Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen3.6, GPT-OSS, Llama, and more!
This report presents the forensic synthetic code analysis of OpenPipe/ART, a Python project with 10,470 GitHub stars. SynthScan v2.0 examined 161,486 lines of code across 555 source files, recording 2895 pattern matches distributed across 19 syntactic categories. The overall adjusted score of 19.8 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).
This chart maps the temporal evolution of the adjusted synthetic code score across successive scan runs. An upward trajectory indicates ongoing incorporation of AI-generated code or expanding LLM-assisted scaffolding; a stable or declining trajectory may reflect active human refactoring, code removal, or the adoption of stricter authorship policies. The dashed secondary line (right axis) independently tracks total raw pattern hit count, which can diverge from the normalised score when codebase size changes significantly between scans.
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 2895 distinct pattern matches across 19 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 | vllm_runtime/src/art_vllm_runtime/patches.py | 14 | def apply_vllm_runtime_patches() -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 38 | def _patch_rope_validation_ignore_keys() -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 76 | def subclass_chat_completion_request() -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 94 | def patch_listen_for_disconnect() -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 103 | async def patched_listen_for_disconnect(request: Any) -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 120 | def patch_tool_parser_manager() -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 148 | def _restore_nccl_unique_id_payload( | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 165 | def _normalize_nccl_comm_init_rank_unique_id(library: Any, unique_id: object) -> object: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 171 | def patch_nccl_unique_id_bootstrap() -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 255 | def patch_art_lora_delta_weight_update() -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 369 | def patch_routed_experts_prefix_cache_sidecar() -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 564 | def require_no_unmet_prefix_route_needs(self: Any, req_id: str) -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/dedicated_server.py | 37 | def _patch_art_runtime_routes() -> None: | CODE |
| LOW | …_runtime/src/art_vllm_runtime/gemma4_moe_lora_patch.py | 20 | def get_causal_expert_mapping( | CODE |
| LOW | …_runtime/src/art_vllm_runtime/gemma4_moe_lora_patch.py | 37 | def get_conditional_expert_mapping( | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/lora_delta.py | 56 | def _merged_delta_skips_experts(adapter_config: dict[str, Any]) -> bool: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/lora_delta.py | 61 | def _iter_lora_checkpoint_deltas( | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/dsv4_patches.py | 19 | def _drop_reload_shadow_attrs(layer: Any, names: Any) -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/dsv4_patches.py | 30 | def patch_layerwise_reload_shadow_attrs() -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/dsv4_patches.py | 317 | def _apply_lora_to_existing_linear_output( | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/dsv4_patches.py | 595 | def _is_active_lora_wrapped_linear(module: Any) -> bool: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/dsv4_patches.py | 971 | def attn_gemm_parallel_execute(self: Any, hidden_states: Any) -> tuple[Any, ...]: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/dsv4_patches.py | 1152 | def attn_gemm_parallel_execute(self: Any, hidden_states: Any) -> tuple[Any, ...]: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/dsv4_patches.py | 1525 | def patch_lora_linear_base_attr_proxy() -> None: | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/dsv4_patches.py | 1546 | def patch_marlin_lora_swiglu_limit() -> None: | CODE |
| LOW | tests/integration.py | 68 | def make_variable_override_source(variables: Dict[str, Any]) -> str: | CODE |
| LOW | tests/integration.py | 82 | def _override_variables_in_notebook(nb, variables: Dict[str, Any]) -> list[str]: | CODE |
| LOW | tests/unit/test_unsloth_metrics.py | 12 | def test_get_log_fn_routes_eval_metrics_to_val_namespace() -> None: | CODE |
| LOW | tests/unit/test_strip_logprobs.py | 15 | def test_strip_dict_with_logprobs(self): | CODE |
| LOW | tests/unit/test_strip_logprobs.py | 46 | def test_strip_list_with_logprobs(self): | CODE |
| LOW | tests/unit/test_strip_logprobs.py | 59 | def test_strip_tuple_with_logprobs(self): | CODE |
| LOW | tests/unit/test_strip_logprobs.py | 73 | def test_strip_object_with_logprobs(self): | CODE |
| LOW | tests/unit/test_strip_logprobs.py | 89 | def test_strip_mixed_nested_structure(self): | CODE |
| LOW⚡ | tests/unit/test_strip_logprobs.py | 109 | def test_strip_empty_structures(self): | CODE |
| LOW⚡ | tests/unit/test_strip_logprobs.py | 115 | def test_strip_none_and_primitives(self): | CODE |
| LOW⚡ | tests/unit/test_strip_logprobs.py | 123 | def test_no_logprobs_unchanged(self): | CODE |
| LOW | tests/unit/test_strip_logprobs.py | 149 | def test_deepcopy_failure_returns_original(self, caplog): | CODE |
| LOW | tests/unit/test_strip_logprobs.py | 175 | def test_deepcopy_failure_with_recursion_error(self, caplog): | CODE |
| LOW | tests/unit/test_tinker_renderers.py | 158 | def test_convert_openai_messages_to_renderer_format_stringifies_dict_arguments() -> ( | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 162 | def test_pack_estimator_matches_ternary_and_random_trees(max_depth: int) -> None: | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 181 | def test_shared_trainable_tokens_accumulate_independent_output_gradients() -> None: | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 206 | def test_planner_handles_vineppo_nested_shape_and_request_mix() -> None: | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 227 | def test_forward_micro_batches_preserves_nested_vineppo_groups( | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 260 | def test_adaptive_planner_materializes_only_final_large_candidate( | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 306 | def test_adaptive_planner_globally_falls_back_when_one_rank_cannot_estimate( | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 329 | def test_adaptive_planner_probes_new_heterogeneous_signatures( | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 359 | def test_adaptive_planner_grows_stable_window_to_largest_aligned_fit( | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 377 | def test_forward_micro_batches_shrinks_when_memory_budget_drops( | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 426 | def test_heterogeneous_slots_split_packing_without_losing_output_estimates( | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 461 | def test_forward_raises_before_expected_oom_with_actionable_context( | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 501 | def test_flatten_rejects_dicts_to_avoid_silent_top_level_shape_changes() -> None: | CODE |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 506 | def test_no_output_requests_do_not_pack_or_consume_compute_memory() -> None: | CODE |
| LOW | tests/unit/test_moe_routing_real_path.py | 44 | def test_align_choice_routes_to_tokenized_result_maps_vllm_routes() -> None: | CODE |
| LOW | tests/unit/test_moe_routing_real_path.py | 66 | def test_align_choice_routes_to_tokenized_result_uses_current_vllm_contract() -> None: | CODE |
| LOW | tests/unit/test_moe_routing_real_path.py | 99 | def test_align_choice_routes_to_tokenized_result_rejects_token_mismatch() -> None: | CODE |
| LOW | tests/unit/test_moe_routing_real_path.py | 152 | def test_pack_carries_routes_through_prefix_tree_splicing() -> None: | CODE |
| LOW | tests/unit/test_moe_routing_real_path.py | 194 | def test_prefix_tree_pack_keeps_trainable_duplicates_in_leaf_metadata() -> None: | CODE |
| LOW | tests/unit/test_moe_routing_real_path.py | 250 | def test_prefix_tree_pack_public_api_emits_nested_metadata() -> None: | CODE |
| LOW | tests/unit/test_moe_routing_real_path.py | 307 | def test_pack_infers_at_least_topk_experts_from_sparse_routes() -> None: | CODE |
| LOW | tests/unit/test_moe_routing_real_path.py | 329 | def test_build_replay_bundle_uses_packed_sequence_sample_calls() -> None: | CODE |
| 1754 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | vllm_runtime/src/art_vllm_runtime/__init__.py | 1 | CODE | |
| LOW | vllm_runtime/src/art_vllm_runtime/__init__.py | 1 | CODE | |
| LOW | vllm_runtime/src/art_vllm_runtime/__init__.py | 1 | CODE | |
| LOW | vllm_runtime/src/art_vllm_runtime/__init__.py | 1 | CODE | |
| LOW | vllm_runtime/src/art_vllm_runtime/__init__.py | 1 | CODE | |
| LOW | tests/unit/test_strip_logprobs.py | 3 | CODE | |
| LOW | tests/unit/test_strip_logprobs.py | 5 | CODE | |
| LOW | tests/unit/test_strip_logprobs.py | 7 | CODE | |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 1 | CODE | |
| LOW | tests/unit/test_trainer_rank_weird_shapes.py | 13 | CODE | |
| LOW | tests/unit/test_moe_routing_real_path.py | 1 | CODE | |
| LOW | tests/unit/test_prefix_tree_grad_parity.py | 1 | CODE | |
| LOW | tests/unit/test_multi_checkpoint_inference.py | 14 | CODE | |
| LOW | tests/unit/test_multi_checkpoint_inference.py | 15 | CODE | |
| LOW | tests/unit/test_multi_checkpoint_inference.py | 19 | CODE | |
| LOW | tests/unit/test_tau_bench_client.py | 1 | CODE | |
| LOW | tests/unit/test_vllm_lora_delta.py | 1 | CODE | |
| LOW | tests/unit/test_dsv4_vllm_runtime_patches.py | 1 | CODE | |
| LOW | tests/unit/test_prefix_tree_attention_builder.py | 1 | CODE | |
| LOW | tests/unit/test_moe_routing_replay.py | 1 | CODE | |
| LOW | tests/unit/test_trainer_rank_validation.py | 1 | CODE | |
| LOW | tests/unit/test_trajectory_copy.py | 9 | CODE | |
| LOW | tests/unit/test_prefix_tree_packing.py | 1 | CODE | |
| LOW | tests/unit/test_trajectory_parquet.py | 13 | CODE | |
| LOW | tests/unit/test_trajectory_parquet.py | 16 | CODE | |
| LOW | tests/unit/test_trajectory_parquet.py | 23 | CODE | |
| LOW | tests/unit/test_trajectory_parquet.py | 44 | CODE | |
| LOW | tests/unit/test_trajectory_parquet.py | 44 | CODE | |
| LOW | tests/unit/test_trajectory_parquet.py | 44 | CODE | |
| LOW | tests/unit/test_trajectory_parquet.py | 44 | CODE | |
| LOW | tests/unit/test_prefix_tree.py | 1 | CODE | |
| LOW | tests/unit/test_megatron_reference_logprobs.py | 1 | CODE | |
| LOW | tests/integration/test_multi_checkpoint_training.py | 18 | CODE | |
| LOW | tests/integration/test_lora_quack_cutover.py | 1 | CODE | |
| LOW | tests/integration/megatron/metrics.py | 1 | CODE | |
| LOW | tests/integration/megatron/metrics.py | 3 | CODE | |
| LOW | tests/integration/megatron/conftest.py | 6 | CODE | |
| LOW | tests/integration/megatron/artifacts.py | 9 | CODE | |
| LOW | tests/integration/megatron/routing_replay/bundle.py | 1 | CODE | |
| LOW | tests/integration/megatron/routing_replay/bundle.py | 3 | CODE | |
| LOW | tests/integration/megatron/routing_replay/trace.py | 1 | CODE | |
| LOW | …n/megatron/cp_attn/megatron_attention_oracle_worker.py | 1 | CODE | |
| LOW | …/megatron/cp_attn/megatron_attention_oracle_harness.py | 1 | CODE | |
| LOW | …/megatron/cp_attn/megatron_attention_oracle_harness.py | 6 | CODE | |
| LOW | …megatron/cp_attn/test_attention_packed_vs_flattened.py | 1 | CODE | |
| LOW | …megatron/cp_attn/test_attention_packed_vs_flattened.py | 14 | CODE | |
| LOW | …n/gdn_shared_prefix/test_qwen35_gdn_topology_oracle.py | 1 | CODE | |
| LOW | tests/integration/megatron/gdn_shared_prefix/oracles.py | 1 | CODE | |
| LOW | tests/integration/megatron/gdn_shared_prefix/metrics.py | 1 | CODE | |
| LOW | …on/gdn_shared_prefix/test_gdn_cp_packed_correctness.py | 1 | CODE | |
| LOW | …on/gdn_shared_prefix/test_gdn_cp_packed_correctness.py | 3 | CODE | |
| LOW | …tron/gdn_shared_prefix/test_fla_cp_native_recurrent.py | 1 | CODE | |
| LOW | …ntegration/megatron/gdn_shared_prefix/parser_import.py | 1 | CODE | |
| LOW | …on/megatron/gdn_shared_prefix/test_real_gdn_tp_lora.py | 1 | CODE | |
| LOW | …sts/integration/megatron/gdn_shared_prefix/__init__.py | 3 | CODE | |
| LOW | tests/integration/megatron/gdn_shared_prefix/cases.py | 1 | CODE | |
| LOW | …gration/megatron/gdn_shared_prefix/distributed_grad.py | 1 | CODE | |
| LOW | …egatron/gdn_shared_prefix/test_gdn_cp_train_prepare.py | 1 | CODE | |
| LOW | …atron/gdn_shared_prefix/test_real_gdn_native_fla_cp.py | 1 | CODE | |
| LOW | …ation/megatron/gdn_shared_prefix/test_gdn_conv_gelu.py | 1 | CODE | |
| 419 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | skypilot-config.yaml | 358 | # ---------------------------------------------------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 22 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 24 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 128 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 130 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 193 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 195 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 249 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 251 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 291 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 293 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 398 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | tests/unit/test_multi_checkpoint_inference.py | 400 | # ============================================================================= | COMMENT |
| MEDIUM | scripts/deploy-model.py | 15 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | scripts/deploy-model.py | 17 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | scripts/deploy-model.py | 45 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | scripts/deploy-model.py | 47 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | dev/sft/sft-warmup.py | 57 | # ======================================================================== | COMMENT |
| MEDIUM⚡ | dev/sft/sft-warmup.py | 59 | # ======================================================================== | COMMENT |
| MEDIUM⚡ | dev/sft/sft-warmup.py | 67 | # ======================================================================== | COMMENT |
| MEDIUM⚡ | dev/sft/sft-warmup.py | 69 | # ======================================================================== | COMMENT |
| MEDIUM⚡ | dev/sft/sft-warmup.py | 84 | # ======================================================================== | COMMENT |
| MEDIUM⚡ | dev/sft/sft-warmup.py | 86 | # ======================================================================== | COMMENT |
| MEDIUM⚡ | dev/sft/sft-warmup.py | 94 | # ======================================================================== | COMMENT |
| MEDIUM⚡ | dev/sft/sft-warmup.py | 96 | # ======================================================================== | COMMENT |
| MEDIUM | dev/sft/sft-warmup.py | 111 | # ======================================================================== | COMMENT |
| MEDIUM | dev/sft/sft-warmup.py | 113 | # ======================================================================== | COMMENT |
| MEDIUM | src/art/types.py | 63 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | src/art/types.py | 65 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | src/art/model.py | 433 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | src/art/model.py | 435 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | src/art/model.py | 1052 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | src/art/model.py | 1054 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | src/art/serverless/backend.py | 713 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | src/art/serverless/backend.py | 715 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | src/art/utils/trajectory_migration.py | 23 | # ============================================================================ | COMMENT |
| MEDIUM⚡ | src/art/utils/trajectory_migration.py | 25 | # ============================================================================ | COMMENT |
| MEDIUM⚡ | src/art/utils/trajectory_migration.py | 139 | # ============================================================================ | COMMENT |
| MEDIUM⚡ | src/art/utils/trajectory_migration.py | 141 | # ============================================================================ | COMMENT |
| MEDIUM | …benchmarking/charts/percentage_comparison_bar_chart.py | 81 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …benchmarking/charts/percentage_comparison_bar_chart.py | 83 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …benchmarking/charts/percentage_comparison_bar_chart.py | 108 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …benchmarking/charts/percentage_comparison_bar_chart.py | 110 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | …benchmarking/charts/percentage_comparison_bar_chart.py | 133 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | …benchmarking/charts/percentage_comparison_bar_chart.py | 135 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | …benchmarking/charts/percentage_comparison_bar_chart.py | 140 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | …benchmarking/charts/percentage_comparison_bar_chart.py | 142 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …benchmarking/charts/percentage_comparison_bar_chart.py | 204 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …benchmarking/charts/percentage_comparison_bar_chart.py | 206 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | …benchmarking/charts/percentage_comparison_bar_chart.py | 226 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | …benchmarking/charts/percentage_comparison_bar_chart.py | 228 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …benchmarking/charts/percentage_comparison_bar_chart.py | 243 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …benchmarking/charts/percentage_comparison_bar_chart.py | 245 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …t/utils/benchmarking/charts/training_progress_chart.py | 74 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …t/utils/benchmarking/charts/training_progress_chart.py | 76 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …t/utils/benchmarking/charts/training_progress_chart.py | 134 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …t/utils/benchmarking/charts/training_progress_chart.py | 136 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …t/utils/benchmarking/charts/training_progress_chart.py | 163 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …t/utils/benchmarking/charts/training_progress_chart.py | 167 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …t/utils/benchmarking/charts/training_progress_chart.py | 180 | # ------------------------------------------------------------------ | COMMENT |
| 31 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 94 | CODE | |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 369 | CODE | |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 103 | CODE | |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 441 | CODE | |
| LOW | vllm_runtime/src/art_vllm_runtime/lora_delta.py | 61 | CODE | |
| LOW | vllm_runtime/src/art_vllm_runtime/dsv4_patches.py | 76 | CODE | |
| LOW | vllm_runtime/src/art_vllm_runtime/dsv4_patches.py | 97 | CODE | |
| LOW | tests/integration.py | 82 | CODE | |
| LOW | tests/integration.py | 176 | CODE | |
| LOW | tests/integration.py | 231 | CODE | |
| LOW | tests/unit/test_frontend_logging.py | 954 | CODE | |
| LOW | tests/unit/test_prefix_tree_attention_builder.py | 564 | CODE | |
| LOW | tests/unit/test_trainer_rank_validation.py | 861 | CODE | |
| LOW | tests/unit/test_prefix_tree.py | 265 | CODE | |
| LOW | tests/unit/test_prefix_tree.py | 367 | CODE | |
| LOW | tests/unit/test_prefix_tree.py | 388 | CODE | |
| LOW | tests/unit/test_prefix_tree.py | 441 | CODE | |
| LOW | …n/megatron/cp_attn/megatron_attention_oracle_worker.py | 43 | CODE | |
| LOW | …atron/gdn_shared_prefix/test_real_gdn_native_fla_cp.py | 601 | CODE | |
| LOW | …shared_prefix/test_real_gdn_cp1_packed_vs_flattened.py | 50 | CODE | |
| LOW | …ntegration/megatron/gdn_shared_prefix/packed_layout.py | 30 | CODE | |
| LOW | …ntegration/megatron/gdn_shared_prefix/packed_layout.py | 212 | CODE | |
| LOW | …egration/megatron/gdn_shared_prefix/real_gdn_oracle.py | 86 | CODE | |
| LOW | …/runtime_isolation/test_live_megatron_backend_smoke.py | 508 | CODE | |
| LOW | …s/integration/megatron/train_inf_mismatch/real_path.py | 429 | CODE | |
| LOW | …s/integration/megatron/train_inf_mismatch/real_path.py | 1109 | CODE | |
| LOW | …tegration/megatron/train_inf_mismatch/output_parity.py | 504 | CODE | |
| LOW | …_inf_mismatch/vllm_forward_trace_site/sitecustomize.py | 454 | CODE | |
| LOW | …ts/integration/megatron/model_support/forward_trace.py | 1744 | CODE | |
| LOW | …ts/integration/megatron/model_support/forward_trace.py | 2023 | CODE | |
| LOW | …ration/megatron/model_support/prefix_tree_workloads.py | 8 | CODE | |
| LOW | …ts/integration/megatron/model_support/oracle_worker.py | 78 | CODE | |
| LOW | …ts/integration/megatron/model_support/oracle_worker.py | 232 | CODE | |
| LOW | …ts/integration/megatron/model_support/oracle_worker.py | 260 | CODE | |
| LOW | …ts/integration/megatron/model_support/oracle_worker.py | 847 | CODE | |
| LOW | …ts/integration/megatron/model_support/oracle_worker.py | 1326 | CODE | |
| LOW | …ts/integration/megatron/model_support/oracle_worker.py | 878 | CODE | |
| LOW | …atron/model_support/test_dsv4_real_path_correctness.py | 41 | CODE | |
| LOW | …egration/megatron/model_support/packed_position_ids.py | 522 | CODE | |
| LOW | …s/integration/megatron/model_support/oracle_harness.py | 2043 | CODE | |
| LOW | …s/integration/megatron/model_support/oracle_harness.py | 1155 | CODE | |
| LOW | …s/integration/megatron/model_support/oracle_harness.py | 1190 | CODE | |
| LOW | …ts/integration/megatron/model_support/lora_coverage.py | 112 | CODE | |
| LOW | …ration/megatron/model_support/routing_replay_bundle.py | 193 | CODE | |
| LOW | tests/integration/megatron/model_support/workflow.py | 466 | CODE | |
| LOW | tests/integration/megatron/model_support/workflow.py | 855 | CODE | |
| LOW | …integration/megatron/model_support/hf_parity_worker.py | 532 | CODE | |
| LOW | …integration/megatron/model_support/hf_parity_worker.py | 935 | CODE | |
| LOW | …integration/megatron/model_support/hf_parity_worker.py | 149 | CODE | |
| LOW | examples/mcp-rl/mcp_rl/checks.py | 24 | CODE | |
| LOW | examples/mcp-rl/mcp_rl/utils.py | 4 | CODE | |
| LOW | examples/mcp-rl/mcp_rl/rollout.py | 43 | CODE | |
| LOW | examples/mcp-rl/mcp_rl/mcp_server.py | 139 | CODE | |
| LOW | examples/mcp-rl/mcp_rl/mcp_server.py | 243 | CODE | |
| LOW | …amples/mcp-rl/mcp_rl/benchmarks/generate_benchmarks.py | 46 | CODE | |
| LOW | …amples/mcp-rl/servers/python/mcp_balldontlie/server.py | 51 | CODE | |
| LOW | …amples/mcp-rl/servers/python/mcp_balldontlie/server.py | 189 | CODE | |
| LOW | examples/mcp-rl/servers/python/mcp_googlemaps/server.py | 180 | CODE | |
| LOW | examples/mcp-rl/servers/python/mcp_googlemaps/server.py | 340 | CODE | |
| LOW | …mples/mcp-rl/servers/python/mcp_alphavantage/server.py | 62 | CODE | |
| 135 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | tests/integration.py | 280 | print(f"Error parsing indexes '{indexes_str}': {e}") | CODE |
| MEDIUM | tests/integration.py | 316 | print("Error: No valid notebook indexes specified") | STRING |
| LOW | tests/integration/test_multi_checkpoint_training.py | 199 | except Exception: | CODE |
| LOW | tests/integration/test_push_and_fork.py | 111 | except Exception: | CODE |
| LOW | tests/integration/test_push_and_fork.py | 191 | except Exception: | CODE |
| LOW | tests/integration/test_push_and_fork.py | 252 | except Exception: | CODE |
| LOW | tests/integration/megatron/routing_replay/trace.py | 342 | except Exception: | CODE |
| LOW | …ts/integration/megatron/gdn_shared_prefix/artifacts.py | 109 | except Exception: | CODE |
| LOW | …sts/integration/megatron/lora/test_lora_disk_codecs.py | 87 | except Exception as exc: | CODE |
| LOW | …s/integration/megatron/train_inf_mismatch/real_path.py | 945 | except Exception: | CODE |
| LOW | …_inf_mismatch/vllm_forward_trace_site/sitecustomize.py | 289 | except Exception as exc: | CODE |
| LOW | …_inf_mismatch/vllm_forward_trace_site/sitecustomize.py | 363 | except Exception as exc: | CODE |
| LOW | …ts/integration/megatron/model_support/forward_trace.py | 82 | except Exception: | CODE |
| LOW | …ts/integration/megatron/model_support/forward_trace.py | 102 | except Exception: | CODE |
| LOW | …s/integration/megatron/model_support/oracle_harness.py | 832 | except Exception: | CODE |
| LOW | …s/integration/megatron/model_support/oracle_harness.py | 1326 | except Exception: | CODE |
| LOW | tests/integration/megatron/model_support/workflow.py | 935 | except Exception as exc: | CODE |
| LOW | tests/integration/megatron/model_support/workflow.py | 963 | except Exception as exc: | CODE |
| LOW | .agents/skills/train-sft/SKILL.md | 55 | except Exception as e: | CODE |
| LOW | examples/2048/utils.py | 150 | except Exception: | CODE |
| LOW | examples/2048/rollout.py | 67 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/test_scenario_generation.py | 145 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/test_scenario_generation.py | 177 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/test_scenario_generation.py | 234 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/test_scenario_generation.py | 253 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/test_scenario_generation.py | 269 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/test_scenario_generation.py | 313 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/mcp_rl/checks.py | 104 | except Exception as e: | CODE |
| MEDIUM | examples/mcp-rl/mcp_rl/checks.py | 105 | print(f"Error during LLM evaluation: {e}") | CODE |
| LOW | examples/mcp-rl/mcp_rl/rollout.py | 182 | except Exception as e: | CODE |
| MEDIUM | examples/mcp-rl/mcp_rl/rollout.py | 183 | print(f"Error checking success: {e}") | CODE |
| LOW | examples/mcp-rl/mcp_rl/rollout.py | 219 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/mcp_rl/rollout.py | 234 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/mcp_rl/rollout.py | 238 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/mcp_rl/train.py | 213 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/mcp_rl/scenario_generator.py | 68 | except Exception: | CODE |
| MEDIUM | examples/mcp-rl/mcp_rl/scenario_generator.py | 221 | print(f"Error: Server params file not found: {args.server_params_path}") | CODE |
| LOW | examples/mcp-rl/mcp_rl/scenario_generator.py | 251 | except Exception as e: | CODE |
| MEDIUM | examples/mcp-rl/mcp_rl/scenario_generator.py | 252 | print(f"Error generating scenarios: {e}") | CODE |
| LOW | …amples/mcp-rl/servers/python/mcp_balldontlie/server.py | 273 | except Exception as e: | CODE |
| LOW | examples/mcp-rl/servers/python/mcp_googlemaps/server.py | 441 | except Exception as e: | CODE |
| LOW | …mples/mcp-rl/servers/python/mcp_alphavantage/server.py | 314 | except Exception as e: | CODE |
| LOW | examples/tic_tac_toe_self_play/game_utils.py | 49 | except Exception: | CODE |
| LOW | examples/tic_tac_toe_self_play/game_utils.py | 61 | except Exception as e: | CODE |
| LOW | examples/tic_tac_toe_self_play/rollout.py | 78 | except Exception as e: | CODE |
| LOW | examples/tic_tac_toe/game_utils.py | 53 | except Exception: | CODE |
| LOW | examples/tic_tac_toe/game_utils.py | 59 | except Exception as e: | CODE |
| LOW | examples/tic_tac_toe/rollout.py | 66 | except Exception as e: | CODE |
| LOW | examples/just-the-facts/test_scraper.py | 21 | except Exception as e: | CODE |
| LOW | examples/just-the-facts/just_the_facts/utils.py | 45 | except Exception as e: | CODE |
| LOW | examples/just-the-facts/just_the_facts/utils.py | 49 | except Exception as fallback_error: | CODE |
| LOW | examples/hn_title_generator/utils.py | 122 | except Exception as e: | CODE |
| MEDIUM | examples/hn_title_generator/utils.py | 123 | print(f"Error connecting to reward model: {str(e)}") | CODE |
| LOW | examples/hn_title_generator/train.py | 54 | except Exception as e: | CODE |
| LOW | examples/hn_title_generator/train.py | 152 | except Exception as e: | CODE |
| LOW⚡ | scripts/bump_version.py | 85 | except Exception as e: | CODE |
| MEDIUM⚡ | scripts/bump_version.py | 86 | print(f"Error: {e}") | CODE |
| LOW | dev/tune_quack_lora_tiles.py | 158 | except Exception as exc: | CODE |
| LOW | dev/yes-no-maybe-fork-pipeline.py | 200 | except Exception as e: | CODE |
| LOW | dev/megatron_review_perf.py | 430 | except Exception as exc: | CODE |
| 71 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | vllm_runtime/src/art_vllm_runtime/__init__.py | 9 | __all__ = [ | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 11 | logger = logging.getLogger(__name__) | CODE |
| LOW | vllm_runtime/src/art_vllm_runtime/patches.py | 268 | def update_weights(self: Any, update_info: dict) -> None: | CODE |
| LOW | tests/unit/test_moe_routing_replay.py | 142 | def set_target_indices(self, topk_indices: torch.Tensor) -> None: | CODE |
| LOW | tests/unit/test_moe_routing_replay.py | 146 | def set_router_replay_action(self, action: Any) -> None: | CODE |
| LOW | tests/integration/megatron/routing_replay/bundle.py | 5 | __all__ = ["build_bundle_from_forward_trace_dir"] | CODE |
| LOW | …tegration/megatron/train_inf_mismatch/output_parity.py | 760 | def _set_seed(seed: int) -> None: | CODE |
| LOW | tests/integration/megatron/trainability/__init__.py | 17 | __all__ = [ | CODE |
| LOW | …ts/integration/megatron/model_support/oracle_worker.py | 159 | def _set_deterministic_seed(seed: int) -> None: | CODE |
| LOW | …s/integration/megatron/model_support/gdn_trace_uids.py | 28 | def set_module_token_uids(self, module: Any, token_uids: Tensor | None) -> None: | CODE |
| LOW⚡ | …s/integration/megatron/model_support/gdn_trace_uids.py | 89 | def set_out_proj_lora_token_uids(self, gdn: Any, hidden_states: Tensor) -> None: | CODE |
| LOW⚡ | …s/integration/megatron/model_support/gdn_trace_uids.py | 98 | def set_out_norm_token_uids(self, gdn: Any, token_uids: Tensor) -> None: | CODE |
| LOW | …s/integration/megatron/model_support/oracle_harness.py | 575 | def update_router_ids(self, reference_ids, candidate_ids) -> None: # type: ignore[no-untyped-def] | CODE |
| LOW | tests/integration/megatron/model_support/hf_parity.py | 406 | __all__ = [ | CODE |
| LOW | …integration/megatron/model_support/hf_parity_worker.py | 140 | def set_active_micro(self, sample_index: int | None, micro_slot: int) -> None: | CODE |
| LOW | examples/mcp-rl/mcp_rl/__init__.py | 6 | __all__ = ["rollout", "McpScenario", "McpServer", "AlphaMcpServer"] | CODE |
| LOW | src/art/metrics.py | 376 | def _update_throughput_metrics(self, result: dict[str, float]) -> None: | CODE |
| LOW | src/art/gather.py | 211 | def update_pbar(self, n: int) -> None: | CODE |
| LOW | src/art/__init__.py | 89 | __all__ = [ | CODE |
| LOW | src/art/auto_trajectory.py | 15 | logger = logging.getLogger(__name__) | CODE |
| LOW | src/art/model.py | 1106 | def set_cost_calculator(self, calculator: CostCalculator | None) -> None: | CODE |
| LOW | src/art/langgraph/__init__.py | 3 | __all__ = ["wrap_rollout", "init_chat_model"] | CODE |
| LOW | src/art/tinker/__init__.py | 5 | __all__ = ["TinkerBackend", "get_renderer_name", "OpenAICompatibleTinkerServer"] | CODE |
| LOW | src/art/tinker_native/__init__.py | 3 | __all__ = ["TinkerNativeBackend"] | CODE |
| LOW | src/art/serverless/__init__.py | 3 | __all__ = ["ServerlessBackend"] | CODE |
| LOW | src/art/tau_bench/__init__.py | 15 | __all__ = [ | CODE |
| LOW | src/art/rewards/__init__.py | 3 | __all__ = ["ruler", "ruler_score_group"] | CODE |
| LOW | src/art/utils/__init__.py | 10 | __all__ = [ | CODE |
| LOW | src/art/utils/deploy_model.py | 36 | __all__ = [ | CODE |
| LOW | src/art/utils/managed_process.py | 24 | def set_parent_death_signal(parent_pid: int, sig: signal.Signals) -> None: | CODE |
| LOW | src/art/utils/strip_logprobs.py | 6 | logger = logging.getLogger(__name__) | CODE |
| LOW | src/art/utils/deployment/__init__.py | 23 | __all__ = [ | CODE |
| LOW | src/art/utils/benchmarking/charts/__init__.py | 4 | __all__ = ["percentage_comparison_bar_chart", "training_progress_chart"] | CODE |
| LOW | src/art/mcp/__init__.py | 12 | __all__ = [ | CODE |
| LOW | src/art/local/backend.py | 17 | logger = logging.getLogger(__name__) | CODE |
| LOW | src/art/local/__init__.py | 3 | __all__ = ["LocalBackend"] | CODE |
| LOW | src/art/pipeline_trainer/__init__.py | 14 | __all__ = [ | CODE |
| LOW | src/art/pipeline_trainer/checkpoint_retention.py | 65 | __all__ = [ | CODE |
| LOW | src/art/pipeline_trainer/status.py | 89 | def set_step(self, step: int) -> None: | CODE |
| LOW | src/art/trainer_rank/__init__.py | 569 | def set_checkpoint(self, name: str | None) -> None: | CODE |
| LOW | src/art/trainer_rank/__init__.py | 572 | def set_lora(self, name: str | None) -> None: | CODE |
| LOW | src/art/trainer_rank/__init__.py | 1019 | def _set_default_slot(self, ref: "LoRASlotRef") -> None: | CODE |
| LOW | src/art/trainer_rank/__init__.py | 2913 | __all__ = [ | CODE |
| LOW | src/art/dev/__init__.py | 22 | __all__ = [ | CODE |
| LOW | src/art/weight_transfer/__init__.py | 9 | __all__ = [ | CODE |
| LOW | src/art/megatron/service.py | 568 | async def _set_served_model_name(self, step: int) -> None: | CODE |
| LOW | src/art/megatron/__init__.py | 3 | __all__ = ["MegatronBackend"] | CODE |
| LOW⚡ | src/art/megatron/routing_replay.py | 29 | logger = logging.getLogger(__name__) | CODE |
| LOW | src/art/megatron/routing_replay.py | 738 | def update_bundle(self, *, bundle: MoeRoutingReplayBundle, strict: bool) -> None: | CODE |
| LOW | src/art/megatron/routing_replay.py | 969 | def set_active_token_uid_key(self, token_uid_key: str) -> None: | CODE |
| LOW | src/art/megatron/train.py | 126 | __all__ = [ | CODE |
| LOW | src/art/megatron/weights/merged_weight_export.py | 516 | __all__ = [ | CODE |
| LOW | src/art/megatron/training/streaming_weight_offload.py | 18 | logger = logging.getLogger(__name__) | CODE |
| LOW | src/art/megatron/training/offload.py | 11 | logger = logging.getLogger(__name__) | CODE |
| LOW | src/art/megatron/runtime/bridge_runtime.py | 408 | def _set_tp_attrs(model: list[MegatronModule]) -> None: | CODE |
| LOW | src/art/megatron/runtime/runtime_env.py | 4 | def _set_cache_dir(env_var: str, default_path: str) -> None: | CODE |
| LOW | src/art/megatron/dsv4/deepseek_v4.py | 442 | def set_position_ids(self, position_ids: torch.Tensor | None) -> None: | CODE |
| LOW | src/art/megatron/dsv4/hf_config.py | 99 | __all__ = [ | CODE |
| LOW | src/art/megatron/dsv4/layer.py | 92 | def set_input_ids(self, input_ids: Tensor | None) -> None: | CODE |
| LOW | src/art/megatron/dsv4/layer.py | 200 | def set_input_ids(self, input_ids: Tensor | None) -> None: | CODE |
| 9 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | tests/test_backend_train_api.py | 6 | cd /workspace/ART && source .venv/bin/activate | STRING |
| HIGH⚡ | tests/unit/test_auto_trajectory.py | 119 | mock_stream_response = b"""data: {"id":"chatcmpl-aa0d1e3261414f53acafc2f8e53bf9d6","object":"chat.completion.chunk","cre | CODE |
| HIGH⚡ | tests/unit/test_auto_trajectory.py | 121 | data: {"id":"chatcmpl-aa0d1e3261414f53acafc2f8e53bf9d6","object":"chat.completion.chunk","created":1755831263,"model":"t | CODE |
| HIGH⚡ | tests/unit/test_auto_trajectory.py | 123 | data: {"id":"chatcmpl-aa0d1e3261414f53acafc2f8e53bf9d6","object":"chat.completion.chunk","created":1755831263,"model":"t | CODE |
| HIGH⚡ | tests/unit/test_auto_trajectory.py | 125 | data: {"id":"chatcmpl-aa0d1e3261414f53acafc2f8e53bf9d6","object":"chat.completion.chunk","created":1755831263,"model":"t | CODE |
| HIGH⚡ | tests/unit/test_auto_trajectory.py | 127 | data: {"id":"chatcmpl-aa0d1e3261414f53acafc2f8e53bf9d6","object":"chat.completion.chunk","created":1755831263,"model":"t | CODE |
| HIGH⚡ | tests/unit/test_auto_trajectory.py | 129 | data: {"id":"chatcmpl-aa0d1e3261414f53acafc2f8e53bf9d6","object":"chat.completion.chunk","created":1755831263,"model":"t | CODE |
| HIGH⚡ | scripts/bump_version.py | 78 | f"1. Commit the change: git add pyproject.toml uv.lock && git commit -m 'Bump version to {new}'" | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | skypilot-config.yaml | 235 | # \# Create a private VPC image and paste its ID in the following format: | COMMENT |
| MEDIUM | skypilot-config.yaml | 317 | # \# Create a S3 bucket named sky-dataset, uploads the contents of | COMMENT |
| MEDIUM | tests/unit/test_frontend_logging.py | 276 | # Create a mock trainable model with step > 0 | COMMENT |
| MEDIUM | tests/unit/test_trajectory_copy.py | 145 | # Create a memo dict with a circular reference | COMMENT |
| MEDIUM | tests/unit/test_trajectory_parquet.py | 458 | # Create a JSONL file with enough content to benefit from compression | COMMENT |
| MEDIUM | examples/mcp-rl/mcp_rl/scenario_generator.py | 113 | # Define the JSON schema for the response | COMMENT |
| MEDIUM | src/art/trajectories.py | 184 | # Create a new instance using the constructor | COMMENT |
| MEDIUM | src/art/trajectories.py | 209 | # Create a new instance with deep copies | COMMENT |
| MEDIUM | src/art/test/test_step_skipping.py | 39 | # Create a temporary directory for the test | COMMENT |
| MEDIUM | src/art/test/test_step_skipping.py | 45 | # Create a test model | COMMENT |
| MEDIUM | src/art/tinker_native/backend.py | 1086 | # Create a sampler client from the new weights | COMMENT |
| MEDIUM | src/art/serverless/backend.py | 779 | # Create a temporary directory that won't be cleaned up automatically | COMMENT |
| MEDIUM | src/art/rewards/ruler.py | 321 | # Create a new trajectory with the same data but fresh objects | STRING |
| MEDIUM | src/art/utils/litellm.py | 83 | # Create the OpenAI Choice object | COMMENT |
| MEDIUM | src/art/utils/trajectory_migration.py | 285 | # Define the message struct schema | COMMENT |
| MEDIUM | src/art/utils/old_benchmarking/generate_line_graphs.py | 55 | plt.figure() # Create a new figure for each metric | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | examples/mcp-rl/mcp_rl/mcp_server.py | 140 | Apply a tool with given arguments. Args: tool_name: Name of the tool to call arguments: | STRING |
| HIGH | examples/mcp-rl/mcp_rl/mcp_server.py | 244 | Execute a tool via the API server. Args: tool_name: Name of the tool to call arguments: | STRING |
| HIGH | examples/just-the-facts/just_the_facts/utils.py | 10 | Scrape article text from a news URL (Fox News, MSNBC, etc.) Args: url: The URL of the news article to | STRING |
| HIGH | src/art/openai.py | 20 | Consume a chat completion stream and build a complete ChatCompletion object. This function processes a stream of Ch | STRING |
| HIGH | src/art/utils/model_config.py | 28 | Detect instruction and response parts from a tokenizer's chat template. Args: tokenizer: A tokenizer with a | STRING |
| HIGH | src/art/utils/model_config.py | 76 | Get instruction and response parts for a model. Checks for explicit model configuration first, then falls back to | STRING |
| HIGH | src/art/utils/sft.py | 53 | Count the number of non-empty rows in a JSONL file. Args: file_path: Path to JSONL file Returns: | STRING |
| HIGH | src/art/utils/sft.py | 282 | Stream trajectories from a JSONL file for one or more epochs. Uses buffer-based shuffling to randomize order w | STRING |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | …s/integration/megatron/model_support/oracle_harness.py | 2004 | CODE | |
| LOW | …s/integration/megatron/model_support/oracle_harness.py | 1063 | CODE | |
| LOW | tests/integration/megatron/model_support/workflow.py | 855 | CODE | |
| LOW | dev/trainer_rank_check.py | 40 | CODE | |
| LOW | dev/megatron_review_perf.py | 39 | CODE | |
| LOW | src/art/_backend_training.py | 15 | CODE | |
| LOW | src/art/model.py | 238 | CODE | |
| LOW | src/art/model.py | 283 | CODE | |
| LOW | src/art/model.py | 1067 | CODE | |
| LOW | src/art/tinker_native/backend.py | 306 | CODE | |
| LOW | src/art/serverless/backend.py | 216 | CODE | |
| LOW | src/art/tau_bench/client.py | 110 | CODE | |
| LOW | src/art/tau_bench/rollout.py | 23 | CODE | |
| LOW | src/art/tau_bench/rollout.py | 40 | CODE | |
| LOW | src/art/tau_bench/rollout.py | 59 | CODE | |
| LOW | src/art/utils/sft.py | 145 | CODE | |
| LOW | src/art/utils/sft.py | 342 | CODE | |
| LOW | src/art/utils/strip_logprobs.py | 32 | CODE | |
| LOW | src/art/local/backend.py | 787 | CODE | |
| LOW | src/art/local/backend.py | 1439 | CODE | |
| LOW | src/art/pipeline_trainer/trainer.py | 73 | CODE | |
| LOW | src/art/megatron/prefix_tree_state.py | 61 | CODE | |
| LOW | src/art/megatron/train.py | 334 | CODE | |
| LOW | src/art/megatron/runtime/bridge_runtime.py | 481 | CODE | |
| LOW | src/art/megatron/dsv4/deepseek_v4.py | 465 | CODE | |
| LOW | src/art/megatron/dsv4/layer.py | 265 | CODE | |
| LOW | src/art/megatron/flex_attn/attention.py | 260 | CODE | |
| LOW | src/art/megatron/flex_attn/flash_dlse_patch.py | 218 | CODE | |
| LOW | src/art/megatron/context_parallel/core_attention.py | 137 | CODE | |
| LOW | src/art/megatron/model_support/handlers/gemma4.py | 627 | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | skypilot-config.yaml | 1 | # Title: Task YAML — SkyPilot documentation | COMMENT |
| LOW | skypilot-config.yaml | 21 | # \# which \`sky\` is called. | COMMENT |
| LOW | skypilot-config.yaml | 41 | # \# if this is specified. | COMMENT |
| LOW | skypilot-config.yaml | 61 | # \# Format: {<name\>:<count\>, ...} | COMMENT |
| LOW | skypilot-config.yaml | 81 | # \# E.g., 32+ means first try to find an instance type with \>\= 32 GiB. If | COMMENT |
| LOW | skypilot-config.yaml | 101 | # \# If \`EAGER\_NEXT\_REGION\` is specified, the job will go to the next region | COMMENT |
| LOW | skypilot-config.yaml | 121 | # \# low: 1000 IOPS; read 90 MB/s; write 90 MB/s | COMMENT |
| LOW | skypilot-config.yaml | 141 | # \# be removed until the cluster is terminated. | COMMENT |
| LOW | skypilot-config.yaml | 161 | # # | COMMENT |
| LOW | skypilot-config.yaml | 181 | # \# Docker support | COMMENT |
| LOW | skypilot-config.yaml | 201 | # \# custom images in multiple regions): | COMMENT |
| LOW | skypilot-config.yaml | 221 | # \# image\_id: skypilot:gpu-ubuntu-2004 | COMMENT |
| LOW | skypilot-config.yaml | 241 | # \# To use a more limited but easier to manage tool: | COMMENT |
| LOW | skypilot-config.yaml | 261 | # \# outside of \`any\_of\`, \`ordered\` will be used as the default values for | COMMENT |
| LOW | skypilot-config.yaml | 281 | # \# \`sky launch/exec --env ENV=val\` (if ENV is present). | COMMENT |
| LOW | skypilot-config.yaml | 301 | # envs: | COMMENT |
| LOW | skypilot-config.yaml | 321 | # # | COMMENT |
| LOW | skypilot-config.yaml | 341 | # # | COMMENT |
| LOW | skypilot-config.yaml | 361 | COMMENT | |
| LOW | …tegration/megatron/train_inf_mismatch/output_parity.py | 21 | # These gates are intentionally bf16-scale, not fp32 oracle-scale. A 2026-05-18 | COMMENT |
| LOW | requirements/backend.vcs.txt | 1 | # Pinned backend dependencies that must come from VCS (not allowed in PyPI metadata). | COMMENT |
| LOW | dev/yes-no-maybe.py | 41 | load_dotenv() | COMMENT |
| LOW | src/art/tinker/server.py | 521 | base_model = model_path_or_base_model | COMMENT |
| LOW | …benchmarking/charts/percentage_comparison_bar_chart.py | 241 | ax.spines["right"].set_visible(False) | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | examples/mcp-rl/servers/python/mcp_googlemaps/README.md | 39 | python server.py --api-key YOUR_API_KEY | CODE |
| HIGH | examples/mcp-rl/servers/python/mcp_googlemaps/README.md | 44 | export GOOGLE_MAPS_API_KEY=your_api_key | CODE |
| HIGH | …mples/mcp-rl/servers/python/mcp_alphavantage/README.md | 25 | python server.py --api-key YOUR_API_KEY | CODE |
| HIGH | …mples/mcp-rl/servers/python/mcp_alphavantage/README.md | 30 | export ALPHAVANTAGE_API_KEY=your_api_key | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW⚡ | .agents/skills/train-sft/SKILL.md | 14 | ## Step 1: Determine Training Scenario | COMMENT |
| LOW⚡ | .agents/skills/train-sft/SKILL.md | 22 | ## Step 2: Determine Backend | COMMENT |
| LOW⚡ | .agents/skills/train-sft/SKILL.md | 28 | ## Step 3: Select and Validate Dataset (JSONL scenario) | COMMENT |
| LOW | .agents/skills/train-sft/SKILL.md | 68 | ## Step 4: Gather Base Parameters | COMMENT |
| LOW | .agents/skills/train-sft/SKILL.md | 84 | ## Step 5: Gather Hyperparameters | COMMENT |
| LOW | .agents/skills/train-sft/SKILL.md | 111 | ## Step 6: Generate the Training Script | COMMENT |
| LOW | .agents/skills/train-sft/SKILL.md | 287 | ## Step 7: Write and Offer to Run | STRING |
| LOW⚡ | src/art/local/backend.py | 1395 | # Step 1: Ensure checkpoint exists at original_checkpoint_dir | COMMENT |
| LOW | src/art/local/backend.py | 1419 | # Step 2: Handle local_path if provided | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | examples/hn_title_generator/reference_grpo_trainer.py | 289 | if scores: # Check if scores list is not empty | CODE |
| LOW | examples/hn_title_generator/reference_grpo_trainer.py | 310 | if v_list: # Check if list for metric is not empty | CODE |
| LOW | scripts/publish.sh | 8 | # Check if PYPI_ART_TOKEN is set | COMMENT |
| LOW | src/art/utils/s3_checkpoint_utils.py | 166 | # Check if it's a 4-digit checkpoint directory (old format) | COMMENT |
| LOW | src/art/utils/s3_checkpoint_utils.py | 192 | # Check if already exists in new location | COMMENT |
| LOW | src/art/utils/s3.py | 133 | # Check if bucket exists | COMMENT |
| LOW⚡ | src/art/local/backend.py | 1390 | # Check if checkpoint exists in the original training location | COMMENT |
| LOW | src/art/megatron/dsv4/encoding.py | 347 | # Check if previous message has a task - if so, this is a task output (no thinking) | COMMENT |
| LOW | src/mp_actors/move.py | 256 | # Check if it's a method or property | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH⚡ | …benchmarking/charts/percentage_comparison_bar_chart.py | 231 | # Remove y-axis label, ticks, and spine as requested | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | tests/unit/test_frontend_logging.py | 1152 | CODE | |
| MEDIUM | src/art/tinker/service.py | 198 | CODE | |
| MEDIUM | src/art/tinker_native/backend.py | 471 | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW⚡ | src/art/local/backend.py | 1395 | # Step 1: Ensure checkpoint exists at original_checkpoint_dir | COMMENT |
| LOW | src/art/local/backend.py | 1419 | # Step 2: Handle local_path if provided | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | examples/hn_title_generator/train.py | 138 | # Be robust to variations like "True.", " False", etc. | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | skypilot-config.yaml | 156 | # \# Example usage: | COMMENT |