuple of neighbors) pairs to represent the graph | COMMENT |
| MEDIUM | rlinf/scheduler/collective/multi_channel_pg.py | 640 | # Create the global rank to group rank mapping | COMMENT |
| MEDIUM | rlinf/scheduler/collective/collective_group.py | 1236 | # Create the group if I'm the master worker | COMMENT |
| MEDIUM | rlinf/scheduler/manager/worker_manager.py | 296 | # Create a new root node if the parent is not found | COMMENT |
| MEDIUM | rlinf/scheduler/dynamic_scheduler/utils.py | 27 | """This function is used to get the valid data parallel sizes for the Actor based on the constraints of batch and gr | STRING |
| MEDIUM | rlinf/agents/wideseek_r1/wideseek_r1.py | 413 | # Initialize tracking variables | COMMENT |
| MEDIUM | rlinf/agents/wideseek_r1/utils/webpage.py | 15 | # This file contains code copied from the Asearcher project: https://github.com/inclusionAI/ASearcher | COMMENT |
| MEDIUM | rlinf/utils/initialize.py | 328 | # Create a MPI process group to help with TP communication overlap bootstrap. | COMMENT |
| MEDIUM | rlinf/utils/metric_utils.py | 706 | # Create the line with exactly 3 columns | COMMENT |
| MEDIUM | rlinf/utils/distributed.py | 839 | """This function is meant to wrap an arbitary function to only call the function | STRING |
| MEDIUM | …inf/models/embodiment/openvla_oft/official/__init__.py | 63 | # Create a new OpenVLAOFTRLConfig from the loaded config | COMMENT |
| MEDIUM | rlinf/models/embodiment/dreamzero/dreamzero_policy.py | 61 | # This method is called in FSDPModelManager.setup_model_and_optimizer | COMMENT |
| MEDIUM | …dels/embodiment/gr00t/gr00t_n1d5/gr00t_action_model.py | 585 | # Create a copy to avoid mutating input | COMMENT |
| MEDIUM | rlinf/runners/reasoning_runner.py | 106 | # Create a local channel (i.e., a channel that is different in every process) | COMMENT |
| MEDIUM | rlinf/runners/reasoning_eval_runner.py | 62 | # Create a local channel (i.e., a channel that is different in every process) | COMMENT |
| MEDIUM | rlinf/runners/coding_online_rl_runner.py | 68 | # Create a local channel (i.e., a channel that is different in every process) | COMMENT |
| MEDIUM | rlinf/envs/habitat/extensions/maps.py | 15 | # This file contains code copied from the VLN-CE project: https://github.com/jacobkrantz/VLN-CE. | COMMENT |
| MEDIUM | rlinf/envs/habitat/extensions/utils.py | 15 | # This file contains code copied from the VLN-CE project: https://github.com/jacobkrantz/VLN-CE. | COMMENT |
| MEDIUM | …f/envs/realworld/franka/tasks/franka_bin_relocation.py | 132 | # Define the parameterized line segment | COMMENT |
| MEDIUM | rlinf/data/datasets/vlm/robo2vlm.py | 264 | # Create a formatted question with choices | COMMENT |
| MEDIUM | rlinf/data/datasets/vlm/robo2vlm.py | 269 | # Create the answer using the correct choice (matching finetune example format) | COMMENT |
| MEDIUM | …/agent/tools/search_local_server_qdrant/build_index.py | 192 | # Create a process pool with the specified number of parallel workers | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | tests/parity_tests/parse_success_once.py | 286 | # Print results | COMMENT |
| LOW | tests/parity_tests/run_all.sh | 181 | # Check if any log for this experiment exists and determine status | COMMENT |
| LOW | tests/parity_tests/run_all.sh | 419 | # Check if received signal changed mid-task (though probability is low) | COMMENT |
| LOW | tests/parity_tests/plot_success_once.py | 259 | # Print results | COMMENT |
| LOW⚡ | tests/parity_tests/analyze_logs.py | 131 | # Check if baseline directory exists for comparison plots | COMMENT |
| LOW | tests/e2e_tests/embodied/genesis_cubepick_ppo_cnn.yaml | 118 | model_path: "RLinf-ResNet10-pretrained" # Set this to the directory containing resnet10_pretrained.pt | CODE |
| LOW | toolkits/standalone_eval_scripts/dexbotic/__init__.py | 86 | # Set HF_HUB_OFFLINE to prevent any network access during model loading | COMMENT |
| LOW | toolkits/standalone_eval_scripts/openpi/__init__.py | 82 | # Check if this is a PyTorch model by looking for model.safetensors | COMMENT |
| LOW | requirements/install.sh | 990 | # Check if pip is available | COMMENT |
| LOW | requirements/install.sh | 2318 | # Check if ROS_CATKIN_PATH is set or serl_franka_controllers is already built | COMMENT |
| LOW | requirements/embodied/ros_install.sh | 6 | # Check if apt is available | COMMENT |
| LOW | requirements/embodied/ros_install.sh | 14 | # Check if already running as root | COMMENT |
| LOW | requirements/embodied/ros_install.sh | 63 | # Check if the source already exists anywhere under /etc/apt | COMMENT |
| LOW | ray_utils/check_ray.sh | 3 | # Check if the required GPU count parameter is provided | COMMENT |
| LOW | ray_utils/check_ray.sh | 15 | # Check if the ray status command executed successfully | COMMENT |
| LOW | ray_utils/check_ray.sh | 29 | # Check if the required GPU count has been reached | COMMENT |
| LOW | rlinf/scheduler/collective/multi_channel_pg.py | 68 | # Check if all workers have the same accelerator type | COMMENT |
| LOW | rlinf/scheduler/collective/collective.py | 60 | # Check if the group already exists in the global collective manager | COMMENT |
| LOW⚡ | rlinf/scheduler/collective/collective_group.py | 1463 | # Check if the peer is on the same node | COMMENT |
| LOW⚡ | rlinf/scheduler/collective/collective_group.py | 1470 | # Check if the two device list has intersection | COMMENT |
| LOW | rlinf/scheduler/channel/channel.py | 192 | # Set max_concurrency to a high value to avoid large number of gets blocking puts | COMMENT |
| LOW | rlinf/scheduler/placement/placement.py | 369 | # Check if it's comma-separated | COMMENT |
| LOW | rlinf/agents/rstar2/http_code_judge_tool.py | 151 | # Check if the last node is a pure expression (not a call) | COMMENT |
| LOW | rlinf/agents/wideseek_r1/utils/reward.py | 328 | # Check if response has required columns | COMMENT |
| LOW | rlinf/utils/data_iter_utils.py | 599 | # Check if attention_mask exists and calculate sequence lengths | COMMENT |
| LOW | rlinf/utils/placement.py | 713 | # Check if actor and reward are on the same GPUs (all GPUs) | COMMENT |
| LOW⚡ | rlinf/utils/placement.py | 727 | # Check if all rollout components' GPUs are subsets of actor's GPUs | COMMENT |
| LOW⚡ | rlinf/utils/placement.py | 736 | # Check if rollout components cover all actor's GPUs | COMMENT |
| LOW | …f/utils/ckpt_convertor/convert_openpi_jax_to_python.py | 371 | # Check if we have Dense layers (for pi05/adaptive normalization) or scale layers (for regular pi0) | COMMENT |
| LOW | rlinf/models/embodiment/dreamzero/dreamzero_policy.py | 150 | # Check if relative_action is enabled and convert relative to absolute | COMMENT |
| LOW | rlinf/models/embodiment/dreamzero/dreamzero_policy.py | 183 | # Check if the state data shape matches the action shape | COMMENT |
| LOW | rlinf/models/embodiment/openpi/__init__.py | 54 | # Check if this is a checkpoint directory (saved by FSDP) | COMMENT |
| LOW | rlinf/models/embodiment/openpi/openpi_action_model.py | 287 | # Set _fsdp_wrap_name to the last part of the path (e.g., "model.action_in_proj" -> "action_in_proj") | COMMENT |
| LOW | rlinf/models/embodiment/openpi/openpi_action_model.py | 1722 | # Check if in [-1, 1] range | COMMENT |
| LOW | rlinf/workers/rollout/server/server_rollout_worker.py | 92 | # Check if we need a new file | COMMENT |
| LOW | rlinf/workers/rollout/server/online_router_worker.py | 134 | # Check if sync is in progress | COMMENT |
| LOW | rlinf/workers/rollout/server/online_router_worker.py | 241 | # Set event to indicate old requests are complete | COMMENT |
| LOW | rlinf/workers/actor/async_fsdp_sac_policy_worker.py | 99 | # Check if replay buffer has enough samples | COMMENT |
| LOW | rlinf/workers/actor/fsdp_sac_policy_worker.py | 709 | # Check if replay buffer has enough samples | COMMENT |
| LOW | rlinf/envs/world_model/world_model_wan_env.py | 222 | # Check if any reward in the chunk exceeds the threshold | COMMENT |
| LOW | rlinf/envs/world_model/world_model_opensora_env.py | 382 | # Check if any reward in the chunk exceeds the threshold | COMMENT |
| LOW | rlinf/envs/realworld/realworld_env.py | 96 | # Check if the path is valid | COMMENT |
| LOW | rlinf/envs/realworld/franka/franka_env.py | 414 | # Check if current state meets the success threshold | COMMENT |
| LOW | rlinf/envs/realworld/common/ros/ros_controller.py | 40 | # Check if the path is valid | COMMENT |
| LOW | rlinf/hybrid_engines/megatron/megatron_model_manager.py | 95 | # Check if FUSCO is available | COMMENT |
| LOW | rlinf/data/storage/replay/dataset.py | 242 | # Check if thread died with exception | COMMENT |
| LOW | …_labeling/recap/process/visualize_advantage_dataset.py | 120 | # Check if it looks like an image (3D or 4D with channel dim) | COMMENT |
| LOW | …/agent/tools/search_local_server_qdrant/build_index.py | 96 | # Check if collection exists, if not, build it from corpus | COMMENT |
| LOW | examples/sft/config/behavior_pi05_vla.yaml | 32 | # Set this to your local BEHAVIOR streaming dataset folder (the | COMMENT |
| LOW | …amples/embodiment/config/genesis_cubepick_ppo_cnn.yaml | 120 | model_path: "RLinf-ResNet10-pretrained" # Set this to the directory containing resnet10_pretrained.pt | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | tests/parity_tests/parse_success_once.py | 147 | CODE | |
| LOW | tests/parity_tests/analyze_logs.py | 60 | CODE | |
| LOW | rlinf/algorithms/utils.py | 280 | CODE | |
| LOW | rlinf/algorithms/losses.py | 27 | CODE | |
| LOW | rlinf/algorithms/losses.py | 170 | CODE | |
| LOW | rlinf/scheduler/collective/multi_channel_pg.py | 499 | CODE | |
| LOW | rlinf/scheduler/worker/worker.py | 849 | CODE | |
| LOW | rlinf/scheduler/worker/worker.py | 1040 | CODE | |
| LOW | rlinf/agents/wideseek_r1/wideseek_r1.py | 776 | CODE | |
| LOW | rlinf/utils/distributed.py | 866 | CODE | |
| LOW | rlinf/utils/timers.py | 196 | CODE | |
| LOW | rlinf/utils/profiler.py | 47 | CODE | |
| LOW | …inf/utils/ckpt_convertor/openpi/jax_to_openpi_rlinf.py | 275 | CODE | |
| LOW | …iment/openvla_oft/official/openvla_oft_action_model.py | 46 | CODE | |
| LOW | …iment/openvla_oft/official/openvla_oft_action_model.py | 705 | CODE | |
| LOW | …iment/openvla_oft/official/openvla_oft_action_model.py | 581 | CODE | |
| LOW | …bodiment/openvla_oft/rlinf/openvla_oft_action_model.py | 576 | CODE | |
| LOW | …bodiment/openvla_oft/rlinf/openvla_oft_action_model.py | 466 | CODE | |
| LOW | rlinf/models/embodiment/evo1/evo1_action_model.py | 62 | CODE | |
| LOW | …f/models/embodiment/value_model/recap/configuration.py | 164 | CODE | |
| LOW | …models/embodiment/value_model/recap/modeling_critic.py | 592 | CODE | |
| LOW | …f/models/embodiment/value_model/steam/configuration.py | 71 | CODE | |
| LOW | …models/embodiment/value_model/steam/modeling_critic.py | 491 | CODE | |
| LOW | rlinf/models/embodiment/openvla/openvla_action_model.py | 61 | CODE | |
| LOW | rlinf/models/embodiment/openvla/openvla_action_model.py | 608 | CODE | |
| LOW | rlinf/models/embodiment/openvla/openvla_action_model.py | 517 | CODE | |
| LOW | rlinf/models/embodiment/modules/flow_actor.py | 28 | CODE | |
| LOW | rlinf/models/embodiment/modules/flow_actor.py | 230 | CODE | |
| LOW | …inf/models/embodiment/modules/rlt_token_transformer.py | 101 | CODE | |
| LOW | …inf/models/embodiment/modules/rlt_token_transformer.py | 190 | CODE | |
| LOW | …inf/models/embodiment/modules/rlt_token_transformer.py | 271 | CODE | |
| LOW | rlinf/models/embodiment/modules/resnet_utils.py | 95 | CODE | |
| LOW | rlinf/models/embodiment/modules/mlp.py | 34 | CODE | |
| LOW | rlinf/models/embodiment/modules/explore_noise_net.py | 103 | CODE | |
| LOW | rlinf/workers/agent/agent_loop.py | 155 | CODE | |
| LOW | rlinf/workers/actor/fsdp_iql_policy_worker.py | 670 | CODE | |
| LOW | rlinf/envs/wrappers/collect_episode.py | 108 | CODE | |
| LOW | rlinf/envs/habitat/extensions/utils.py | 383 | CODE | |
| LOW | rlinf/envs/robocasa365/robocasa365_env.py | 337 | CODE | |
| LOW | rlinf/envs/realworld/franka/franka_controller.py | 39 | CODE | |
| LOW | rlinf/envs/realworld/common/pico/pico_expert.py | 93 | CODE | |
| LOW | rlinf/hybrid_engines/sglang/common/sgl_scheduler.py | 271 | CODE | |
| LOW | rlinf/hybrid_engines/megatron/megatron_model_manager.py | 623 | CODE | |
| LOW | rlinf/data/datasets/world_model.py | 111 | CODE | |
| LOW | rlinf/data/datasets/recap/value_dataset.py | 204 | CODE | |
| LOW | rlinf/data/datasets/steam/pair_dataset.py | 630 | CODE | |
| LOW | rlinf/data/datasets/dagger/dataset.py | 107 | CODE | |
| LOW | rlinf/data/datasets/dagger/dataloader.py | 144 | CODE | |
| LOW | …datasets/openpi_rlinf/behavior/behavior_sft_dataset.py | 596 | CODE | |
| LOW | rlinf/data/datasets/reasoning/dataset.py | 254 | CODE | |
| LOW | rlinf/data/datasets/reasoning/wideseek_r1.py | 139 | CODE | |
| LOW | rlinf/data/datasets/dreamzero/lerobot_dataset.py | 67 | CODE | |
| LOW | rlinf/data/storage/lerobot/writer.py | 59 | CODE | |
| LOW | …rl/advantage_labeling/recap/process/compute_returns.py | 195 | CODE | |
| LOW | …ls/search_local_server_faiss/local_retrieval_server.py | 356 | CODE | |
| LOW | …/agent/tools/search_local_server_qdrant/build_index.py | 278 | CODE | |
| LOW | …s/search_local_server_qdrant/local_retrieval_server.py | 266 | CODE | |
| LOW | …les/agent/rstar2/data_process/process_train_dataset.py | 51 | CODE | |
| LOW | examples/reward/preprocess_vlm_trend_reward_dataset.py | 166 | CODE | |
| LOW | examples/reward/preprocess_vlm_trend_reward_dataset.py | 417 | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW⚡ | tests/parity_tests/analyze_logs.py | 102 | # Step 1: Parse logs | COMMENT |
| LOW⚡ | tests/parity_tests/analyze_logs.py | 122 | # Step 2: Plot curves (each experiment vs baseline in separate figures) | COMMENT |
| LOW | tests/parity_tests/analyze_logs.py | 166 | # Step 3: Compare with baseline | COMMENT |
| LOW | …olkits/standalone_eval_scripts/dexbotic/libero_eval.py | 138 | # IMPORTANT: Do nothing for the first few timesteps because the simulator drops objects | COMMENT |
| LOW | toolkits/standalone_eval_scripts/openpi/libero_eval.py | 138 | # IMPORTANT: Do nothing for the first few timesteps because the simulator drops objects | COMMENT |
| LOW | docs/README.md | 9 | ### Step 1: Set Environment Variables | COMMENT |
| LOW | docs/README.md | 22 | ### Step 2: Install Dependencies | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 195 | # Step 1: Average rewards per trajectory. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 198 | # Step 2: reshape to [num_questions, group_size] for per-question GRPO. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 201 | # Step 3: compute per-question mean and std. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 209 | # Step 4: normalize within each question group. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 214 | # Step 5: flatten back to [num_trajectories]. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 217 | # Step 6: broadcast trajectory advantages to all turns in that trajectory. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 222 | # Step 1: map each turn to its owning question. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 229 | # Step 2: normalize turn rewards within each question group. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 234 | # Step 3: compute mean and std for all turns in this question. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 238 | # Step 4: normalize turn rewards within the question. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 243 | # Step 5: write normalized turn-level advantages back. | COMMENT |
| LOW⚡ | rlinf/models/embodiment/openpi/openpi_action_model.py | 863 | # Step 1: SAC agent outputs noise | COMMENT |
| LOW⚡ | rlinf/models/embodiment/openpi/openpi_action_model.py | 870 | # Step 2: Use noise to sample actual actions from diffusion model | COMMENT |
| LOW⚡ | rlinf/models/embodiment/openpi/openpi_action_model.py | 878 | # Step 3: Extract actual actions for environment interaction | COMMENT |
| LOW | …_labeling/recap/process/visualize_advantage_dataset.py | 888 | # Step 1: Create distribution plot | COMMENT |
| LOW | …_labeling/recap/process/visualize_advantage_dataset.py | 899 | # Step 2: Create episode visualizations | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | .cursor/skills/test-install/driver.py | 161 | # Line-based scan: robust to Hydra structure, catches `key: /abs/path`. | COMMENT |
| MEDIUM | .cursor/skills/install-check/check.sh | 5 | # This is the harness for the `install-check` skill. It statically scans | COMMENT |
| MEDIUM | tests/parity_tests/run_all.sh | 344 | # Run comprehensive log analysis using analyze_logs.py | COMMENT |
| MEDIUM | .claude/skills/test-install/driver.py | 161 | # Line-based scan: robust to Hydra structure, catches `key: /abs/path`. | COMMENT |
| MEDIUM | .claude/skills/install-check/check.sh | 5 | # This is the harness for the `install-check` skill. It statically scans | COMMENT |
| MEDIUM | .codex/skills/test-install/driver.py | 161 | # Line-based scan: robust to Hydra structure, catches `key: /abs/path`. | COMMENT |
| MEDIUM | .codex/skills/install-check/check.sh | 5 | # This is the harness for the `install-check` skill. It statically scans | COMMENT |
| LOW | rlinf/scheduler/collective/multi_channel_pg.py | 603 | # which may just pass their timeout value (or None) | COMMENT |
| LOW | rlinf/scheduler/channel/channel.py | 170 | # Local channel does not need to be launched, just create a local channel object | COMMENT |
| MEDIUM | rlinf/models/embodiment/mlp_policy/iql_mlp_policy.py | 145 | # Keep shape aligned with generic MLPPolicy for seamless | COMMENT |
| LOW | rlinf/workers/sft/fsdp_vlm_sft_worker.py | 84 | # set the padding side to left for the tokenizer, QWEN 2.5 VL just use left padding | COMMENT |
| MEDIUM | rlinf/runners/reasoning_runner.py | 164 | # Use a sampler to facilitate checkpoint resumption. | COMMENT |
| MEDIUM⚡ | rlinf/envs/frankasim/frankasim_env.py | 27 | # Obs utils: object scalar unwrap + robust flatten | COMMENT |
| MEDIUM | rlinf/envs/genesis/tasks/cube_pick.py | 302 | """Compute reward and robust success signals for one step.""" | STRING |
| LOW | rlinf/hybrid_engines/weight_syncer/patch_syncer.py | 447 | # needs participate in the all-gather, but we just create | COMMENT |
| LOW | rlinf/hybrid_engines/weight_syncer/patch_syncer.py | 655 | # needs participate in the all-gather, but we just create | COMMENT |
| MEDIUM | …_labeling/recap/process/visualize_advantage_dataset.py | 152 | """Create comprehensive advantage distribution plots from meta/advantages.parquet.""" | STRING |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW⚡ | tests/parity_tests/analyze_logs.py | 102 | # Step 1: Parse logs | COMMENT |
| LOW⚡ | tests/parity_tests/analyze_logs.py | 122 | # Step 2: Plot curves (each experiment vs baseline in separate figures) | COMMENT |
| LOW | tests/parity_tests/analyze_logs.py | 166 | # Step 3: Compare with baseline | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 195 | # Step 1: Average rewards per trajectory. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 198 | # Step 2: reshape to [num_questions, group_size] for per-question GRPO. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 201 | # Step 3: compute per-question mean and std. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 209 | # Step 4: normalize within each question group. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 214 | # Step 5: flatten back to [num_trajectories]. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 217 | # Step 6: broadcast trajectory advantages to all turns in that trajectory. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 222 | # Step 1: map each turn to its owning question. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 229 | # Step 2: normalize turn rewards within each question group. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 234 | # Step 3: compute mean and std for all turns in this question. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 238 | # Step 4: normalize turn rewards within the question. | COMMENT |
| LOW⚡ | rlinf/algorithms/advantages.py | 243 | # Step 5: write normalized turn-level advantages back. | COMMENT |
| LOW⚡ | rlinf/models/embodiment/openpi/openpi_action_model.py | 863 | # Step 1: SAC agent outputs noise | COMMENT |
| LOW⚡ | rlinf/models/embodiment/openpi/openpi_action_model.py | 870 | # Step 2: Use noise to sample actual actions from diffusion model | COMMENT |
| LOW⚡ | rlinf/models/embodiment/openpi/openpi_action_model.py | 878 | # Step 3: Extract actual actions for environment interaction | COMMENT |
| LOW | …_labeling/recap/process/visualize_advantage_dataset.py | 888 | # Step 1: Create distribution plot | COMMENT |
| LOW | …_labeling/recap/process/visualize_advantage_dataset.py | 899 | # Step 2: Create episode visualizations | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | tests/e2e_tests/agent/rstar2/qwen2.5-1.5b-megatron.yaml | 124 | # demo specific parameters. not used in this demo. you can add your own parameters here. | COMMENT |
| MEDIUM | …ls/embodiment/openpi/dataconfig/robocasa_dataconfig.py | 53 | # these keys are features in LerobotDataset, so you can use it to make SFT right. | COMMENT |
| MEDIUM | …/agent/rstar2/config/rstar2-qwen2.5-1.5b-megatron.yaml | 124 | # demo specific parameters. not used in this demo. you can add your own parameters here. | COMMENT |
| MEDIUM | …es/agent/rstar2/config/rstar2-qwen2.5-7b-megatron.yaml | 124 | # demo specific parameters. not used in this demo. you can add your own parameters here. | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | …fline_rl/config/steam_compute_advantages_ensemble.yaml | 1 | # Usage: | COMMENT |
| LOW | …amples/offline_rl/config/recap_compute_advantages.yaml | 1 | # Usage: | COMMENT |
| LOW | examples/offline_rl/config/recap_compute_returns.yaml | 1 | # Usage: | COMMENT |
| LOW | …dvantage_labeling/recap/process/run_compute_returns.sh | 4 | # Usage: | COMMENT |
| LOW | …ntage_labeling/recap/process/run_compute_advantages.sh | 4 | # Usage: | COMMENT |
| LOW | …eling/steam/process/run_compute_advantages_ensemble.sh | 9 | # Usage: | COMMENT |