A PyTorch native platform for training generative AI models
This report summarizes scanner evidence for pytorch/torchtitan, a Python project with 5,796 GitHub stars. SynthScan examined 250,991 lines of code across 942 source files, recording 4768 pattern matches distributed across 21 syntactic categories. The legacy adjusted pattern signal is 21.3 (Elevated pattern signal); it is not an authorship estimate.
Use the findings to ask whether maintainers can explain intent, verify dependencies, simplify unnecessary scaffolding, and focus review effort. Risk Payload v1 adds ownership, supply-chain, over-engineering, and review-effort vectors when available. These are prompts for human verification, not probabilities, security findings, or measured maintenance costs.
Use each vector to guide manual verification of the underlying findings.
Can maintainers explain intent, assumptions, and trade-offs? 3539 finding(s).
Are imports declared and package/API names verified? 652 finding(s).
Are wrappers, guards, and repeated structures necessary? 3569 finding(s).
Which changes need focused human review? 4768 finding(s).
This chart tracks the legacy adjusted pattern score across scan runs. Changes can reflect source changes, scan coverage, or detector/scoring versions; they do not measure AI adoption or prove that a team refactored code. Compare runs only when their scoring versions are compatible. The dashed line tracks the raw pattern-hit count.
Groups findings by configured severity weight. Severity is not a confidence estimate, security rating, or proof that a package is missing. Verify imports against the project's manifests and inspect each finding in its local design context.
This horizontal bar chart shows where configured pattern scores accumulate across directories. Treat the bars as locations to inspect, not as evidence of authorship or code quality. Use the file-level findings to decide whether a directory needs a focused ownership, dependency, complexity, or review check.
The scanner identified 4768 distinct pattern matches across 21 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a detector match to inspect in context. 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 configured finding weight (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — legacy comment-context matches receive a ×1.5 score weight. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier under the legacy scoring model.
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | …ts/unit_tests/gpu/test_deepep_activation_checkpoint.py | 123 | def test_dispatch_and_combine_are_saved_not_replayed(self): | CODE |
| LOW | …ts/unit_tests/gpu/test_deepep_activation_checkpoint.py | 137 | def test_region_ac_replay_is_exact(self): | CODE |
| LOW | …ts/unit_tests/gpu/test_deepep_activation_checkpoint.py | 146 | def _check_deepep_activation_checkpointing( | CODE |
| LOW | tests/unit_tests/gpu/test_gated_rmsnorm.py | 37 | def test_default_forward_and_backward_emit_triton(self): | CODE |
| LOW | tests/unit_tests/gpu/test_kda_attention.py | 84 | def test_fragments_follow_the_explicit_permutation(self): | CODE |
| LOW | tests/unit_tests/gpu/test_kda_attention.py | 94 | def test_backend_builds_routing_from_global_varlen_metadata(self): | CODE |
| LOW | tests/unit_tests/gpu/test_kda_attention.py | 127 | def test_metadata_annotates_cp_routing_tensors(self): | CODE |
| LOW | tests/unit_tests/gpu/test_kda_attention.py | 175 | def test_varlen_matches_independent_documents(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 65 | def test_construction_does_not_crash_on_dtensor_pin_memory(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 79 | def test_step_updates_offloaded_values_correctly(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 104 | def test_scratch_buffer_reused_across_steps(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 122 | def test_scratch_is_bounded_well_below_parameter_memory(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 146 | def test_compute_stream_waits_for_the_offload_read(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 194 | def test_materialized_dtensor_is_safe_to_read_from_another_stream(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 220 | def test_state_dict_leaves_the_container_pinned(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 241 | def test_state_dict_failure_leaves_the_container_usable(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 274 | def test_offload_and_plain_load_agree_on_a_partial_state_dict(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 296 | def test_offload_state_dict_key_layout_matches_super(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 315 | def test_save_load_round_trip_stays_offloaded(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 349 | def test_multi_copy_round_trip_through_offload_paths(self): | CODE |
| LOW | tests/unit_tests/gpu/test_ema.py | 406 | def _build_sharded_model_with_buffer(self) -> nn.Module: | CODE |
| LOW⚡ | tests/unit_tests/gpu/test_fused_swiglu.py | 74 | def test_replaces_only_activation(self): | CODE |
| LOW⚡ | tests/unit_tests/gpu/test_fused_swiglu.py | 79 | def test_gate_up_projection_is_linear(self): | CODE |
| LOW⚡ | tests/unit_tests/gpu/test_fused_swiglu.py | 88 | def test_preserves_converted_linear_config(self): | CODE |
| LOW | tests/unit_tests/gpu/test_fused_swiglu.py | 102 | def test_saves_native_layout_without_copy(self): | CODE |
| LOW | tests/unit_tests/gpu/test_fused_swiglu.py | 109 | def test_triton_checkpoint_loads_into_native(self): | CODE |
| LOW | tests/unit_tests/gpu/test_fused_swiglu.py | 120 | def test_native_checkpoint_loads_into_triton(self): | CODE |
| LOW | tests/unit_tests/gpu/test_fused_swiglu.py | 138 | def test_strict_load_reports_missing(self): | CODE |
| LOW | tests/unit_tests/gpu/test_fused_swiglu.py | 162 | def test_dist_gemm_config_keeps_overlap(self): | CODE |
| LOW | tests/unit_tests/gpu/test_fused_swiglu.py | 193 | def test_hf_adapter_roundtrip(self): | CODE |
| LOW | tests/unit_tests/gpu/test_cuda_graph_group_replay.py | 21 | def test_group_graph_replay_accumulates_existing_gradients() -> None: | CODE |
| LOW | tests/unit_tests/gpu/test_cuda_graph_group_replay.py | 60 | def test_group_graph_replay_accumulates_reduced_fsdp_gradients() -> None: | CODE |
| LOW | tests/unit_tests/gpu/test_cuda_graph_group_replay.py | 96 | def _run_real_fsdp_group_graph_numerics( | CODE |
| LOW | tests/unit_tests/gpu/test_cuda_graph_group_replay.py | 186 | def test_group_graph_replay_matches_real_fsdp_numerics() -> None: | CODE |
| LOW | tests/unit_tests/gpu/test_hi_mid_lo_linear.py | 39 | def test_hi_mid_lo_linear_compiles(input_dtype, weight_dtype, backward_mode): | CODE |
| LOW | tests/unit_tests/gpu/test_hi_mid_lo_linear.py | 135 | def test_compiled_backward_keeps_mid(): | CODE |
| LOW | tests/unit_tests/gpu/test_hi_mid_lo_linear.py | 154 | def test_third_piece_keeps_what_two_pieces_drop(num_pieces, num_tokens, compile): | CODE |
| LOW | tests/unit_tests/gpu/test_hi_mid_lo_linear.py | 189 | def test_compiled_split_matches_eager_split(num_pieces, dim): | CODE |
| LOW | tests/unit_tests/gpu/test_hi_mid_lo_linear.py | 209 | def test_backward_compiles_the_split_once_per_pieces_and_layout(): | CODE |
| LOW | tests/unit_tests/gpu/test_hi_mid_lo_linear.py | 244 | def test_backward_traces_with_make_fx(tracing_mode, out_features): | CODE |
| LOW | tests/unit_tests/gpu/test_hi_mid_lo_linear.py | 290 | def test_backward_handles_zero_tokens(): | CODE |
| LOW | tests/unit_tests/gpu/test_hi_mid_lo_linear.py | 386 | def test_backward_accumulates_into_running_weight_grad( | CODE |
| LOW | tests/unit_tests/gpu/test_hi_mid_lo_linear.py | 443 | def test_inplace_wgrad_accum_raises_under_compile(): | CODE |
| LOW | tests/unit_tests/gpu/test_hi_mid_lo_linear.py | 449 | def _run_fsdp_accumulates_chunk_weight_grads(rank, world_size, port): | CODE |
| LOW | tests/unit_tests/gpu/test_hi_mid_lo_linear.py | 510 | def test_fsdp_accumulates_chunk_weight_grads(): | CODE |
| LOW | tests/unit_tests/gpu/test_fsdp_cpu_offload.py | 152 | def test_vision_encoder_gets_the_same_offload_policy_as_the_decoder(self) -> None: | CODE |
| LOW | tests/unit_tests/gpu/test_fsdp_cpu_offload.py | 159 | def test_backward_runs_with_cpu_offload_on_a_model_with_a_vision_encoder( | CODE |
| LOW | tests/unit_tests/gpu/test_fsdp_cpu_offload.py | 178 | def test_no_offload_policy_when_cpu_offload_is_off(self) -> None: | CODE |
| LOW | tests/unit_tests/gpu/test_fsdp_cpu_offload.py | 189 | def _run_mixed_rank_image_presence(self, *, cpu_offload: bool) -> None: | CODE |
| LOW | tests/unit_tests/gpu/test_fsdp_cpu_offload.py | 236 | def test_mixed_rank_image_presence_completes_backward(self) -> None: | CODE |
| LOW | tests/unit_tests/gpu/test_fsdp_cpu_offload.py | 240 | def test_mixed_rank_image_presence_completes_backward_with_cpu_offload( | CODE |
| LOW | tests/unit_tests/gpu/test_fsdp_cpu_offload.py | 252 | def test_image_then_empty_accumulation(self) -> None: | CODE |
| LOW | tests/unit_tests/gpu/test_kimi_k3_attention_residual.py | 55 | def test_forward_and_backward_emit_triton(self): | CODE |
| LOW | tests/unit_tests/gpu/test_kimi_k3_attention_residual.py | 77 | def test_forward_and_backward_match_reference(self): | CODE |
| LOW | tests/unit_tests/gpu/test_kimi_k3_attention_residual.py | 159 | def test_forward_and_backward_are_batch_invariant(self): | CODE |
| LOW | tests/unit_tests/gpu/test_nvfp4_linear.py | 55 | def _install_unsharded_weight(linear: NVFP4Linear) -> NVFP4Linear: | CODE |
| LOW | tests/unit_tests/gpu/test_remat_all_to_all.py | 143 | def test_collective_replay_follows_region_policy(self): | CODE |
| LOW | tests/unit_tests/gpu/test_remat_all_to_all.py | 288 | def test_torchao_padded_dispatch_matches_unpadded(self): | CODE |
| LOW | tests/unit_tests/gpu/test_mxfp8_fsdp.py | 63 | def _run_reshard_after_forward( | CODE |
| LOW | tests/unit_tests/gpu/test_mxfp8_fsdp.py | 274 | def _run_cuda_graph_cache_lifecycle( | CODE |
| 2912 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM⚡ | tests/unit_tests/cpu/test_tp_kv_heads_validation.py | 117 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | tests/unit_tests/cpu/test_tp_kv_heads_validation.py | 119 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | tests/unit_tests/cpu/test_tp_kv_heads_validation.py | 127 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | tests/unit_tests/cpu/test_tp_kv_heads_validation.py | 129 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | tests/unit_tests/cpu/test_tp_kv_heads_validation.py | 137 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | tests/unit_tests/cpu/test_tp_kv_heads_validation.py | 139 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 140 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 142 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 373 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 375 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 603 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 605 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 764 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 766 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 806 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 808 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 863 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 865 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 916 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 918 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 1696 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | …nit_tests/cpu/observability/test_structured_logging.py | 1698 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 58 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 60 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 352 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 354 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 637 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 639 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 953 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 955 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 1161 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 1163 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 1380 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM | …nit_tests/cpu/observability/test_structured_logging.py | 1386 | # ------------------------------------------------------------------ | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 95 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 97 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 310 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 312 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 458 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 460 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 525 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 527 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 564 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 566 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/cpu/test_metrics.py | 291 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/cpu/test_metrics.py | 293 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/cpu/test_metrics.py | 362 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/cpu/test_metrics.py | 364 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/cpu/test_metrics.py | 478 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/cpu/test_metrics.py | 480 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/cpu/test_metrics.py | 487 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/cpu/test_metrics.py | 489 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/cpu/test_metrics.py | 716 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | tests/rl/unit_tests/cpu/test_metrics.py | 718 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/rl/unit_tests/cpu/test_metrics.py | 23 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM | tests/rl/unit_tests/cpu/test_metrics.py | 25 | # --------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | scripts/loss_compare.py | 316 | # ============================================================================= | COMMENT |
| MEDIUM⚡ | scripts/loss_compare.py | 318 | # ============================================================================= | COMMENT |
| MEDIUM | scripts/loss_compare.py | 84 | # ============================================================================= | COMMENT |
| MEDIUM | scripts/loss_compare.py | 86 | # ============================================================================= | COMMENT |
| 151 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | tests/rl/unit_tests/cpu/test_rollout_recorder.py | 9 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_dapo_math.py | 9 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_alphabet_sort.py | 9 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_metrics.py | 9 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_metrics.py | 611 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_train_loop.py | 9 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_muse_glimmer_renderer.py | 20 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_intra_generator_router.py | 9 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_rollout_utils.py | 9 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_verifiers.py | 9 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_runtime_env.py | 7 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_engine_loop.py | 14 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_verifiers_example.py | 9 | CODE | |
| LOW | …checkpoint_conversion/numerical_tests_qwen3_5_shard.py | 29 | CODE | |
| LOW | torchtitan_recipes/overrides/helion_rope.py | 35 | CODE | |
| LOW | torchtitan_recipes/overrides/moe_token_dispatcher.py | 40 | CODE | |
| LOW | torchtitan_recipes/rl/verifiers_dapo_math.py | 9 | CODE | |
| LOW | torchtitan_recipes/rl/dapo_math.py | 9 | CODE | |
| LOW | torchtitan_recipes/rl/search_r1.py | 17 | CODE | |
| LOW | torchtitan/training_engine.py | 54 | CODE | |
| LOW | torchtitan/tools/filesystem.py | 15 | CODE | |
| LOW | torchtitan/distributed/spmd_types.py | 9 | CODE | |
| LOW | torchtitan/distributed/fsdp.py | 42 | CODE | |
| LOW | torchtitan/distributed/fsdp.py | 43 | CODE | |
| LOW | torchtitan/distributed/__init__.py | 7 | CODE | |
| LOW | torchtitan/distributed/__init__.py | 8 | CODE | |
| LOW | torchtitan/distributed/__init__.py | 8 | CODE | |
| LOW | torchtitan/distributed/utils.py | 7 | CODE | |
| LOW | torchtitan/distributed/context_parallel.py | 9 | CODE | |
| LOW | torchtitan/distributed/parallelism_context.py | 13 | CODE | |
| LOW | …n/distributed/flex_shard/_optimizer_reshard_runtime.py | 16 | CODE | |
| LOW | torchtitan/distributed/flex_shard/dist_muon.py | 9 | CODE | |
| LOW | torchtitan/distributed/flex_shard/__init__.py | 9 | CODE | |
| LOW | torchtitan/distributed/flex_shard/__init__.py | 10 | CODE | |
| LOW | torchtitan/distributed/flex_shard/__init__.py | 10 | CODE | |
| LOW | torchtitan/distributed/flex_shard/__init__.py | 10 | CODE | |
| LOW | torchtitan/distributed/flex_shard/__init__.py | 10 | CODE | |
| LOW | …/distributed/flex_shard/_optimizer_reshard_schedule.py | 9 | CODE | |
| LOW | torchtitan/distributed/flex_shard/optimizer_reshard.py | 9 | CODE | |
| LOW | torchtitan/config/__init__.py | 15 | CODE | |
| LOW | torchtitan/config/__init__.py | 15 | CODE | |
| LOW | torchtitan/config/__init__.py | 15 | CODE | |
| LOW | torchtitan/config/__init__.py | 16 | CODE | |
| LOW | torchtitan/config/__init__.py | 17 | CODE | |
| LOW | torchtitan/config/__init__.py | 18 | CODE | |
| LOW | torchtitan/config/__init__.py | 19 | CODE | |
| LOW | torchtitan/config/__init__.py | 19 | CODE | |
| LOW | torchtitan/config/__init__.py | 19 | CODE | |
| LOW | torchtitan/config/__init__.py | 19 | CODE | |
| LOW | torchtitan/config/__init__.py | 19 | CODE | |
| LOW | torchtitan/config/__init__.py | 19 | CODE | |
| LOW | torchtitan/config/__init__.py | 27 | CODE | |
| LOW | torchtitan/config/__init__.py | 27 | CODE | |
| LOW | torchtitan/config/loader.py | 9 | CODE | |
| LOW | torchtitan/config/override.py | 32 | CODE | |
| LOW | torchtitan/config/validation.py | 9 | CODE | |
| LOW | torchtitan/config/validation.py | 22 | CODE | |
| LOW | torchtitan/config/transform/dist_moe.py | 9 | CODE | |
| LOW | torchtitan/config/transform/__init__.py | 9 | CODE | |
| LOW | torchtitan/config/transform/__init__.py | 9 | CODE | |
| 492 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW⚡ | tests/unit_tests/cpu/test_fsdp.py | 30 | def set_force_sum_reduction_for_comms(self, enable: bool) -> None: | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_fsdp.py | 33 | def set_symm_mem_for_comm(self) -> None: | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_trainer.py | 1241 | def set_requires_all_reduce(self, flag: bool, *, recurse: bool = True) -> None: | STRING |
| LOW⚡ | tests/unit_tests/cpu/test_trainer.py | 1245 | def set_is_last_backward(self, flag: bool) -> None: | STRING |
| LOW⚡ | tests/unit_tests/cpu/test_trainer.py | 1248 | def set_reshard_after_backward(self, flag: bool) -> None: | STRING |
| LOW⚡ | tests/unit_tests/cpu/test_trainer.py | 1251 | def set_requires_gradient_sync(self, flag: bool, *, recurse: bool = True) -> None: | STRING |
| LOW | tests/integration_tests/run_tests.py | 33 | logger = logging.getLogger(__name__) | CODE |
| LOW | tests/integration_tests/__init__.py | 14 | __all__ = [ | CODE |
| LOW | tests/integration_tests/flux.py | 23 | logger = logging.getLogger(__name__) | CODE |
| LOW⚡ | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 89 | logger = logging.getLogger(__name__) | CODE |
| LOW | tests/rl/unit_tests/gpu/test_bitwise_parity.py | 178 | def _set_generator_determinism(debug) -> None: | CODE |
| LOW | …s/rl/unit_tests/cpu/test_vllm_dist_moe_padding_mask.py | 51 | def set_num_local_input_tokens_per_call(self, num_tokens: int) -> None: | CODE |
| LOW | …s/rl/unit_tests/cpu/test_vllm_dist_moe_padding_mask.py | 152 | def set_num_valid_tokens(self, n: int) -> None: | CODE |
| LOW | tests/rl/integration_tests/rl.py | 48 | logger = logging.getLogger(__name__) | CODE |
| LOW | …skills/numerics_debugging/scripts/activation_tracer.py | 61 | logger = logging.getLogger(__name__) | CODE |
| LOW⚡ | …skills/numerics_debugging/scripts/activation_tracer.py | 381 | def set_numerics_capture_active(active: bool) -> None: | CODE |
| LOW | …skills/numerics_debugging/scripts/activation_tracer.py | 877 | def _setup(self) -> None: | CODE |
| LOW | …ripts/checkpoint_conversion/numerical_tests_example.py | 19 | logger = logging.getLogger(__name__) | CODE |
| LOW | scripts/checkpoint_conversion/convert_ema_to_hf.py | 30 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan_recipes/overrides/fused_mla.py | 86 | __all__ = [ | CODE |
| LOW | torchtitan_recipes/overrides/kimi_attention_residual.py | 41 | __all__ = [ | CODE |
| LOW | torchtitan_recipes/overrides/helion_rope.py | 51 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan_recipes/overrides/helion_rope.py | 71 | __all__ = ["HelionComplexRoPE"] | CODE |
| LOW | torchtitan_recipes/overrides/fused_swiglu.py | 30 | __all__ = [ | CODE |
| LOW | torchtitan_recipes/tests/__init__.py | 20 | def _set_spmd_typechecking(config: Trainer.Config, *, typechecking: bool) -> None: | CODE |
| LOW | torchtitan_recipes/tests/suites/features.py | 58 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan_recipes/tests/models/multimodal.py | 66 | def set_rank_conditional_image_presence(config: Trainer.Config) -> None: | CODE |
| LOW | torchtitan/training_engine.py | 57 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/train.py | 19 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/trainer.py | 37 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/tools/utils.py | 19 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/tools/garbage_collector.py | 18 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/distributed/spmd_types.py | 27 | __all__ = [ | CODE |
| LOW | torchtitan/distributed/fsdp.py | 27 | __all__ = [ | CODE |
| LOW | torchtitan/distributed/fsdp.py | 38 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/distributed/cuda_graph.py | 23 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/distributed/offloading.py | 9 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/distributed/activation_checkpoint.py | 23 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/distributed/__init__.py | 13 | __all__ = [ | CODE |
| LOW | torchtitan/distributed/local_compile.py | 83 | __all__ = ["apply_local_compile", "local_compile"] | CODE |
| LOW | torchtitan/distributed/utils.py | 30 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/distributed/batch_invariant.py | 13 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/distributed/batch_invariant.py | 73 | def set_batch_invariance(enable: bool) -> None: | CODE |
| LOW | torchtitan/distributed/context_parallel.py | 39 | __all__ = [ | CODE |
| LOW | torchtitan/distributed/pipeline_parallel.py | 41 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/distributed/pipeline_parallel.py | 44 | __all__ = [ | CODE |
| LOW | torchtitan/distributed/parallelism_context.py | 34 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/distributed/parallelism_context.py | 37 | __all__ = [ | CODE |
| LOW | torchtitan/distributed/flex_shard/dist_muon.py | 55 | __all__ = [ | CODE |
| LOW | torchtitan/distributed/flex_shard/__init__.py | 12 | __all__ = [ | CODE |
| LOW | torchtitan/distributed/flex_shard/optimizer_reshard.py | 18 | __all__ = ["BlockShard", "BucketConfig", "ComputeLayout", "Owned"] | CODE |
| LOW | torchtitan/distributed/deepep/hybridep.py | 30 | logger = logging.getLogger(__name__) | CODE |
| LOW⚡ | torchtitan/distributed/deepep/hybridep.py | 584 | __all__ = [ | CODE |
| LOW | torchtitan/config/configurable.py | 15 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/config/__init__.py | 29 | __all__ = [ | CODE |
| LOW | torchtitan/config/loader.py | 23 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/config/loader.py | 179 | def _set_attr(config: object, name: str, value: Any) -> None: | CODE |
| LOW | torchtitan/config/override.py | 42 | logger = logging.getLogger(__name__) | CODE |
| LOW | torchtitan/config/validation.py | 24 | __all__ = [ | CODE |
| LOW | torchtitan/config/transform/dist_moe.py | 27 | __all__ = ["DistMoeTransform"] | CODE |
| 235 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | …ts/unit_tests/gpu/test_deepep_activation_checkpoint.py | 146 | CODE | |
| LOW | tests/unit_tests/gpu/test_ema.py | 349 | CODE | |
| LOW | tests/unit_tests/gpu/test_remat.py | 866 | CODE | |
| LOW | tests/unit_tests/gpu/test_remat.py | 947 | CODE | |
| LOW | …ts/unit_tests/gpu/test_tensor_parallel_feed_forward.py | 40 | CODE | |
| LOW | tests/unit_tests/gpu/test_fsdp_moe_sharding.py | 207 | CODE | |
| LOW | tests/unit_tests/gpu/test_fsdp_embedding.py | 39 | CODE | |
| LOW | tests/unit_tests/cpu/test_ema.py | 267 | CODE | |
| LOW | tests/unit_tests/cpu/test_state_dict_adapter.py | 255 | CODE | |
| LOW | tests/unit_tests/cpu/test_state_dict_adapter.py | 310 | CODE | |
| LOW | tests/unit_tests/cpu/test_tokenizer.py | 118 | CODE | |
| LOW | tests/unit_tests/cpu/test_tokenizer.py | 474 | CODE | |
| LOW | tests/unit_tests/cpu/test_tokenizer.py | 496 | CODE | |
| LOW | tests/unit_tests/cpu/test_tokenizer.py | 519 | CODE | |
| LOW | tests/unit_tests/cpu/test_dataset_checkpointing.py | 35 | CODE | |
| LOW | tests/unit_tests/cpu/test_convert_ema_to_hf.py | 135 | CODE | |
| LOW | tests/unit_tests/cpu/test_convert_ema_to_hf.py | 167 | CODE | |
| LOW | tests/unit_tests/cpu/test_kimi_k3_pp_block_grads.py | 223 | CODE | |
| LOW | tests/unit_tests/cpu/test_kimi_k3_pp_block_grads.py | 248 | CODE | |
| LOW | tests/unit_tests/cpu/test_deepseek_v4_dsa.py | 97 | CODE | |
| LOW | tests/unit_tests/cpu/test_embedding.py | 224 | CODE | |
| LOW | tests/unit_tests/cpu/test_validator_frequency.py | 13 | CODE | |
| LOW | tests/unit_tests/cpu/test_loss.py | 421 | CODE | |
| LOW | tests/unit_tests/cpu/test_loss.py | 546 | CODE | |
| LOW | tests/unit_tests/cpu/test_linear.py | 392 | CODE | |
| LOW | tests/unit_tests/cpu/test_linear.py | 426 | CODE | |
| LOW | tests/unit_tests/cpu/observability/test_sdc_replayer.py | 162 | CODE | |
| LOW | tests/unit_tests/cpu/observability/test_sdc_replayer.py | 171 | CODE | |
| LOW | tests/integration_tests/run_tests.py | 139 | CODE | |
| LOW | tests/integration_tests/run_tests.py | 380 | CODE | |
| LOW | tests/assets/cc12m_test/pack_test_dataset.py | 13 | CODE | |
| LOW | tests/rl/unit_tests/gpu/test_gdn_cudagraph.py | 152 | CODE | |
| LOW | tests/rl/unit_tests/gpu/test_gdn_cudagraph.py | 266 | CODE | |
| LOW | tests/rl/unit_tests/cpu/test_batcher.py | 181 | CODE | |
| LOW | …h_trace_compaction/scripts/compact_cuda_graph_trace.py | 693 | CODE | |
| LOW | …h_trace_compaction/scripts/compact_cuda_graph_trace.py | 743 | CODE | |
| LOW | …skills/numerics_debugging/scripts/activation_tracer.py | 428 | CODE | |
| LOW | …skills/numerics_debugging/scripts/activation_tracer.py | 459 | CODE | |
| LOW | …/skills/numerics_debugging/scripts/compare_numerics.py | 82 | CODE | |
| LOW | …/skills/numerics_debugging/scripts/compare_numerics.py | 341 | CODE | |
| LOW | …/skills/numerics_debugging/scripts/compare_numerics.py | 489 | CODE | |
| LOW | …/skills/numerics_debugging/scripts/compare_numerics.py | 1084 | CODE | |
| LOW | …/skills/numerics_debugging/scripts/compare_numerics.py | 573 | CODE | |
| LOW | scripts/loss_compare.py | 572 | CODE | |
| LOW | scripts/download_hf_assets.py | 12 | CODE | |
| LOW | .github/scripts/todo_debt_tracker.py | 120 | CODE | |
| LOW | .github/scripts/todo_debt_tracker.py | 237 | CODE | |
| LOW | .github/scripts/todo_debt_tracker.py | 338 | CODE | |
| LOW | torchtitan_recipes/tests/suites/features.py | 80 | CODE | |
| LOW | torchtitan/trainer.py | 465 | CODE | |
| LOW | torchtitan/trainer.py | 116 | CODE | |
| LOW | torchtitan/tools/utils.py | 112 | CODE | |
| LOW | torchtitan/distributed/spmd_types.py | 231 | CODE | |
| LOW | torchtitan/distributed/fsdp.py | 225 | CODE | |
| LOW | torchtitan/distributed/pipeline_parallel.py | 482 | CODE | |
| LOW | torchtitan/distributed/pipeline_parallel.py | 595 | CODE | |
| LOW | …n/distributed/flex_shard/_optimizer_reshard_runtime.py | 128 | CODE | |
| LOW | …n/distributed/flex_shard/_optimizer_reshard_runtime.py | 420 | CODE | |
| LOW | torchtitan/distributed/flex_shard/dist_muon.py | 1104 | CODE | |
| LOW | torchtitan/distributed/flex_shard/dist_muon.py | 1375 | CODE | |
| 128 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | .ci/docker/common/install_conda.sh | 58 | # WARNING: This is a HACK from PyTorch core to be able to build PyTorch on 22.04. | COMMENT |
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 366 | # Step 2: Load tokenizer using our function | COMMENT |
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 372 | # Step 3: Load tokenizer using official Tokenizer library (if available) | COMMENT |
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 379 | # Step 4: Load tokenizer using transformers AutoTokenizer (if available) | COMMENT |
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 388 | # Step 5: Compare underlying tokenizer attributes (only if official tokenizer is available) | COMMENT |
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 394 | # Step 6: Compare with transformers tokenizer if available | COMMENT |
| LOW | tests/unit_tests/cpu/test_tokenizer.py | 351 | # Step 1: Download tokenizer files | COMMENT |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 80 | 0.5, # Step 0: 50% of max LR (warmup) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 81 | 1.0, # Step 1: 100% of max LR (warmup complete) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 82 | 1.0, # Step 2: We maunally added step of stable phase, to prevent LR from dropping to 0 at last step | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 83 | 7.0 / 8.0, # Step 3: 7/8 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 84 | 6.0 / 8.0, # Step 4: 3/4 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 85 | 5.0 / 8.0, # Step 5: 5/8 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 86 | 4.0 / 8.0, # Step 6: 1/2 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 87 | 3.0 / 8.0, # Step 7: 3/8 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 88 | 2.0 / 8.0, # Step 8: 1/4 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 89 | 1.0 / 8.0, # Step 9: 1/8 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 123 | 0.5, # Step 0: 50% of max LR (warmup) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 124 | 1.0, # Step 1: 100% of max LR (warmup complete) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 125 | 1.0, # Step 2: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 126 | 1.0, # Step 3: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 127 | 1.0, # Step 4: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 128 | 1.0, # Step 5: We maunally added step of stable phase, to prevent LR from dropping to 0 at last step | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 129 | 0.8, # Step 6: Linear decay starts (80% of max LR) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 130 | 0.6, # Step 7: 60% of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 131 | 0.4, # Step 8: 40% of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 132 | 0.2, # Step 9: 20% of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 234 | 0.2, # Step 0: 50% of max LR (warmup) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 235 | 0.4, # Step 1: 100% of max LR (warmup complete) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 236 | 0.6, # Step 2: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 237 | 0.8, # Step 3: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 238 | 1.0, # Step 4: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 271 | 0.5, # Step 0: 50% of max LR (warmup) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 272 | 1.0, # Step 1: 100% of max LR (warmup complete) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 273 | 1.0, # Step 2: We maunally added step of stable phase, to prevent LR from dropping to 0 at last step | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 274 | 1.0, # Step 3: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 275 | 1.0, # Step 4: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 276 | 1.0, # Step 5: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 277 | 1.0, # Step 6: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 278 | 1.0, # Step 7: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 279 | 1.0, # Step 8: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 280 | 1.0, # Step 9: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 314 | 0.2, # Step 0: 50% of max LR (warmup) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 315 | 0.4, # Step 1: 100% of max LR (warmup complete) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 316 | 0.6, # Step 2: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 317 | 0.8, # Step 3: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 318 | 1.0, # Step 4: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 319 | 1.0, # Step 5: We maunally added step of stable phase, to prevent LR from dropping to 0 at last step | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 320 | 0.8, # Step 6: Linear decay starts (80% of max LR) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 321 | 0.6, # Step 7: 60% of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 322 | 0.4, # Step 8: 40% of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 323 | 0.2, # Step 9: 20% of max LR | CODE |
| LOW | docs/release.md | 18 | ## Step 1 - Validate `main` and create the release branch | COMMENT |
| LOW | docs/release.md | 83 | ## Step 2 - Set the RC version | COMMENT |
| LOW | docs/release.md | 104 | ## Step 3 - Stage the RC on TestPyPI | COMMENT |
| LOW | docs/release.md | 144 | ## Step 4 - Validate the RC installation | COMMENT |
| LOW | docs/release.md | 246 | ## Step 5 - Triage an RC validation failure | COMMENT |
| LOW | docs/release.md | 297 | ## Step 6 - Finalize the version | COMMENT |
| LOW | docs/release.md | 312 | ## Step 7 - Cut the GitHub Release and publish to PyPI | COMMENT |
| LOW | docs/release.md | 324 | ## Step 8 - Verify the production PyPI release | COMMENT |
| 40 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 366 | # Step 2: Load tokenizer using our function | COMMENT |
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 372 | # Step 3: Load tokenizer using official Tokenizer library (if available) | COMMENT |
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 379 | # Step 4: Load tokenizer using transformers AutoTokenizer (if available) | COMMENT |
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 388 | # Step 5: Compare underlying tokenizer attributes (only if official tokenizer is available) | COMMENT |
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 394 | # Step 6: Compare with transformers tokenizer if available | COMMENT |
| LOW | tests/unit_tests/cpu/test_tokenizer.py | 351 | # Step 1: Download tokenizer files | COMMENT |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 80 | 0.5, # Step 0: 50% of max LR (warmup) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 81 | 1.0, # Step 1: 100% of max LR (warmup complete) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 82 | 1.0, # Step 2: We maunally added step of stable phase, to prevent LR from dropping to 0 at last step | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 83 | 7.0 / 8.0, # Step 3: 7/8 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 84 | 6.0 / 8.0, # Step 4: 3/4 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 85 | 5.0 / 8.0, # Step 5: 5/8 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 86 | 4.0 / 8.0, # Step 6: 1/2 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 87 | 3.0 / 8.0, # Step 7: 3/8 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 88 | 2.0 / 8.0, # Step 8: 1/4 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 89 | 1.0 / 8.0, # Step 9: 1/8 of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 123 | 0.5, # Step 0: 50% of max LR (warmup) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 124 | 1.0, # Step 1: 100% of max LR (warmup complete) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 125 | 1.0, # Step 2: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 126 | 1.0, # Step 3: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 127 | 1.0, # Step 4: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 128 | 1.0, # Step 5: We maunally added step of stable phase, to prevent LR from dropping to 0 at last step | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 129 | 0.8, # Step 6: Linear decay starts (80% of max LR) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 130 | 0.6, # Step 7: 60% of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 131 | 0.4, # Step 8: 40% of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 132 | 0.2, # Step 9: 20% of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 234 | 0.2, # Step 0: 50% of max LR (warmup) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 235 | 0.4, # Step 1: 100% of max LR (warmup complete) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 236 | 0.6, # Step 2: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 237 | 0.8, # Step 3: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 238 | 1.0, # Step 4: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 271 | 0.5, # Step 0: 50% of max LR (warmup) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 272 | 1.0, # Step 1: 100% of max LR (warmup complete) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 273 | 1.0, # Step 2: We maunally added step of stable phase, to prevent LR from dropping to 0 at last step | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 274 | 1.0, # Step 3: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 275 | 1.0, # Step 4: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 276 | 1.0, # Step 5: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 277 | 1.0, # Step 6: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 278 | 1.0, # Step 7: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 279 | 1.0, # Step 8: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 280 | 1.0, # Step 9: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 314 | 0.2, # Step 0: 50% of max LR (warmup) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 315 | 0.4, # Step 1: 100% of max LR (warmup complete) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 316 | 0.6, # Step 2: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 317 | 0.8, # Step 3: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 318 | 1.0, # Step 4: Stable phase | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 319 | 1.0, # Step 5: We maunally added step of stable phase, to prevent LR from dropping to 0 at last step | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 320 | 0.8, # Step 6: Linear decay starts (80% of max LR) | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 321 | 0.6, # Step 7: 60% of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 322 | 0.4, # Step 8: 40% of max LR | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 323 | 0.2, # Step 9: 20% of max LR | CODE |
| LOW | scripts/checkpoint_compat_test.py | 221 | # Step 1: Reference run at load_commit (full training from scratch) | COMMENT |
| LOW | scripts/checkpoint_compat_test.py | 243 | # Step 2: Save run at save_commit (partial training + checkpoint) | COMMENT |
| LOW | scripts/checkpoint_compat_test.py | 272 | # Step 3: Resume run at load_commit (load checkpoint, train to end) | COMMENT |
| LOW | scripts/checkpoint_compat_test.py | 300 | # Step 4: Compare | COMMENT |
| LOW⚡ | .github/workflows/integration_test_8gpu_torchft.yaml | 32 | # Step 1: Dynamically compute the matrix based on conditions | COMMENT |
| LOW⚡ | .github/workflows/integration_test_8gpu_torchft.yaml | 40 | # Step 2: Use the dynamic matrix in the build-test job | COMMENT |
| LOW⚡ | .github/workflows/integration_test_8gpu_rl.yaml | 25 | # Step 1: Dynamically compute the matrix based on conditions | COMMENT |
| LOW⚡ | .github/workflows/integration_test_8gpu_rl.yaml | 33 | # Step 2: Use the dynamic matrix in the build-test job | COMMENT |
| LOW⚡ | …tegration_test_8gpu_transformers_modeling_backend.yaml | 31 | # Step 1: Dynamically compute the matrix based on conditions | COMMENT |
| 26 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | …h_trace_compaction/scripts/compact_cuda_graph_trace.py | 185 | Extract measured GPU slices that have a profiler stream. Args: events: Trace event list. Returns: | STRING |
| HIGH | …h_trace_compaction/scripts/compact_cuda_graph_trace.py | 466 | Verify slice identity and non-overlap of every compacted lane. Args: original_events: Unmodified trace even | STRING |
| HIGH | …h_trace_compaction/scripts/compact_cuda_graph_trace.py | 750 | Remap GPU flows and annotations, and replace old stream metadata. Args: events: Events whose measured slice | STRING |
| HIGH | …h_trace_compaction/scripts/compact_cuda_graph_trace.py | 813 | Compact CUDA graph streams in a parsed PyTorch profiler trace. Args: trace: Parsed profiler trace. | STRING |
| HIGH | torchtitan/distributed/utils.py | 510 | Clip the gradient norm of an iterable of parameters. Gradient norm clipping requires computing the gradient no | STRING |
| HIGH | torchtitan/distributed/pipeline_parallel.py | 808 | Create pipeline stages based on specified module names for each stage. Also used by Graph PP to split the model int | STRING |
| HIGH | torchtitan/distributed/parallelism_context.py | 410 | Get a device mesh by axis name(s), returning None if not enabled. Args: axes: Names of the mesh axe | STRING |
| HIGH | torchtitan/distributed/parallelism_context.py | 471 | Get a device mesh by axis name(s), raising if not available. Args: axes: Names of the mesh axes. Va | STRING |
| HIGH | …/experiments/transformers_modeling_backend/pipeline.py | 159 | This API creates pipeline stages based on specified module names for each stage. Some model restrictions inclu | STRING |
| HIGH | …experiments/graph_trainer/paged_stash_memory_policy.py | 2853 | Megatron's ``_pp_schedule`` for one rank: every paged layer's forward and backward, in the order the rank runs them. | STRING |
| HIGH | …itan/experiments/graph_trainer/graph_pp/split_di_dw.py | 204 | Split a backward graph into input-gradient and weight-gradient graphs. Contract: Input: backward(save | STRING |
| HIGH | torchtitan/experiments/graph_trainer/graph_pp/runner.py | 1065 | Register graph action handlers on a runtime schedule. Args: schedule (_PipelineScheduleRuntime): Runtime pi | STRING |
| HIGH | …htitan/experiments/graph_trainer/graph_pp/partition.py | 475 | Partition a post-pass GraphTrainer joint graph into GraphPP callables. The caller passes a flat ``minimal_fx_tracer | STRING |
| HIGH | …ments/graph_trainer/graph_pp/split_fsdp_collectives.py | 200 | Extract FSDP parameter all-gather chains from a graph. Contract: unshard(param_shards_and_replicated_params) | STRING |
| HIGH | …ments/graph_trainer/graph_pp/split_fsdp_collectives.py | 407 | Extract FSDP/DDP/HSDP reduce-grad epilogues from a graph. Contract: compute(original_inputs) -> leadi | STRING |
| HIGH | …/experiments/graph_trainer/graph_pp/graph_multiplex.py | 48 | Concatenate backward and forward graphs into one boxed GraphPP callable. Contract: ``OVERLAP_F_B`` schedule a | STRING |
| HIGH | torchtitan/experiments/torchft/diloco/utils.py | 23 | This API creates fragments based on specified module names for each fragment. This method updates the model in | STRING |
| HIGH | torchtitan/models/common/multimodal.py | 139 | Locate each visual item's placeholder run in the token sequence. Args: tokens: ``(T,)`` token IDs. | STRING |
| HIGH | torchtitan/models/common/attention/cp_attention.py | 101 | Prepare a rank-local BlockMask for K/V all-gather CP. The returned mask covers the current rank's Q shard and t | STRING |
| HIGH | torchtitan/hf_datasets/multimodal/utils/video.py | 46 | Load a deterministic uint8 frame array from a NumPy file. Args: path: Path to a ``.npy`` array shaped ``(T, | STRING |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | run_train.sh | 1 | #!/usr/bin/bash | COMMENT |
| LOW | .ci/docker/common/utils.sh | 1 | #!/bin/bash | COMMENT |
| LOW | .ci/docker/common/install_deepep.sh | 1 | #!/bin/bash | COMMENT |
| LOW | tests/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | tests/unit_tests/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | tests/unit_tests/gpu/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | tests/unit_tests/gpu/flex_shard/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | tests/unit_tests/cpu/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | tests/unit_tests/cpu/flex_shard/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | tests/unit_tests/cpu/observability/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | tests/rl/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | tests/rl/unit_tests/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | tests/rl/unit_tests/gpu/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | tests/rl/unit_tests/cpu/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | tests/rl/integration_tests/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | …skills/numerics_debugging/scripts/activation_tracer.py | 341 | # | COMMENT |
| LOW | …skills/numerics_debugging/scripts/activation_tracer.py | 561 | # record_ids populates DebugMode's _tensor_memo with a stable | COMMENT |
| LOW | …skills/numerics_debugging/scripts/activation_tracer.py | 641 | COMMENT | |
| LOW | …/skills/numerics_debugging/scripts/compare_numerics.py | 61 | stats: dict[str, str] = field(default_factory=dict) | COMMENT |
| LOW | …/skills/numerics_debugging/scripts/compare_numerics.py | 181 | COMMENT | |
| LOW | …/skills/numerics_debugging/scripts/compare_numerics.py | 661 | # log_tensor_hashes. Comma-separated lists of floats — we | COMMENT |
| LOW | .github/scripts/run_8xgpu_integration_tests.sh | 1 | #!/bin/bash | COMMENT |
| LOW | .github/scripts/integration_test_real_pg.sh | 1 | #!/usr/bin/env bash | COMMENT |
| LOW | torchtitan_recipes/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan_recipes/graph_trainer/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan_recipes/tests/graph_trainer/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan_recipes/tests/suites/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan_recipes/tests/datasets/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan_recipes/tests/models/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan_recipes/tests/models/kimi_k2_7.py | 541 | # No PP gate: DistMuon is PP-safe. The one precondition -- every stage | COMMENT |
| LOW | torchtitan_recipes/tests/torchft/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan_recipes/tests/rl/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan_recipes/models/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan_recipes/torchft/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan_recipes/rl/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan/tools/utils.py | 141 | else: # for H100 SXM and other variants | COMMENT |
| LOW | torchtitan/tools/utils.py | 181 | # Dot Product Accumulate Systolic (DPAS): | COMMENT |
| LOW | torchtitan/distributed/deepep/deepep.py | 81 | COMMENT | |
| LOW | torchtitan/distributed/deepep/deepep.py | 101 | "-> (Tensor, Tensor, Tensor, Tensor, Tensor)" | COMMENT |
| LOW | torchtitan/experiments/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | …tan/experiments/transformers_modeling_backend/model.py | 781 | except Exception as e: | COMMENT |
| LOW | …periments/transformers_modeling_backend/parallelize.py | 101 | 4. Single model._parallelize(parallelism_context) call -- shards states, wraps forward | COMMENT |
| LOW | …rmers_modeling_backend/tests/test_flex_cp_numerical.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | …iments/transformers_modeling_backend/tests/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | torchtitan/experiments/graph_trainer/make_fx_tracer.py | 581 | # engine dispatches backward to a worker thread that has a | COMMENT |
| LOW | torchtitan/experiments/graph_trainer/__init__.py | 1 | # Copyright (c) Meta Platforms, Inc. and affiliates. | COMMENT |
| LOW | …itan/experiments/graph_trainer/run_train_precompile.sh | 1 | #!/usr/bin/bash | COMMENT |
| LOW | torchtitan/experiments/graph_trainer/fsdp_patterns.py | 441 | # cast_grad = grad.to(reduce_dtype) | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 261 | from torchtitan.components.optim.optimizer import OptimizersContainer | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 541 | # ``[cuda, host]`` freelist state, and the same preallocated reset values so | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 901 | # individual stash rounds up to a page boundary: ``_live_pages`` is the | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 1421 | # Custom ops that page MoE activations in and out of the stash buffers | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 1441 | # dx = paged_stash.reload_next(dx, slot, prefetch) | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 1461 | # only thing that crosses into backward, in place of the padded activation. | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 1481 | # so this only bounds parallelism, not correctness. | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 1961 | # Transformer Engine tagged with ``grouped_tensor_scale_inv`` inside that context | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 2281 | layer_id, | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 2301 | # Rewrite performed for each paged layer:: | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 2321 | # ``num_tokens`` -- a few kilobytes -- are what cross into backward. Under | COMMENT |
| LOW | …experiments/graph_trainer/paged_stash_memory_policy.py | 2761 | # =========================================================================== | COMMENT |
| 40 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | …rchtitan/experiments/graph_trainer/subgraph_regions.py | 97 | if node.op in ("placeholder", "output", "get_attr"): | CODE |
| LOW | …rchtitan/experiments/graph_trainer/subgraph_regions.py | 343 | if any(node.op in ("placeholder", "output") for node in ordered_nodes): | CODE |
| LOW | torchtitan/experiments/graph_trainer/decompositions.py | 49 | placeholders = gm.graph.find_nodes(op="placeholder") | CODE |
| LOW | torchtitan/experiments/graph_trainer/fsdp_passes.py | 119 | placeholders = gm.graph.find_nodes(op="placeholder") | CODE |
| LOW | torchtitan/experiments/graph_trainer/fsdp_passes.py | 171 | if n.all_input_nodes[0].op == "placeholder": | CODE |
| LOW | torchtitan/experiments/graph_trainer/fsdp_passes.py | 1572 | if n in visited or n.op in ("placeholder", "get_attr"): | CODE |
| LOW⚡ | torchtitan/experiments/graph_trainer/memory_policy.py | 147 | gm.graph.find_nodes(op="placeholder") | CODE |
| LOW | torchtitan/experiments/graph_trainer/memory_policy.py | 520 | node for node in required_fw_nodes if node.op in ("placeholder", "get_attr") | CODE |
| LOW | torchtitan/experiments/graph_trainer/memory_policy.py | 594 | if node not in required_fw_nodes or node.op in ("placeholder", "get_attr"): | CODE |
| LOW⚡ | …htitan/experiments/graph_trainer/wgrad_accumulation.py | 137 | or accumulator.op != "placeholder" | CODE |
| LOW⚡ | torchtitan/experiments/graph_trainer/inductor_passes.py | 68 | if node.op == "placeholder" and "val" in node.meta: | CODE |
| LOW⚡ | torchtitan/experiments/graph_trainer/inductor_passes.py | 84 | if node.op == "placeholder": | CODE |
| LOW⚡ | torchtitan/experiments/graph_trainer/inductor_passes.py | 105 | original_placeholders = list(gm.graph.find_nodes(op="placeholder")) | CODE |
| LOW⚡ | torchtitan/experiments/graph_trainer/inductor_passes.py | 114 | if node.op == "placeholder": | CODE |
| LOW | torchtitan/experiments/graph_trainer/inductor_passes.py | 387 | if node.op in ("placeholder", "output"): | CODE |
| LOW | torchtitan/experiments/graph_trainer/fsdp_patterns.py | 228 | for placeholder in gm.graph.find_nodes(op="placeholder"): | CODE |
| LOW | torchtitan/experiments/graph_trainer/fsdp_patterns.py | 317 | if node.op == "placeholder" or node.meta.get("autograd_backward", False): | CODE |
| LOW | …chtitan/experiments/graph_trainer/grad_accumulation.py | 79 | first_compute = next(node for node in gm.graph.nodes if node.op != "placeholder") | CODE |
| LOW⚡ | torchtitan/experiments/graph_trainer/ep_overlap_pass.py | 84 | _GRAPH_BOUNDARY_OPS = {"placeholder", "get_attr"} | CODE |
| LOW | …ph_trainer/spmd_gradient_accumulation_graph_builder.py | 296 | placeholders = call_spec.module.graph.find_nodes(op="placeholder") | CODE |
| LOW | torchtitan/experiments/graph_trainer/debug_utils.py | 36 | if node.op == "placeholder": | CODE |
| LOW | …n/experiments/graph_trainer/tests/test_trace_module.py | 97 | if node.op != "placeholder": | CODE |
| LOW | …n/experiments/graph_trainer/tests/test_trace_module.py | 400 | if node.op == "placeholder" | CODE |
| LOW | …n/experiments/graph_trainer/tests/test_trace_module.py | 655 | if node.op == "placeholder" | CODE |
| LOW | …n/experiments/graph_trainer/tests/test_trace_module.py | 679 | if node.op == "placeholder" | CODE |
| LOW | …n/experiments/graph_trainer/tests/test_trace_module.py | 700 | if node.op == "placeholder" | CODE |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 3582 | self.assertEqual(node.args[0].op, "placeholder") | CODE |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 5073 | if node.op == "placeholder": | CODE |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 5180 | if node.op == "placeholder": | CODE |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 5199 | self.assertEqual(node.args[0].op, "placeholder") | CODE |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 5277 | self.assertEqual(relu_node.args[0].op, "placeholder") | CODE |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 5376 | if node.op == "placeholder": | CODE |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 5476 | if node.op == "placeholder": | CODE |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 5517 | if node.op == "placeholder": | CODE |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 5847 | node.target for node in compile_gm.graph.find_nodes(op="placeholder") | CODE |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 5850 | node.meta for node in compile_gm.graph.find_nodes(op="placeholder") | CODE |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 5887 | node.target for node in wrapped.graph.find_nodes(op="placeholder") | CODE |
| LOW | …xperiments/graph_trainer/tests/test_graph_pp_runner.py | 619 | for node in traced.gm.graph.find_nodes(op="placeholder") | CODE |
| LOW | …xperiments/graph_trainer/tests/test_graph_pp_runner.py | 1630 | for node in gm.graph.find_nodes(op="placeholder"): | CODE |
| LOW | …xperiments/graph_trainer/tests/test_graph_pp_runner.py | 1786 | bw_gm.graph.find_nodes(op="placeholder")[0].meta["val"] = bw_fake | CODE |
| LOW | …xperiments/graph_trainer/tests/test_graph_pp_runner.py | 1787 | fw_gm.graph.find_nodes(op="placeholder")[0].meta["val"] = fw_fake | CODE |
| LOW | …xperiments/graph_trainer/tests/test_graph_pp_runner.py | 1802 | placeholders = multiplexed.graph.find_nodes(op="placeholder") | CODE |
| LOW | …xperiments/graph_trainer/tests/test_graph_pp_runner.py | 1846 | fw_gm.graph.find_nodes(op="placeholder")[0].meta["val"] = fw_fake | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2120 | sharded_param, replicated_param, _ = gm.graph.find_nodes(op="placeholder") | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2132 | sharded_param = gm.graph.find_nodes(op="placeholder")[0] | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2357 | sharded_param = gm.graph.find_nodes(op="placeholder")[0] | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2366 | sharded_param, reconstruction_input = graph.find_nodes(op="placeholder") | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2392 | sharded_param = gm.graph.find_nodes(op="placeholder")[0] | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2403 | sharded_param, unrelated = gm.graph.find_nodes(op="placeholder") | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2413 | sharded_param = graph.find_nodes(op="placeholder")[0] | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2441 | sharded_param = gm.graph.find_nodes(op="placeholder")[0] | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2453 | sharded_param = gm.graph.find_nodes(op="placeholder")[0] | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2502 | sharded_param = gm.graph.find_nodes(op="placeholder")[0] | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2514 | sharded_param, unrelated = gm.graph.find_nodes(op="placeholder") | CODE |
| LOW⚡ | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2526 | sharded_param = graph.find_nodes(op="placeholder")[0] | CODE |
| LOW | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 345 | node.name: node for node in gm.graph.find_nodes(op="placeholder") | CODE |
| LOW | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 389 | node.name: node for node in view_gm.graph.find_nodes(op="placeholder") | CODE |
| LOW | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 1261 | return tuple(node.name for node in gm.graph.find_nodes(op="placeholder")) | CODE |
| LOW | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2223 | sharded_param = gm.graph.find_nodes(op="placeholder")[0] | CODE |
| LOW | …xperiments/graph_trainer/tests/test_graph_pp_passes.py | 2328 | sharded_param = gm.graph.find_nodes(op="placeholder")[0] | CODE |
| 30 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | tests/unit_tests/cpu/test_set_determinism.py | 46 | # Create a world_mesh mock | COMMENT |
| MEDIUM⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 27 | # Create a simple model with parameters | COMMENT |
| MEDIUM⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 29 | # Create an optimizer | COMMENT |
| MEDIUM⚡ | tests/unit_tests/cpu/test_lr_scheduler.py | 36 | # Create an optimizer container | COMMENT |
| MEDIUM | tests/unit_tests/cpu/test_lr_scheduler.py | 63 | # Create a job config with 10 steps, 2 warmup steps, and linear decay | COMMENT |
| MEDIUM | tests/unit_tests/cpu/test_lr_scheduler.py | 106 | # Create a job config with 10 steps, 2 warmup steps, 3 stable steps, and 5 decay steps | COMMENT |
| MEDIUM | tests/unit_tests/cpu/test_lr_scheduler.py | 148 | # Create a job config with a minimum learning rate | COMMENT |
| MEDIUM | tests/unit_tests/cpu/test_lr_scheduler.py | 217 | # Create a job config where warmup steps > training steps | COMMENT |
| MEDIUM | tests/unit_tests/cpu/test_lr_scheduler.py | 254 | # Create a job config with 10 steps, 2 warmup steps, and no decay phase | COMMENT |
| MEDIUM | tests/unit_tests/cpu/test_lr_scheduler.py | 296 | # Create a job config where warmup + decay steps > training steps | COMMENT |
| MEDIUM | tests/unit_tests/cpu/test_checkpoint.py | 978 | # Create a fake model with freqs_cis and other parameters | COMMENT |
| MEDIUM | tests/assets/cc12m_test/pack_test_dataset.py | 23 | # Create a TarWriter object to write the dataset to a tar archive | COMMENT |
| MEDIUM⚡ | scripts/loss_compare.py | 334 | # Create the output folder | COMMENT |
| MEDIUM⚡ | scripts/loss_compare.py | 903 | # Create a test case | COMMENT |
| MEDIUM | …ripts/checkpoint_conversion/numerical_tests_example.py | 151 | # Define the set of outputs to test loss for | COMMENT |
| MEDIUM | .github/scripts/update_version.sh | 6 | # Create a variable named __version__ at the end of __init__.py | COMMENT |
| MEDIUM | torchtitan/distributed/activation_checkpoint.py | 7 | # This file provides the util functions to apply activation checkpointing to the model. | COMMENT |
| MEDIUM | torchtitan/distributed/pipeline_parallel.py | 614 | # Create a set of modules to keep for faster lookup | COMMENT |
| MEDIUM | …/experiments/transformers_modeling_backend/pipeline.py | 199 | # Create a set of modules to keep for faster lookup | COMMENT |
| MEDIUM | …chtitan/experiments/graph_trainer/tests/test_passes.py | 1039 | # Create a second PG to simulate expert-FSDP. | COMMENT |
| MEDIUM | …experiments/graph_trainer/tests/test_custom_codegen.py | 206 | # Create main module | COMMENT |
| MEDIUM | torchtitan/experiments/torchft/manager.py | 217 | # Create the outer optimizer based on the inner optimizer parameters. | COMMENT |
| MEDIUM | torchtitan/experiments/torchft/diloco/utils.py | 52 | # Create a set of modules to keep for faster lookup | COMMENT |
| MEDIUM | torchtitan/models/common/moe.py | 172 | """This class implements token-choice routing. In token-choice top-K routing, each token is | STRING |
| MEDIUM | torchtitan/rl/rollout/environment/token.py | 227 | # Create a new list to avoid mutating previous states | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | tests/unit_tests/cpu/test_tp_kv_heads_validation.py | 40 | except Exception: | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 376 | except Exception as e: | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_tokenizer.py | 385 | except Exception as e: | CODE |
| LOW⚡ | tests/unit_tests/cpu/test_checkpoint.py | 1129 | except Exception as e: | CODE |
| LOW | tests/integration_tests/run_tests.py | 441 | except Exception as e: | CODE |
| LOW | tests/integration_tests/run_tests.py | 455 | except Exception as e: | CODE |
| LOW | tests/integration_tests/flux.py | 125 | except Exception as e: | CODE |
| LOW | tests/rl/unit_tests/cpu/test_muse_glimmer_renderer.py | 220 | except Exception as exc: # offline, or no access to the Hub | CODE |
| MEDIUM⚡ | scripts/loss_compare.py | 309 | log_print(f"Error: Import file does not exist: {import_result}") | CODE |
| MEDIUM⚡ | scripts/loss_compare.py | 330 | log_print(f"Error: Output folder '{output_folder}' already exists") | CODE |
| MEDIUM⚡ | scripts/loss_compare.py | 892 | log_print(f"Error: No {metric_name} values found in baseline") | CODE |
| MEDIUM⚡ | scripts/loss_compare.py | 896 | log_print(f"Error: No {metric_name} values found in test") | CODE |
| MEDIUM⚡ | scripts/loss_compare.py | 900 | log_print(f"Error: No {metric_name} values found in import file") | CODE |
| MEDIUM | scripts/loss_compare.py | 276 | log_print("Error: All settings are identical") | CODE |
| MEDIUM | scripts/loss_compare.py | 287 | log_print(f"Error: --steps must be a positive integer, got: {steps}") | CODE |
| MEDIUM | scripts/loss_compare.py | 292 | log_print("Error: --import-result requires --assert-equal") | CODE |
| MEDIUM | scripts/loss_compare.py | 782 | log_print("Error: baseline-only mode requires --import-result") | CODE |
| MEDIUM | scripts/loss_compare.py | 794 | log_print(f"Error: metric '{name}' missing from {other_name} results") | CODE |
| MEDIUM | scripts/rl/loss_compare.py | 302 | log_print("Error: no losses found in TensorBoard events") | CODE |
| LOW | .github/scripts/todo_debt_tracker.py | 45 | except Exception: | CODE |
| LOW | .github/scripts/todo_debt_tracker.py | 116 | except Exception as e: | CODE |
| LOW | .github/scripts/todo_debt_tracker.py | 183 | except Exception: | CODE |
| LOW | .github/scripts/todo_debt_tracker.py | 303 | except Exception as e: | CODE |
| MEDIUM | .github/scripts/todo_debt_tracker.py | 304 | print(f"Error reading {file_path}: {e}", file=sys.stderr) | CODE |
| LOW | torchtitan/__init__.py | 11 | except Exception: | CODE |
| LOW | torchtitan/train.py | 59 | except Exception: | CODE |
| LOW | torchtitan/trainer.py | 237 | except Exception: | CODE |
| LOW | …tan/experiments/transformers_modeling_backend/model.py | 781 | except Exception as e: | CODE |
| LOW | …_moe_model/scripts/numerical_equivalence_pretrained.py | 47 | except Exception: | CODE |
| LOW | …_moe_model/scripts/numerical_equivalence_pretrained.py | 192 | except Exception as e: | CODE |
| LOW | …/skills/add_moe_model/scripts/numerical_equivalence.py | 572 | except Exception as e: | CODE |
| LOW | …/skills/add_moe_model/scripts/numerical_equivalence.py | 622 | except Exception as e: | CODE |
| LOW | torchtitan/experiments/graph_trainer/custom_codegen.py | 67 | except Exception: | CODE |
| LOW | torchtitan/experiments/graph_trainer/cpu_offload.py | 174 | except Exception: | CODE |
| LOW | torchtitan/experiments/graph_trainer/cpu_offload.py | 186 | except Exception: | CODE |
| LOW | …eriments/graph_trainer/log_activation_memory_policy.py | 77 | except Exception: | CODE |
| LOW | torchtitan/experiments/graph_trainer/graph_pp/utils.py | 203 | except Exception as exc: | CODE |
| LOW | torchtitan/experiments/graph_trainer/graph_pp/utils.py | 223 | except Exception as exc: | CODE |
| LOW | torchtitan/models/common/dist_moe/runtime.py | 413 | except Exception: | CODE |
| LOW | torchtitan/observability/metrics.py | 491 | except Exception as e: | CODE |
| LOW | torchtitan/observability/structured_logger/README.md | 151 | except Exception: | CODE |
| LOW | torchtitan/components/data/loader.py | 187 | except Exception: | CODE |
| LOW | torchtitan/components/checkpointer/base.py | 68 | except Exception as error: | CODE |
| LOW | torchtitan/hf_datasets/multimodal/utils/video.py | 157 | except Exception as e: | CODE |
| LOW | torchtitan/hf_datasets/multimodal/utils/image.py | 280 | except Exception as e: | CODE |
| LOW | torchtitan/rl/controller.py | 433 | except Exception: | CODE |
| LOW | torchtitan/rl/controller.py | 438 | except Exception: | CODE |
| LOW⚡ | torchtitan/rl/controller.py | 456 | except Exception: | CODE |
| LOW⚡ | torchtitan/rl/controller.py | 461 | except Exception: | CODE |
| LOW⚡ | torchtitan/rl/controller.py | 467 | except Exception: | CODE |
| LOW | torchtitan/rl/controller.py | 997 | except Exception: | CODE |
| LOW | torchtitan/rl/generator.py | 1369 | except Exception as exc: | CODE |
| LOW | torchtitan/rl/generator.py | 1601 | except Exception: | CODE |
| LOW | torchtitan/rl/observability/vllm.py | 263 | except Exception as e: | CODE |
| LOW | torchtitan/rl/observability/vllm.py | 368 | except Exception as e: | CODE |
| LOW | torchtitan/rl/observability/metrics/processor.py | 220 | except Exception: | CODE |
| LOW | torchtitan/rl/observability/metrics/processor.py | 231 | except Exception: | CODE |
| LOW | torchtitan/rl/rollout/rollouter.py | 466 | except Exception: | CODE |
| LOW | torchtitan/rl/rollout/environment/token.py | 166 | except Exception: | CODE |
| LOW | torchtitan/rl/examples/search_r1/env.py | 91 | except Exception as exc: | CODE |
| 3 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | torchtitan/tools/filesystem.py | 95 | except FileNotFoundError: | CODE |
| HIGH | torchtitan/experiments/graph_trainer/custom_codegen.py | 155 | except SyntaxError: | CODE |
| HIGH | torchtitan/experiments/graph_trainer/custom_codegen.py | 629 | except AttributeError: | CODE |
| HIGH | torchtitan/experiments/graph_trainer/custom_codegen.py | 70 | except OSError: | CODE |
| HIGH | torchtitan/experiments/graph_trainer/memory_policy.py | 107 | except AttributeError: | CODE |
| HIGH | torchtitan/experiments/graph_trainer/common_utils.py | 398 | except ValueError: | CODE |
| HIGH | …chtitan/experiments/graph_trainer/tests/test_passes.py | 3305 | except ValueError: | CODE |
| HIGH | torchtitan/rl/generator.py | 683 | except (asyncio.CancelledError, Exception): | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | tests/unit_tests/cpu/test_override.py | 477 | parse_cli_imports(['mod={"block_size": 256, "flag": null}']), | CODE |
| HIGH | …/skills/numerics_debugging/scripts/compare_numerics.py | 1016 | const hidden = el.dataset.hideByCategory || el.dataset.hideByOpFilter; | CODE |
| HIGH | …/skills/numerics_debugging/scripts/compare_numerics.py | 1028 | let _opFilterTimer = null; | CODE |
| HIGH | …/skills/numerics_debugging/scripts/compare_numerics.py | 1037 | let re = null; | CODE |
| HIGH | scripts/loss_compare.py | 437 | log_print(" - To commit: git add -A && git commit -m 'message'") | CODE |
| HIGH⚡ | …ripts/checkpoint_conversion/numerical_tests_qwen3_5.py | 53 | """KL(softmax(b) || softmax(a)) — F.kl_div(log Q, P) computes KL(P || Q).""" | STRING |
| HIGH | …htitan/rl/examples/verifiers/terminal_bench/harness.py | 118 | 'tmux kill-server >/dev/null 2>&1 || true; rm -rf "$TMUX_TMPDIR"', | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| CRITICAL | tests/unit_tests/cpu/test_validate.py | 312 | assert validator.pp_schedule.eval.call_args.kwargs.get("return_outputs") is False | CODE |
| CRITICAL | …n/experiments/graph_trainer/tests/test_trace_module.py | 2274 | code_lines = trainer.engine._traced_step.gm.graph.python_code( | CODE |
| CRITICAL | …n/experiments/graph_trainer/tests/test_trace_module.py | 2309 | code = trainer.engine._traced_step.gm.graph.python_code("self").src | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | tests/unit_tests/gpu/test_helion_rope.py | 28 | # tests below stay robust if a sibling test (e.g. test_override.py) calls | COMMENT |
| MEDIUM | tests/unit_tests/cpu/test_tokenizer.py | 178 | # 2. Compare vocabularies with more comprehensive sampling | COMMENT |
| MEDIUM | tests/unit_tests/cpu/test_tokenizer.py | 203 | # Test a random sample of tokens (more comprehensive than just common words) | COMMENT |
| MEDIUM | torchtitan/experiments/graph_trainer/trainer.py | 51 | # harness builds without running initialization -- see it as disabled. | COMMENT |
| MEDIUM | torchtitan/components/tokenizer.py | 378 | # Process added_tokens_decoder (comprehensive special token definitions) | COMMENT |
| MEDIUM | torchtitan/rl/distributed/routing/strategies.py | 198 | # easy to implement, but not robust for all scenarios. Revisit with an | COMMENT |
| MEDIUM | torchtitan/rl/distributed/routing/strategies.py | 199 | # more robust approach. | COMMENT |
| MEDIUM | torchtitan/rl/examples/verifiers/rollouter.py | 290 | # One VerifiersEnvClient.run executes a complete rollout. The harness | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | tests/unit_tests/gpu/test_qwen3_5_deltanet.py | 350 | CODE | |
| LOW | torchtitan_recipes/overrides/fused_mla.py | 842 | CODE | |
| LOW | torchtitan/distributed/fsdp.py | 225 | CODE | |
| LOW | torchtitan/distributed/flex_shard/dist_muon.py | 168 | CODE | |
| LOW | …periments/transformers_modeling_backend/parallelize.py | 198 | CODE | |
| LOW | …formers_modeling_backend/tests/test_moe_parallelism.py | 91 | CODE | |
| LOW | …formers_modeling_backend/tests/test_moe_parallelism.py | 127 | CODE | |
| LOW | …ents/graph_trainer/tests/test_bitwise_deterministic.py | 189 | CODE | |
| LOW | …chtitan/experiments/graph_trainer/tests/test_passes.py | 391 | CODE | |
| LOW | …experiments/graph_trainer/tests/_trainer_test_utils.py | 77 | CODE | |
| LOW | …riments/graph_trainer/tests/test_wgrad_accumulation.py | 55 | CODE | |
| LOW | torchtitan/models/kimi_k2_7/model.py | 266 | CODE | |
| LOW | torchtitan/models/kimi_k3/model.py | 616 | CODE | |
| LOW | torchtitan/models/common/decoder.py | 287 | CODE | |
| LOW | torchtitan/models/muse_glimmer/model.py | 546 | CODE | |
| LOW | torchtitan/models/muse_glimmer/flavors.py | 276 | CODE | |
| LOW | torchtitan/models/qwen3_5/model.py | 696 | CODE | |
| LOW | torchtitan/quantization/nvfp4/linear.py | 525 | CODE | |
| LOW | torchtitan/quantization/mxfp8/linear.py | 500 | CODE | |
| LOW | torchtitan/hf_datasets/multimodal/utils/image.py | 202 | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | tests/unit_tests/cpu/test_checkpoint.py | 138 | # Check if this is async_with_pinned_mem mode by looking for async_stager parameter | COMMENT |
| LOW | tests/unit_tests/cpu/test_checkpoint.py | 919 | cfg.interval = 10 # Set interval to 10 so step 1 wouldn't normally trigger save | CODE |
| LOW | tests/assets/cc12m_test/pack_test_dataset.py | 38 | # Write the file and its metadata to the TarWriter | COMMENT |
| LOW⚡ | scripts/loss_compare.py | 328 | # Check if output folder already exists | COMMENT |
| LOW | scripts/loss_compare.py | 740 | # Check if losses were extracted successfully | COMMENT |
| LOW | scripts/loss_compare.py | 1269 | # Check if git working directory is clean before switching commits | COMMENT |
| LOW | …chtitan/experiments/torchft/tests/integration_tests.py | 104 | # Check if we have enough GPUs | COMMENT |
| LOW | torchtitan/models/flux/validate.py | 148 | # Set model to eval mode | COMMENT |
| LOW | torchtitan/models/flux/model/state_dict_adapter.py | 41 | # Check if safetensors index file exists | COMMENT |
| LOW | torchtitan/observability/metrics.py | 263 | # Check if the loss-visible rank is included in LOG_RANK environment variable | COMMENT |
| LOW | torchtitan/observability/metrics.py | 443 | # Check if any logging backend is enabled | COMMENT |
| LOW | torchtitan/components/validate.py | 153 | # Set model to eval mode | COMMENT |
| LOW | torchtitan/rl/generate.py | 236 | # Print results | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | tests/rl/unit_tests/cpu/test_muse_glimmer_renderer.py | 358 | # Header scaffolding must never be trainable. | COMMENT |
| MEDIUM | tests/rl/unit_tests/cpu/test_muse_glimmer_renderer.py | 396 | # ...and header scaffolding is not content on any role. | COMMENT |
| MEDIUM | torchtitan/rl/examples/search_r1/README.md | 1 | # Search-R1: multi-turn retrieval-augmented GRPO | COMMENT |
| MEDIUM | torchtitan/rl/model/muse_glimmer/renderer.py | 472 | # emits its own stop token; on history roles it is template scaffolding. | COMMENT |
| MEDIUM | torchtitan/rl/model/muse_glimmer/renderer.py | 624 | # header scaffolding is excluded on every role. | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | torchtitan/rl/generator.py | 711 | # The loop now executes the decision. In this example: STEP. | STRING |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | torchtitan/rl/examples/verifiers/data.py | 46 | # TODO: implement this as a SourceConfig and reuse SingleDatasetConfig once | COMMENT |