pytorch/executorch

11.4

Adjusted Score

11.4

Raw Score

100%

Time Factor

2026-07-14

Last Push

4.8K

Stars

Python

Language

1.4M

Lines of Code

7.1K

Files

11.9K

Pattern Hits

2026-07-14

Scan Date

0.06

HC Hit Rate

What These Metrics Mean

Adjusted Score: Primary synthetic code indicator. Raw score normalised per 1,000 lines of code and multiplied by the temporal discount factor. This is the definitive comparative metric — use it to rank repositories by AI authorship density.
Raw Score: The unmodified sum of all severity-weighted, context-multiplied pattern match scores before temporal discounting. Reflects the absolute signal strength independent of when the repository was last active.
Time Factor: The temporal discount multiplier (0–100%) applied to the raw score. Repositories last updated before ChatGPT's launch (Nov 2022) receive a 5% factor. Full signal is only assigned to repositories active in the post-adoption era (Jan 2024+).
Pattern Hits: Total count of individual pattern matches across all files and categories. A high hit count with a low score may indicate a very large codebase with isolated AI snippets; a low count with a high score indicates dense, concentrated AI signatures.
HC Hit Rate: High+Critical pattern hits per file, averaged across the repository. This orthogonal signal catches repositories where a few files are densely packed with high-severity AI tells — a strong indicator even when the normalised score appears moderate due to codebase size.
Lines of Code / Files: Total lines and files analysed. The scanner examines 94 file extensions. These denominators are used to normalise the score, enabling fair comparison between repositories of vastly different sizes.

Score History

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.

Severity Breakdown

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.

CRITICAL 28HIGH 429MEDIUM 1210LOW 10250

Directory Score Breakdown

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.

Pattern Findings

The scanner identified 11917 distinct pattern matches across 25 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.

Hyper-Verbose Identifiers5276 hits · 5024 pts

Severity	File	Line	Snippet	Context
LOW	install_executorch.py	111	def get_required_submodule_paths():	CODE
LOW	install_executorch.py	132	def check_and_update_submodules():	CODE
LOW	install_requirements.py	109	def install_optional_example_requirements(use_pytorch_nightly):	CODE
LOW	setup.py	615	def copy_extensions_to_source(self) -> None:	CODE
LOW	install_utils.py	164	def _cuda_version_to_pytorch_suffix(major, minor):	CODE
LOW	kernels/test/gen_supported_features.py	51	def generate_header_entry_text(namespace: str, feature: str, properties: dict):	CODE
LOW	kernels/test/gen_supported_features.py	93	def generate_definition_entry(d: dict):	STRING
LOW	kernels/test/gen_supported_features.py	105	def generate_definition_entry_text(namespace: str, feature: str, value: Any):	STRING
LOW⚡	kernels/test/test_bitwise_shift.py	115	def test_left_shift_tensor_tensor(self):	CODE
LOW⚡	kernels/test/test_bitwise_shift.py	122	def test_left_shift_tensor_scalar(self):	CODE
LOW⚡	kernels/test/test_bitwise_shift.py	128	def test_right_shift_tensor_tensor(self):	CODE
LOW⚡	kernels/test/test_bitwise_shift.py	135	def test_right_shift_tensor_scalar(self):	CODE
LOW⚡	kernels/test/test_case_gen.py	36	def make_out_dynamic_shape_bound_shape_same(tensor: torch.Tensor):	CODE
LOW⚡	kernels/test/test_case_gen.py	43	def make_out_dynamic_shape_bound_shape_larger(tensor: torch.Tensor):	CODE
LOW⚡	kernels/test/test_case_gen.py	51	def make_out_dynamic_shape_unbound_shape(tensor: torch.Tensor):	CODE
LOW	kernels/test/test_case_gen.py	78	def make_test_cases_dynamic_shape(*args):	CODE
LOW	kernels/test/test_case_gen.py	104	def make_test_cases_broadcast_two_input_tensor(x, y, cpp_args, torch_args, torch_fn):	CODE
LOW	kernels/test/test_case_gen.py	282	def make_simple_generated_case(*args, torch_fn):	CODE
LOW	kernels/test/test_case_gen.py	362	def gen_test_case_op_as_strided_copy():	STRING
LOW	kernels/test/test_case_gen.py	367	def gen_test_case_op_bitwise_not():	STRING
LOW	kernels/test/test_case_gen.py	436	def gen_test_case_op_detach_copy():	STRING
LOW	kernels/test/test_case_gen.py	457	def gen_test_case_op_full_like():	STRING
LOW	kernels/test/test_case_gen.py	520	def gen_test_case_op_log_softmax():	STRING
LOW	kernels/test/test_case_gen.py	661	def gen_test_case_op_transpose():	STRING
LOW	kernels/test/test_case_gen.py	666	def gen_test_case_op_unsqueeze():	STRING
LOW	kernels/test/test_case_gen.py	724	def gen_test_case_op_embedding():	STRING
LOW	kernels/test/test_case_gen.py	734	def gen_test_case_op_floor_divide():	STRING
LOW	kernels/test/test_case_gen.py	791	def gen_test_case_op_select_scatter():	STRING
LOW	kernels/test/test_case_gen.py	847	def gen_test_case_op_convolution():	STRING
LOW	kernels/test/test_case_gen.py	857	def gen_test_case_op_masked_fill():	STRING
LOW	kernels/quantized/test/test_quant_dequant_per_token.py	34	def test_quantize_per_token_large_tensor(self):	CODE
LOW	kernels/quantized/test/test_quant_dequant_per_token.py	47	def test_quantize_per_token_high_rank(self):	CODE
LOW	kernels/quantized/test/test_quant_dequant_per_token.py	60	def test_quantize_per_token_dynamic(self):	CODE
LOW	kernels/quantized/test/test_quant_dequant_per_token.py	85	def test_dequantize_per_token(self):	CODE
LOW	kernels/quantized/test/test_quant_dequant_per_token.py	100	def test_dequantize_per_token_large_tensor(self):	CODE
LOW	kernels/quantized/test/test_quant_dequant_per_token.py	113	def test_dequantize_per_token_high_rank(self):	CODE
LOW	kernels/quantized/test/test_quant_dequant_per_token.py	126	def test_dequantize_per_token_dynamic(self):	CODE
LOW⚡	kernels/quantized/test/test_out_variants.py	30	def test_choose_qparams_tensor_to_out_variant(self) -> None:	CODE
LOW⚡	kernels/quantized/test/test_out_variants.py	38	def test_dequantize_per_tensor_to_out_variant(self) -> None:	CODE
LOW⚡	kernels/quantized/test/test_out_variants.py	46	def test_dequantize_per_tensor_tensor_to_out_variant(self) -> None:	CODE
LOW⚡	kernels/quantized/test/test_out_variants.py	56	def test_dequantize_per_channel_to_out_variant(self) -> None:	CODE
LOW⚡	kernels/quantized/test/test_out_variants.py	64	def test_mixed_linear_to_out_variant(self) -> None:	CODE
LOW⚡	kernels/quantized/test/test_out_variants.py	70	def test_mixed_mm_to_out_variant(self) -> None:	CODE
LOW⚡	kernels/quantized/test/test_out_variants.py	76	def test_quantize_per_tensor_to_out_variant(self) -> None:	CODE
LOW⚡	kernels/quantized/test/test_out_variants.py	84	def test_quantize_per_tensor_tensor_to_out_variant(self) -> None:	CODE
LOW⚡	kernels/quantized/test/test_out_variants.py	94	def test_quantize_per_channel_to_out_variant(self) -> None:	CODE
LOW	…rnels/portable/test/test_grid_sampler_2d_executorch.py	119	def test_all_mode_combinations(self):	CODE
LOW	…rnels/portable/test/test_grid_sampler_2d_executorch.py	361	def test_exact_boundary_coordinates(self):	CODE
LOW	…rnels/portable/test/test_grid_sampler_2d_executorch.py	397	def test_out_of_bounds_values_in_grid(self):	CODE
LOW	tools/cmake/common/preset_test.py	68	def test_define_overridable_option_invalid_name(self):	STRING
LOW	tools/cmake/common/preset_test.py	80	def test_define_overridable_option_default(self):	STRING
LOW	tools/cmake/common/preset_test.py	103	def test_define_overridable_option_invalid_type(self):	STRING
LOW	tools/cmake/common/preset_test.py	115	def test_define_overridable_option_cli_override(self):	STRING
LOW	tools/cmake/common/preset_test.py	136	def test_define_overridable_option_set_override_before(self):	STRING
LOW	tools/cmake/common/preset_test.py	158	def test_define_overridable_option_set_override_after(self):	STRING
LOW	tools/cmake/common/preset_test.py	181	def test_define_overridable_option_set_override_after_with_cache(self):	STRING
LOW	tools/cmake/common/preset_test.py	204	def test_define_overridable_option_override_existing_cache_with_cli(self):	STRING
LOW	tools/cmake/common/preset_test.py	228	def test_set_overridable_option_before(self):	STRING
LOW	tools/cmake/common/preset_test.py	252	def test_set_overridable_option_after(self):	STRING
LOW	tools/cmake/common/preset_test.py	274	def test_set_overridable_option_loaded_from_file(self):	STRING
5216 more matches not shown…

Decorative Section Separators666 hits · 2008 pts

Severity	File	Line	Snippet	Context
MEDIUM⚡	kernels/test/test_bitwise_shift.py	111	# ==========================================================================	COMMENT
MEDIUM⚡	kernels/test/test_bitwise_shift.py	113	# ==========================================================================	COMMENT
MEDIUM⚡	kernels/test/test_bitwise_shift.py	141	# ==========================================================================	COMMENT
MEDIUM⚡	kernels/test/test_bitwise_shift.py	143	# ==========================================================================	COMMENT
MEDIUM	…a/org/pytorch/executorch/LlmLoraInstrumentationTest.kt	65	// ─── dataFiles constructor variants ─────────────────────────────────────────	COMMENT
MEDIUM	…a/org/pytorch/executorch/LlmLoraInstrumentationTest.kt	111	// ─── LlmModuleConfig with dataPath ──────────────────────────────────────────	COMMENT
MEDIUM	…a/org/pytorch/executorch/LlmLoraInstrumentationTest.kt	154	// ─── Invalid data file paths ────────────────────────────────────────────────	COMMENT
MEDIUM	…a/org/pytorch/executorch/LlmLoraInstrumentationTest.kt	200	// ─── Baseline equivalence ───────────────────────────────────────────────────	COMMENT
MEDIUM	…a/org/pytorch/executorch/LlmLoraInstrumentationTest.kt	227	// ─── LlmModuleConfig builder validation ─────────────────────────────────────	COMMENT
MEDIUM	…a/org/pytorch/executorch/LlmLoraInstrumentationTest.kt	268	// ─── Helpers ────────────────────────────────────────────────────────────────	COMMENT
MEDIUM	…org/pytorch/executorch/AsrModuleInstrumentationTest.kt	42	// ─── Constructor validation ─────────────────────────────────────────────────	COMMENT
MEDIUM	…org/pytorch/executorch/AsrModuleInstrumentationTest.kt	102	// ─── Lifecycle ──────────────────────────────────────────────────────────────	COMMENT
MEDIUM	…org/pytorch/executorch/AsrModuleInstrumentationTest.kt	146	// ─── Transcribe validation ──────────────────────────────────────────────────	COMMENT
MEDIUM	…org/pytorch/executorch/AsrModuleInstrumentationTest.kt	161	// ─── AsrTranscribeConfig ────────────────────────────────────────────────────	COMMENT
MEDIUM	…org/pytorch/executorch/AsrModuleInstrumentationTest.kt	225	// ─── Helpers ────────────────────────────────────────────────────────────────	COMMENT
MEDIUM	…Test/java/org/pytorch/executorch/LlmPerformanceTest.kt	219	// ─── LlmCallback ──────────────────────────────────────────────────────────────────	COMMENT
MEDIUM	…Test/java/org/pytorch/executorch/LlmPerformanceTest.kt	241	// ─── Helpers ─────────────────────────────────────────────────────────────────────	COMMENT
MEDIUM	.ci/scripts/test_torchao_huggingface_checkpoints.sh	4	# -------------------------	COMMENT
MEDIUM	.ci/scripts/test_torchao_huggingface_checkpoints.sh	6	# -------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	12	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	14	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	87	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	89	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	108	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	110	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	121	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	123	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	134	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	136	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	205	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	207	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	222	# ----------------------------	COMMENT
MEDIUM	.ci/scripts/test_wheel_package_qnn.sh	224	# ----------------------------	COMMENT
MEDIUM⚡	backends/apple/metal/tests/test_modules.py	1352	# =============================================================================	COMMENT
MEDIUM⚡	backends/apple/metal/tests/test_modules.py	1354	# =============================================================================	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	138	# =============================================================================	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	140	# =============================================================================	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	155	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	157	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	172	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	191	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	193	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	212	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	229	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	254	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	279	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	304	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	329	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	354	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	356	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	384	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	410	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	436	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	438	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	460	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	484	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	534	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	556	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	576	# -------------------------------------------------------------------------	COMMENT
MEDIUM	backends/apple/metal/tests/test_modules.py	597	# -------------------------------------------------------------------------	COMMENT
606 more matches not shown…

Cross-File Repetition283 hits · 1415 pts

Severity	File	Line	Snippet	Context
HIGH	extension/module/test/resources/gen_bundled_program.py	0	an example model with multi-methods. each method has multiple input and single output	STRING
HIGH	docs/source/bundled-io.md	0	an example model with multi-methods. each method has multiple input and single output	STRING
HIGH	devtools/bundled_program/util/test_util.py	0	an example model with multi-methods. each method has multiple input and single output	STRING
HIGH	backends/apple/mps/mps_preprocess.py	0	returns the padding required to align `offset` to `alignment`.	STRING
HIGH	…kends/xnnpack/serialization/xnnpack_graph_serialize.py	0	returns the padding required to align `offset` to `alignment`.	STRING
HIGH	exir/_serialize/padding.py	0	returns the padding required to align `offset` to `alignment`.	STRING
HIGH	backends/apple/mps/test/test_mps_utils.py	0	helper testing function that asserts that the model output and the reference output are equal with some tolerance. due t	STRING
HIGH	backends/vulkan/test/test_vulkan_delegate.py	0	helper testing function that asserts that the model output and the reference output are equal with some tolerance. due t	STRING
HIGH	backends/xnnpack/test/test_xnnpack_utils.py	0	helper testing function that asserts that the model output and the reference output are equal with some tolerance. due t	STRING
HIGH	backends/apple/mps/test/test_mps_utils.py	0	helper testing function that takes a torch.nn.module and lowers it to xnnpack with the given sample inputs. it then runs	STRING
HIGH	backends/vulkan/test/test_vulkan_delegate.py	0	helper testing function that takes a torch.nn.module and lowers it to xnnpack with the given sample inputs. it then runs	STRING
HIGH	backends/xnnpack/test/test_xnnpack_utils.py	0	helper testing function that takes a torch.nn.module and lowers it to xnnpack with the given sample inputs. it then runs	STRING
HIGH	backends/apple/mps/operators/node_visitor.py	0	create a new class instance at runtime, and put them in a dict	STRING
HIGH	backends/qualcomm/builders/node_visitor_manager.py	0	create a new class instance at runtime, and put them in a dict	STRING
HIGH	backends/samsung/builders/node_visitor.py	0	create a new class instance at runtime, and put them in a dict	STRING
HIGH	backends/xnnpack/operators/node_visitor.py	0	create a new class instance at runtime, and put them in a dict	STRING
HIGH	backends/apple/mps/utils/mps_utils.py	0	returns true if the given node is a get attr node for a tensor of the model.	STRING
HIGH	backends/transforms/utils.py	0	returns true if the given node is a get attr node for a tensor of the model.	STRING
HIGH	backends/xnnpack/utils/utils.py	0	returns true if the given node is a get attr node for a tensor of the model.	STRING
HIGH	backends/arm/_passes/arm_pass_utils.py	0	returns true if the given node is a get attr node for a tensor of the model.	STRING
HIGH	…ends/apple/coreml/quantizer/_coreml_quantizer_utils.py	0	given a list of nodes (that represents an operator pattern), return true if any of the node is annotated, otherwise retu	STRING
HIGH	backends/vulkan/quantizer/vulkan_quantizer_utils.py	0	given a list of nodes (that represents an operator pattern), return true if any of the node is annotated, otherwise retu	STRING
HIGH	backends/xnnpack/quantizer/xnnpack_quantizer_utils.py	0	given a list of nodes (that represents an operator pattern), return true if any of the node is annotated, otherwise retu	STRING
HIGH	backends/qualcomm/quantizer/rules.py	0	given a list of nodes (that represents an operator pattern), return true if any of the node is annotated, otherwise retu	STRING
HIGH	backends/samsung/quantizer/annotator.py	0	given a list of nodes (that represents an operator pattern), return true if any of the node is annotated, otherwise retu	STRING
HIGH	backends/mediatek/quantizer/annotator.py	0	given a list of nodes (that represents an operator pattern), return true if any of the node is annotated, otherwise retu	STRING
HIGH	backends/nxp/tests/models.py	0	call `torch.max_pool2d` that is a noop, but it enforces the channelsfirst format in the `nodeformatinference`.	STRING
HIGH	…ir/converter/node_converter/test_mean_dim_converter.py	0	call `torch.max_pool2d` that is a noop, but it enforces the channelsfirst format in the `nodeformatinference`.	STRING
HIGH	…sts/ir/converter/node_converter/test_amin_converter.py	0	call `torch.max_pool2d` that is a noop, but it enforces the channelsfirst format in the `nodeformatinference`.	STRING
HIGH	…rter/node_converter/test_sum_dim_int_list_converter.py	0	call `torch.max_pool2d` that is a noop, but it enforces the channelsfirst format in the `nodeformatinference`.	STRING
HIGH	…ir/converter/node_converter/test_mean_dim_converter.py	0	when `keep_dim = false`, the `sum` operator changes the rank, so the format have to be explicitly handled. the tests in	STRING
HIGH	…sts/ir/converter/node_converter/test_amin_converter.py	0	when `keep_dim = false`, the `sum` operator changes the rank, so the format have to be explicitly handled. the tests in	STRING
HIGH	…rter/node_converter/test_sum_dim_int_list_converter.py	0	when `keep_dim = false`, the `sum` operator changes the rank, so the format have to be explicitly handled. the tests in	STRING
HIGH	backends/nxp/backend/edge_helper.py	0	get the static data from a given node. if it doesn't have any data, return `none`.	STRING
HIGH	backends/nxp/aten_passes/split_group_convolution.py	0	get the static data from a given node. if it doesn't have any data, return `none`.	STRING
HIGH	…ends/nxp/aten_passes/fuse_batch_norm_with_conv_pass.py	0	get the static data from a given node. if it doesn't have any data, return `none`.	STRING
HIGH	…ds/nxp/aten_passes/fuse_batch_norm_with_linear_pass.py	0	get the static data from a given node. if it doesn't have any data, return `none`.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/LogicalOrOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	…ds/nxp/backend/ir/lib/tflite/StablehloReduceOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/ZerosLikeOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	…/backend/ir/lib/tflite/StablehloReduceWindowOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	…kends/nxp/backend/ir/lib/tflite/ReduceWindowOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/EqualOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/PowOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/LessOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/AddOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	…s/nxp/backend/ir/lib/tflite/StablehloScatterOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	…ends/nxp/backend/ir/lib/tflite/HashtableSizeOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/QuantizeOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/RightShiftOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	…xp/backend/ir/lib/tflite/StablehloCustomCallOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/Pool2DOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/SoftmaxOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/ShapeOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	…ckends/nxp/backend/ir/lib/tflite/SparsityParameters.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	…ds/nxp/backend/ir/lib/tflite/ReverseSequenceOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/LeakyReluOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	…ends/nxp/backend/ir/lib/tflite/LSHProjectionOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	backends/nxp/backend/ir/lib/tflite/Metadata.py	0	this method is deprecated. please switch to getrootas.	STRING
HIGH	…nds/nxp/backend/ir/lib/tflite/StablehloSliceOptions.py	0	this method is deprecated. please switch to getrootas.	STRING
223 more matches not shown…

Unused Imports1283 hits · 1148 pts

Severity	File	Line	Context
LOW	kernels/prim_ops/test/prim_ops_test.py	10	CODE
LOW	kernels/quantized/test/test_quant_dequant_per_token.py	12	CODE
LOW	kernels/quantized/test/test_out_variants.py	10	CODE
LOW	kernels/quantized/test/test_out_variants.py	12	CODE
LOW	kernels/quantized/test/test_out_variants.py	13	CODE
LOW	kernels/quantized/test/test_out_variants.py	15	CODE
LOW	extension/pytree/__init__.py	26	CODE
LOW	extension/pytree/__init__.py	26	CODE
LOW	extension/pytree/__init__.py	26	CODE
LOW	extension/pytree/__init__.py	40	CODE
LOW	extension/pytree/__init__.py	40	CODE
LOW	extension/pytree/__init__.py	40	CODE
LOW	…tension/llm/custom_ops/test_update_cross_attn_cache.py	12	CODE
LOW	extension/llm/custom_ops/test_preprocess_custom_ops.py	14	CODE
LOW	extension/llm/custom_ops/custom_ops.py	32	CODE
LOW	extension/llm/custom_ops/test_sdpa_with_kv_cache.py	14	CODE
LOW	extension/llm/custom_ops/test_update_cache.py	14	CODE
LOW	extension/llm/custom_ops/model_sharding.py	10	CODE
LOW	extension/llm/custom_ops/test_quantized_sdpa.py	14	CODE
LOW	extension/llm/custom_ops/test_quantized_sdpa.py	29	CODE
LOW	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	3	CODE
LOW	extension/llm/runner/__init__.py	18	CODE
LOW	extension/llm/runner/__init__.py	18	CODE
LOW	extension/llm/runner/__init__.py	18	CODE
LOW	extension/llm/runner/__init__.py	18	CODE
LOW	extension/llm/runner/__init__.py	18	CODE
LOW	extension/llm/export/gguf.py	29	CODE
LOW	extension/llm/export/quantize.py	223	CODE
LOW	extension/llm/export/__init__.py	7	CODE
LOW	extension/llm/export/export_passes.py	114	CODE
LOW	extension/llm/modules/__init__.py	7	CODE
LOW	extension/llm/modules/__init__.py	7	CODE
LOW	extension/llm/modules/__init__.py	7	CODE
LOW	extension/llm/modules/__init__.py	7	CODE
LOW	extension/llm/modules/__init__.py	13	CODE
LOW	extension/llm/modules/__init__.py	13	CODE
LOW	extension/llm/modules/__init__.py	14	CODE
LOW	extension/llm/modules/turboquant/__init__.py	7	CODE
LOW	extension/training/__init__.py	9	CODE
LOW	extension/training/__init__.py	11	CODE
LOW	extension/training/__init__.py	11	CODE
LOW	extension/training/__init__.py	11	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
LOW	extension/pybindings/portable_lib.py	80	CODE
1223 more matches not shown…

Self-Referential Comments451 hits · 1055 pts

Severity	File	Line	Snippet	Context
MEDIUM	pytest-windows.ini	3	# This file is intended to be deleted once the enablement is complete.	COMMENT
MEDIUM	setup.py	720	# Create the version file.	COMMENT
MEDIUM	kernels/portable/test/op_grid_sampler_2d_test.py	119	# Create a checkerboard pattern	COMMENT
MEDIUM	kernels/portable/test/op_grid_sampler_2d_test.py	165	# Create a smooth gradient	COMMENT
MEDIUM	kernels/portable/test/op_grid_sampler_2d_test.py	168	# Create a fine grid for upsampling	COMMENT
MEDIUM	tools/cmake/common/__init__.py	33	# Create a file tree in the current working directory (cwd). The structure of the	COMMENT
MEDIUM	extension/pytree/__init__.py	11	# Create a logger	COMMENT
MEDIUM	…tension/llm/custom_ops/test_update_cross_attn_cache.py	64	# Define a function that uses the op inside torch.cond	COMMENT
MEDIUM	…tension/llm/custom_ops/test_update_cross_attn_cache.py	115	# Define a module that uses torch.cond with the op	COMMENT
MEDIUM	…tension/llm/custom_ops/test_update_cross_attn_cache.py	224	# Define a function that uses different ops in each branch	COMMENT
MEDIUM	…/llm/custom_ops/spinquant/special_hadamard_code_gen.py	2	# This file is adapted from	COMMENT
MEDIUM	extension/llm/runner/test/test_runner_pybindings.py	167	# Create an image	COMMENT
MEDIUM	extension/llm/runner/test/test_runner_pybindings.py	256	# Create a test image tensor (RGB, CHW)	COMMENT
MEDIUM	extension/llm/export/test/test_export_llm.py	59	# Create a temporary config file	COMMENT
MEDIUM	extension/llm/export/test/test_export_llm.py	111	# Create a temporary config file	COMMENT
MEDIUM⚡	extension/llm/export/test/test_builder.py	18	# Create a mock model	COMMENT
MEDIUM⚡	extension/llm/export/test/test_builder.py	28	# Create a manager with preset dynamic_shapes	COMMENT
MEDIUM	extension/llm/export/test/test_builder.py	46	# Create a manager with enable_dynamic_shape=True and use_kv_cache=False	COMMENT
MEDIUM	extension/llm/export/test/test_builder.py	70	# Create a manager with enable_dynamic_shape=True and use_kv_cache=True	COMMENT
MEDIUM	extension/llm/export/test/test_builder.py	103	# Create a manager with enable_dynamic_shape=False	COMMENT
MEDIUM	extension/pybindings/test/make_test.py	184	# Create the ExecuTorch program from the graph.	COMMENT
MEDIUM	.ci/scripts/setup-ios.sh	21	# Create a temporary keychain	COMMENT
MEDIUM	.ci/scripts/test_coreml_bc.sh	23	# Create a conda environment with Python 3.10 for compatibility with old ET versions	COMMENT
MEDIUM	.ci/scripts/wheel/envvar_linux.sh	7	# This file is sourced into the environment before building a pip wheel. It	COMMENT
MEDIUM	.ci/scripts/wheel/envvar_base.sh	7	# This file is sourced into the environment before building a pip wheel. It	COMMENT
MEDIUM	.ci/scripts/wheel/envvar_macos.sh	7	# This file is sourced into the environment before building a pip wheel. It	COMMENT
MEDIUM	…ripts/benchmark_tooling/analyze_benchmark_stability.py	402	# Create a string buffer to hold the report content	COMMENT
MEDIUM	…ripts/benchmark_tooling/analyze_benchmark_stability.py	665	# Create a string buffer to hold the report content	COMMENT
MEDIUM	…ripts/benchmark_tooling/analyze_benchmark_stability.py	1051	# Create a string buffer to hold the report content	COMMENT
MEDIUM	…ripts/benchmark_tooling/analyze_benchmark_stability.py	1248	# Create a string buffer to hold the report content	COMMENT
MEDIUM	test/end2end/exported_module.py	124	# Create the eager module.	COMMENT
MEDIUM	test/end2end/exported_module.py	224	# Create the ExportedModule.	COMMENT
MEDIUM	backends/apple/coreml/compiler/coreml_preprocess.py	760	# Create a temporary directory for all the mlpackages	COMMENT
MEDIUM	backends/apple/coreml/compiler/coreml_preprocess.py	768	# Create a NamedDataStore to hold the shared multifunction models	COMMENT
MEDIUM	backends/nxp/tests/ops_aliases.py	6	# This file defines ops aliases for shorter and more readable test description. List is sorted alphabetically.	COMMENT
MEDIUM	…ir/converter/node_converter/test_hardtanh_converter.py	69	# Create a RandomDatasetCreator that covers also negative numbers to properly test the operator.	COMMENT
MEDIUM	…/ir/converter/node_converter/test_sigmoid_converter.py	39	# Create a RandomDatasetCreator that covers also negative numbers to properly test the operator.	COMMENT
MEDIUM	…tests/ir/converter/node_converter/test_mm_converter.py	47	# Create a RandomDatasetCreator that covers also negative numbers to properly test the operator.	COMMENT
MEDIUM	…ts/ir/converter/node_converter/test_addmm_converter.py	55	# Create a RandomDatasetCreator that covers also negative numbers to properly test the operator.	COMMENT
MEDIUM	…/converter/node_converter/test_leaky_relu_converter.py	44	# Create a RandomDatasetCreator that covers also negative numbers to properly test the operator.	COMMENT
MEDIUM	…de_converters/ops_converters/permute_copy_converter.py	221	"""This method is currently far more complex than necessary, as Neutron C supports all permutations.	STRING
MEDIUM	…/node_converters/ops_converters/view_copy_converter.py	90	# Create the TFLite Reshape with the new shape	COMMENT
MEDIUM	…ends/nxp/backend/ir/converter/builder/model_builder.py	202	# Create a new one and register it for potential future use.	COMMENT
MEDIUM	…ends/nxp/backend/ir/converter/builder/model_builder.py	249	# Create the output tensor of the 'Pad' operator.	COMMENT
MEDIUM	…ends/nxp/backend/ir/converter/builder/model_builder.py	255	# Create the second input of the 'Pad' operator.	COMMENT
MEDIUM	…ends/nxp/backend/ir/converter/builder/model_builder.py	442	# Create a Transpose operator and replace the graph input	COMMENT
MEDIUM	…ends/nxp/backend/ir/converter/builder/model_builder.py	574	# Create the final edge-to-tflite mapping after model optimization	COMMENT
MEDIUM	…tions/permute_fully_connected_weights_after_reshape.py	105	# Create a new tensor for the data, in case it is used by some other operator as well.	COMMENT
MEDIUM	backends/nxp/backend/ir/lib/tflite/LogicalOrOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
MEDIUM	…ds/nxp/backend/ir/lib/tflite/StablehloReduceOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
MEDIUM	backends/nxp/backend/ir/lib/tflite/ZerosLikeOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
MEDIUM	…/backend/ir/lib/tflite/StablehloReduceWindowOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
MEDIUM	…kends/nxp/backend/ir/lib/tflite/ReduceWindowOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
MEDIUM	backends/nxp/backend/ir/lib/tflite/EqualOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
MEDIUM	backends/nxp/backend/ir/lib/tflite/PowOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
MEDIUM	backends/nxp/backend/ir/lib/tflite/LessOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
MEDIUM	backends/nxp/backend/ir/lib/tflite/AddOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
MEDIUM	…s/nxp/backend/ir/lib/tflite/StablehloScatterOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
MEDIUM	…ends/nxp/backend/ir/lib/tflite/HashtableSizeOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
MEDIUM	backends/nxp/backend/ir/lib/tflite/QuantizeOptions.py	23	"""This method is deprecated. Please switch to GetRootAs."""	STRING
391 more matches not shown…

Over-Commented Block1043 hits · 1020 pts

Severity	File	Line	Snippet	Context
LOW	install_requirements.sh	1	#!/bin/bash	COMMENT
LOW	install_executorch.py	81		COMMENT
LOW	CMakeLists.txt	1	# Copyright (c) Meta Platforms, Inc. and affiliates.	COMMENT
LOW	CMakeLists.txt	21	(rm -rf cmake-out \	COMMENT
LOW	pyproject.toml	101	# include. See also setuptools/discovery.py for custom finders.	COMMENT
LOW	install_requirements.py	21	#	COMMENT
LOW	setup.py	1	# Copyright (c) Meta Platforms, Inc. and affiliates.	COMMENT
LOW	setup.py	21	# this list of conditions and the following disclaimer in the documentation	COMMENT
LOW	install_executorch.sh	1	#!/bin/bash	COMMENT
LOW	kernels/prim_ops/et_copy_index.cpp	21	namespace torch {	COMMENT
LOW	kernels/prim_ops/et_copy_index.cpp	41	// submodule_0 = self.submodule_0	COMMENT
LOW	kernels/prim_ops/test/prim_ops_test.cpp	1	/*	COMMENT
LOW	kernels/prim_ops/test/CMakeLists.txt	1	# Copyright (c) Meta Platforms, Inc. and affiliates.	COMMENT
LOW	kernels/test/CMakeLists.txt	1	# Copyright (c) Meta Platforms, Inc. and affiliates.	COMMENT
LOW	kernels/test/op_index_select_test.cpp	261	// clang-format on	COMMENT
LOW	kernels/test/op_to_copy_test.cpp	1	/*	COMMENT
LOW	kernels/test/op__to_dim_order_copy_test.cpp	1	/*	COMMENT
LOW	kernels/test/op_mean_test.cpp	1	/*	COMMENT
LOW	kernels/test/op_select_copy_test.cpp	41		COMMENT
LOW	kernels/test/op_select_copy_test.cpp	301	#define TEST_ENTRY(ctype, dtype) test_dtype<ctype, ScalarType::dtype>();	COMMENT
LOW	kernels/test/op_select_scatter_test.cpp	501	/// A generic smoke test that works for any dtype that supports ones() and	COMMENT
LOW	kernels/test/UnaryUfuncRealHBBF16ToFloatHBF16Test.h	21	// Generic test harness for ops that use unary_ufunc_realhb_to_floath	COMMENT
LOW	kernels/test/op_linear_test.cpp	1	/*	COMMENT
LOW	kernels/test/op_stack_test.cpp	121		COMMENT
LOW	kernels/test/op_mm_test.cpp	1	/*	COMMENT
LOW	kernels/test/op_amin_test.cpp	1	/*	COMMENT
LOW	kernels/test/op_min_test.cpp	1	/*	COMMENT
LOW	kernels/test/op_native_layer_norm_test.cpp	1	/*	COMMENT
LOW	kernels/test/op_sum_test.cpp	1	/*	COMMENT
LOW	kernels/test/supported_features_skip.h	1	/*	COMMENT
LOW	kernels/test/supported_features_skip.h	21	// fbcode: expands to `if (cond) return;` so the test reports PASS, not SKIP.	COMMENT
LOW	kernels/test/update_supported_feature_summary.sh	1	#!/bin/bash	COMMENT
LOW	kernels/test/op_max_test.cpp	1	/*	COMMENT
LOW	kernels/test/op_clamp_test.cpp	1	/*	COMMENT
LOW	kernels/test/op_amax_test.cpp	1	/*	COMMENT
LOW	kernels/test/op_arange_test.cpp	1	/*	COMMENT
LOW	kernels/test/op_var_test.cpp	1	/*	COMMENT
LOW	kernels/quantized/CMakeLists.txt	1	# Copyright (c) Meta Platforms, Inc. and affiliates.	COMMENT
LOW	kernels/portable/custom_ops.yaml	1	# Copyright (c) Meta Platforms, Inc. and affiliates.	COMMENT
LOW	kernels/portable/functions.yaml	1	# Copyright (c) Meta Platforms, Inc. and affiliates.	COMMENT
LOW	kernels/portable/cpu/math_constants.h	21	#define M_E 2.7182818284590452354 /* e */	COMMENT
LOW	kernels/portable/cpu/math_constants.h	41	#define M_PI 3.14159265358979323846 /* pi */	COMMENT
LOW	kernels/portable/cpu/math_constants.h	61	#define M_2_SQRTPI 1.12837916709551257390 /* 2/sqrt(pi) */	COMMENT
LOW	kernels/portable/cpu/util/transpose_util.h	101	auto a_sizes = a.sizes();	COMMENT
LOW	kernels/portable/cpu/util/repeat_util.cpp	241	// 'in_offset' indicates the offset (in bytes) from the origin to an self	COMMENT
LOW	kernels/portable/cpu/util/grid_sampler_2d_util.h	21	enum class GridSamplerInterpolation { Bilinear, Nearest, Bicubic };	COMMENT
LOW	kernels/optimized/blas/BlasKernel.cpp	1	/*	COMMENT
LOW	kernels/optimized/blas/BlasKernel.cpp	101		COMMENT
LOW	kernels/optimized/cpu/op_grid_sampler_2d.cpp	1	/*	COMMENT
LOW	kernels/optimized/cpu/op_grid_sampler_2d.cpp	21	// instructions on older ones.	COMMENT
LOW	kernels/optimized/cpu/op_grid_sampler_2d_fp16_hw.cpp	1	/*	COMMENT
LOW	kernels/optimized/utils/llvmMathExtras.h	1	/*	COMMENT
LOW	kernels/optimized/utils/llvmMathExtras.h	21	#pragma once	COMMENT
LOW	kernels/optimized/utils/llvmMathExtras.h	41	#if defined(__GNUC__) && defined(__GNUC_MINOR__) && defined(__GNUC_PATCHLEVEL__)	COMMENT
LOW	kernels/optimized/utils/llvmMathExtras.h	701	return 1ull << (63 - countLeadingZeros(A, ZB_Undefined));	COMMENT
LOW	kernels/optimized/utils/llvmMathExtras.h	721	/// \code	COMMENT
LOW	tools/cmake/common/__init__.py	41	# "README.md": "this is a read me file",	COMMENT
LOW	extension/aten_util/make_aten_functor_from_et_functor.h	1	/*	COMMENT
LOW	extension/aten_util/make_aten_functor_from_et_functor.h	21	#include <ATen/native/Resize.h>	COMMENT
LOW	…sion/apple/ExecuTorch/Exported/ExecuTorch+Tensor.swift	121	/// The shape of the tensor.	COMMENT
983 more matches not shown…

Deep Nesting807 hits · 728 pts

Severity	File	Line	Context
LOW	install_executorch.py	111	CODE
LOW	install_utils.py	243	CODE
LOW	kernels/test/summarize_supported_features.py	20	CODE
LOW	kernels/test/test_case_gen.py	243	CODE
LOW	…rnels/portable/test/test_grid_sampler_2d_executorch.py	361	CODE
LOW	tools/cmake/cmake_cache.py	43	CODE
LOW	extension/llm/custom_ops/model_sharding.py	35	CODE
LOW	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	650	CODE
LOW	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	725	CODE
LOW	extension/llm/export/gguf.py	514	CODE
LOW	extension/llm/export/quantize.py	28	CODE
LOW	extension/llm/export/builder.py	288	CODE
LOW	extension/llm/export/builder.py	377	CODE
LOW	extension/llm/export/builder.py	470	CODE
LOW	extension/llm/export/builder.py	301	CODE
LOW	extension/llm/export/export_passes.py	44	CODE
LOW	extension/llm/export/quantizer_lib.py	260	CODE
LOW	extension/training/examples/CIFAR/train_utils.py	51	CODE
LOW	extension/training/examples/CIFAR/train_utils.py	188	CODE
LOW	.ci/scripts/gather_benchmark_configs.py	163	CODE
LOW	.ci/scripts/gather_benchmark_configs.py	316	CODE
LOW	.ci/scripts/test_example_imports.py	212	CODE
LOW	.ci/scripts/test_example_imports.py	410	CODE
LOW	.ci/scripts/test_example_imports.py	31	CODE
LOW	.ci/scripts/test_huggingface_optimum_model.py	133	CODE
LOW	…ripts/benchmark_tooling/analyze_benchmark_stability.py	68	CODE
LOW	…ripts/benchmark_tooling/analyze_benchmark_stability.py	655	CODE
LOW	…ripts/benchmark_tooling/analyze_benchmark_stability.py	1242	CODE
LOW	…ripts/benchmark_tooling/get_benchmark_analysis_data.py	282	CODE
LOW	util/collect_env.py	184	CODE
LOW	util/collect_env.py	344	CODE
LOW	util/collect_env.py	378	CODE
LOW	util/activation_memory_profiler.py	51	CODE
LOW	backends/apple/mps/utils/mps_utils.py	44	CODE
LOW	backends/apple/metal/passes/decompose_linear_pass.py	24	CODE
LOW	backends/apple/coreml/recipes/coreml_recipe_provider.py	42	CODE
LOW	backends/apple/coreml/partition/coreml_partitioner.py	181	CODE
LOW	backends/apple/coreml/partition/coreml_partitioner.py	343	CODE
LOW	…ckends/apple/coreml/compiler/enumerated_shape_utils.py	102	CODE
LOW	backends/nxp/neutron_partitioner.py	370	CODE
LOW	backends/nxp/tests/models.py	548	CODE
LOW	backends/nxp/tests/dataset_creator.py	312	CODE
LOW	backends/nxp/tests/nsys_testing.py	322	CODE
LOW	backends/nxp/tests/nsys_testing.py	788	CODE
LOW	backends/nxp/backend/node_format_inference.py	131	CODE
LOW	backends/nxp/backend/node_format_inference.py	303	CODE
LOW	backends/nxp/backend/neutron_map.py	227	CODE
LOW	backends/nxp/backend/neutron_map.py	337	CODE
LOW	backends/nxp/backend/neutron_map.py	369	CODE
LOW	backends/nxp/backend/edge_program_converter.py	140	CODE
LOW	backends/nxp/backend/edge_program_converter.py	166	CODE
LOW	backends/nxp/backend/edge_program_converter.py	228	CODE
LOW	backends/nxp/backend/edge_helper.py	311	CODE
LOW	backends/nxp/backend/ir/tflite_generator/meta/types.py	138	CODE
LOW	backends/nxp/backend/ir/tflite_generator/meta/types.py	162	CODE
LOW	…ode_converters/ops_converters/convolution_converter.py	417	CODE
LOW	…de_converters/ops_converters/permute_copy_converter.py	218	CODE
LOW	…nd/ir/converter/node_converters/shared/reduce_utils.py	67	CODE
LOW	…verter/node_converters/shared/reshape_transposition.py	139	CODE
LOW	…ends/nxp/backend/ir/converter/conversion/translator.py	522	CODE
747 more matches not shown…

Cross-Language Confusion94 hits · 466 pts

Severity	File	Line	Snippet	Context
HIGH	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	757	"cd measurements && make run_%s" % type_name, shell=True, stdout=devnull	CODE
HIGH	extension/llm/export/quantizer_lib.py	112	"rm -rf cmake-out && mkdir cmake-out && (cd cmake-out && cmake -DEXECUTORCH_BUILD_KERNELS_QUANTIZED_	CODE
HIGH	backends/qualcomm/tests/utils.py	249	adb.push(	CODE
HIGH	backends/qualcomm/tests/utils.py	622	adb.push(	CODE
HIGH	backends/qualcomm/tests/rework/conftest.py	355	adb.push()	CODE
HIGH	backends/qualcomm/tests/rework/conftest.py	397	adb.push(inputs=[inputs], init_env=False)	CODE
HIGH	backends/qualcomm/debugger/utils.py	297	self.adb.push(	STRING
HIGH	backends/samsung/test/utils/runtime_executor.py	129	test_manager.push()	CODE
HIGH	backends/vulkan/test/op_tests/utils/gen_benchmark_vk.py	230	if (high == 1.0 && low == 0.0)	CODE
HIGH	…nds/vulkan/test/op_tests/utils/gen_correctness_base.py	317	if (high == 1.0 && low == 0.0)	CODE
HIGH	…kends/vulkan/test/op_tests/utils/gen_correctness_vk.py	151	if (t1.scalar_type() == at::kLong \|\| t2.scalar_type() == at::kLong) {	CODE
HIGH	…kends/vulkan/test/op_tests/utils/gen_correctness_vk.py	155	if (!is_close && t1.numel() < 500) {	CODE
HIGH	…s/xnnpack/_passes/channels_last_tagged_reshape_pass.py	148	# then we want to see this is nchw hence return false	COMMENT
HIGH	backends/arm/_passes/insert_rescales_pass.py	218	# zero point is non-null, the result will be in the range [-255;	COMMENT
HIGH	backends/arm/scripts/aot_arm_compiler.py	804	intermediates_path, f"input_{method_index}.pt" # type: ignore[possibly-undefined]	CODE
HIGH	backends/mlx/custom_kernel_ops/gguf/q5k/linear.py	166	if (tiisg == 0 && r < N) {{	CODE
HIGH	backends/mlx/custom_kernel_ops/gguf/q5k/linear.py	314	if (j < nr1 && i < nr0) {{	CODE
HIGH	backends/mlx/custom_kernel_ops/gguf/q6k/linear.py	142	if (tiisg == 0 && r < N) {{	CODE
HIGH	backends/mlx/custom_kernel_ops/gguf/q6k/linear.py	289	if (j < nr1 && i < nr0) {{	CODE
HIGH⚡	backends/mlx/custom_kernel_ops/gguf/q4k/embedding.py	48	if (row < 0 \|\| row >= V) {	CODE
HIGH	backends/mlx/custom_kernel_ops/gguf/q4k/linear.py	133	if (tiisg == 0 && r < N) {{	CODE
HIGH	backends/mlx/custom_kernel_ops/gguf/q4k/linear.py	279	if (j < nr1 && i < nr0) {{	CODE
HIGH	codegen/tools/test/test_gen_selected_op_variants.py	77	&& (scalar_type == executorch::aten::ScalarType::Float \|\| scalar_type == executorch::aten::ScalarType::Int))	STRING
HIGH	codegen/tools/test/test_gen_selected_op_variants.py	149	&& (scalar_type == executorch::aten::ScalarType::Float \|\| scalar_type == executorch::aten::ScalarType::Int))	STRING
HIGH	devtools/inspector/tests/inspector_test.py	2212	1. The output tensor has a valid (non-null) data pointer	STRING
HIGH	examples/qualcomm/qaihub_scripts/utils/export.py	194	adb.push(files=files)	CODE
HIGH	examples/qualcomm/qaihub_scripts/utils/export.py	323	adb.push(inputs=inputs)	CODE
HIGH	…ub_scripts/stable_diffusion/qaihub_stable_diffusion.py	323	adb.push(inputs=input_unet, files=files)	CODE
HIGH	…alcomm/qaihub_scripts/llama/llama3/qaihub_llama3_8b.py	164	adb.push(files=custom_files)	CODE
HIGH	…alcomm/qaihub_scripts/llama/llama2/qaihub_llama2_7b.py	164	adb.push(files=custom_files)	CODE
HIGH	…ualcomm/util_scripts/qnn_intermediate_debugger_demo.py	85	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/util_scripts/cli.py	372	adb.push(inputs=user_inputs, backends=[backend_type])	CODE
HIGH	examples/qualcomm/util_scripts/gen_etrecord.py	75	adb.push(inputs=[sample_input])	CODE
HIGH	examples/qualcomm/scripts/wav2letter.py	151	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/scripts/deeplab_v3.py	105	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/scripts/mobilebert_fine_tune.py	292	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/scripts/edsr.py	140	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/scripts/mobilenet_v2.py	72	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/scripts/inception_v4.py	71	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/scripts/mobilenet_v3.py	79	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/scripts/inception_v3.py	73	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/scripts/torchvision_vit.py	121	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/custom_op/custom_ops_1.py	214	adb.push(inputs=sample_input, files=op_package_paths)	CODE
HIGH	examples/qualcomm/custom_op/custom_ops_2.py	224	adb.push(inputs=sample_input, files=op_package_paths)	CODE
HIGH	examples/qualcomm/oss_scripts/deit.py	94	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/ssd300_vgg16.py	157	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/pvt.py	77	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/dino_v2.py	79	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/convnext_small.py	80	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/vit_b_16.py	81	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/regnet.py	77	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/maxvit_t.py	188	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/roberta.py	125	adb.push(inputs=[sample_input])	CODE
HIGH	examples/qualcomm/oss_scripts/roberta.py	137	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/squeezenet.py	68	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/distilbert.py	109	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/fastvit.py	138	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/cvt.py	141	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/mobilevit_v1.py	105	adb.push(inputs=inputs)	CODE
HIGH	examples/qualcomm/oss_scripts/albert.py	109	adb.push(inputs=inputs)	CODE
34 more matches not shown…

Structural Annotation Overuse261 hits · 449 pts

Severity	File	Line	Snippet	Context
LOW⚡	install_executorch.py	225	# Step 1: Install core dependencies first	COMMENT
LOW⚡	install_executorch.py	228	# Step 2: Install core package	COMMENT
LOW	install_executorch.py	249	# Step 3: Extra (optional) packages that is only useful for running examples.	COMMENT
LOW	.wiki/troubleshooting/runtime-errors.md	181	### Step 1: Profile	COMMENT
LOW⚡	.wiki/troubleshooting/runtime-errors.md	192	### Step 2: Analyze	COMMENT
LOW⚡	.wiki/troubleshooting/runtime-errors.md	200	### Step 3: Identify Bottlenecks	COMMENT
LOW⚡	.wiki/troubleshooting/runtime-errors.md	207	### Step 4: Advanced Profiling	COMMENT
LOW	.wiki/quantization/debugging.md	102	### Step 1: Export Both Versions	COMMENT
LOW	.wiki/quantization/debugging.md	121	### Step 2: Run Both and Compare	COMMENT
LOW	.wiki/quantization/debugging.md	139	### Step 3: For CoreML, Extract and Compare	COMMENT
LOW	extension/module/test/resources/gen_bundled_program.py	14	# Step 1: ExecuTorch Program Export	COMMENT
LOW	extension/module/test/resources/gen_bundled_program.py	54	# Step 2: Construct MethodTestSuite for Each Method	COMMENT
LOW⚡	extension/module/test/resources/gen_bundled_program.py	86	# Step 3: Generate BundledProgram	COMMENT
LOW⚡	extension/module/test/resources/gen_bundled_program.py	89	# Step 4: Serialize BundledProgram to flatbuffer.	COMMENT
LOW⚡	extension/gguf_util/load_gguf.py	96	# Step 1: Build GGUFModelArgs	COMMENT
LOW⚡	extension/gguf_util/load_gguf.py	100	# Step 2: Build GGUFWeights	COMMENT
LOW⚡	extension/gguf_util/convert_main.py	32	# Step 1: Load the GGUF file	COMMENT
LOW⚡	extension/gguf_util/convert_main.py	35	# Step 2: Convert the GGUF model to PTE	COMMENT
LOW	extension/gguf_util/convert_main.py	47	# Step 3: Save the PTE program so that	COMMENT
LOW⚡	extension/gguf_util/converters/llama_converter.py	110	# Step 1: Create the PyTorch model	COMMENT
LOW⚡	extension/gguf_util/converters/llama_converter.py	116	# Step 2: Load the weights into the PyTorch model	COMMENT
LOW⚡	extension/gguf_util/converters/llama_converter.py	120	# Step 3: Export to ExecuTorch	COMMENT
LOW	extension/image/image_processor.cpp	323	// Step 1: orientation (orient -> ROI -> resize). Produce an oriented copy of	COMMENT
LOW	extension/image/image_processor.cpp	347	// Step 2: ROI crop (pointer arithmetic).	COMMENT
LOW	extension/image/image_processor.cpp	368	// Step 3: Swizzle BGRA/RGBA → RGB (alpha discarded).	COMMENT
LOW	extension/image/image_processor.cpp	382	// Step 4: Resize.	COMMENT
LOW	extension/image/image_processor.cpp	403	// Step 5: Normalize + layout into the caller's CHW output (padded).	COMMENT
LOW	extension/image/image_processor_apple.cpp	694	// Step 2: orientation. Rotate the BGRA buffer (vImage) so ROI/resize run in	COMMENT
LOW⚡	extension/image/image_processor_apple.cpp	720	// Step 3: ROI crop (pointer arithmetic on BGRA data).	COMMENT
LOW⚡	extension/image/image_processor_apple.cpp	723	// Step 4: resize. Letterbox padding is applied during normalization.	COMMENT
LOW	extension/image/image_processor_apple.cpp	752	// Step 5: normalize BGRA → CHW float buffer.	COMMENT
LOW	.ci/docker/common/install_conda.sh	59	# WARNING: This is a HACK from PyTorch core to be able to build PyTorch on 22.04.	COMMENT
LOW	.ci/scripts/gather_test_models.py	8	# WARNING: The CI runner logic should directly be in the corresponding yml files	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	61	# Step 1: Checkout old version of ExecuTorch	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	66	# Step 2: Install the old version	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	75	# Step 3: Export model	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	124	# Step 4: Checkout current version (main)	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	130	# Step 5: Install current version	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	135	# Step 6: Run the old pte file	COMMENT
LOW	backends/qualcomm/_passes/decompose_diagonal.py	95	# Step 1: Permute dim1, dim2 to last positions (if needed)	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_diagonal.py	107	# Step 2: Reshape [..., M, N] -> [..., M*N]	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_diagonal.py	117	# Step 3: Compute flat diagonal indices	COMMENT
LOW	backends/qualcomm/_passes/decompose_diagonal.py	142	# Step 4: index_select on last dim	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_cdist.py	19	# Step 1: Compute differences	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_cdist.py	22	# Step 2: Square differences	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_cdist.py	25	# Step 3: Sum of squares	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_cdist.py	28	# Step 4: Square root	COMMENT
LOW	backends/qualcomm/_passes/decompose_var.py	106	# Step 1: mean_x = mean(x, dim, keepdim=True)	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_var.py	118	# Step 2: diff = x - mean_x	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_var.py	126	# Step 3: sq = diff * diff (more efficient than pow(diff, 2))	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_var.py	134	# Step 4: var = mean(sq, dim, keepdim)	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_var.py	140	# Step 5: Apply correction factor if needed	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_div_mode.py	55	# Step 1: div_result = div(x, y)	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_div_mode.py	61	# Step 2: Apply rounding mode if needed	COMMENT
LOW	backends/qualcomm/_passes/convert_linear_to_conv2d.py	134	# Step 1: reshape input	COMMENT
LOW	backends/qualcomm/_passes/convert_linear_to_conv2d.py	168	# Step 2: reshape weight	COMMENT
LOW	backends/qualcomm/_passes/convert_linear_to_conv2d.py	202	# Step 3: restore shape	COMMENT
LOW	backends/cuda/runtime/cuda_backend.cpp	1028	// Step 1: Enumerate constants and partition into cached/uncached	COMMENT
LOW	backends/cuda/runtime/cuda_backend.cpp	1071	// Step 2: Load uncached constants from blob (if any).	COMMENT
LOW	backends/cuda/runtime/cuda_backend.cpp	1157	// Step 3: Point the container to cached tensors via user_managed pairs	COMMENT
201 more matches not shown…

Excessive Try-Catch Wrapping323 hits · 370 pts

Severity	File	Line	Snippet	Context
LOW	install_utils.py	55	except Exception:	CODE
LOW	install_utils.py	218	except Exception as err:	CODE
LOW	install_utils.py	283	except Exception as e:	CODE
LOW	kernels/portable/__init__.py	22	except Exception: # noqa: E722	CODE
LOW	…rnels/portable/test/test_grid_sampler_2d_executorch.py	112	except Exception as e:	CODE
LOW	tools/cmake/buck_util.py	57	except Exception as e:	CODE
LOW	extension/llm/custom_ops/test_update_cache.py	226	except Exception:	CODE
LOW	extension/llm/custom_ops/test_update_cache.py	246	except Exception:	CODE
LOW	extension/llm/custom_ops/test_update_cache.py	346	except Exception:	CODE
LOW	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	624	except Exception:	CODE
LOW	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	640	except Exception as e:	CODE
LOW	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	655	except Exception:	CODE
LOW	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	673	except Exception:	CODE
LOW	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	690	except Exception:	CODE
LOW	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	793	except Exception:	CODE
LOW	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	828	except Exception as e:	CODE
MEDIUM	…nsion/llm/custom_ops/spinquant/third-party/FFHT/gen.py	621	def greedy_merged(type_name, log_n, composite_step):	CODE
LOW	extension/pybindings/portable_lib.py	59	except Exception as e:	CODE
LOW	extension/pybindings/portable_lib.py	69	except Exception as e:	CODE
LOW	extension/pybindings/test/test_pybindings.py	39	except Exception:	CODE
LOW	extension/pybindings/test/test_pybindings.py	49	except Exception:	CODE
LOW	extension/pybindings/test/test_pybindings.py	121	except Exception:	CODE
LOW	extension/pybindings/test/test_pybindings.py	380	except Exception:	CODE
LOW	extension/export_util/utils.py	144	except Exception as e:	CODE
MEDIUM	.ci/scripts/cuda_benchmark.py	100	print(f"Error parsing PyTorchObserver log: {e}", file=sys.stderr)	CODE
LOW	.ci/scripts/cuda_benchmark.py	134	except Exception as e:	CODE
LOW	.ci/scripts/cuda_benchmark.py	202	except Exception as e:	CODE
MEDIUM	.ci/scripts/cuda_benchmark.py	203	print(f"Error: Failed to set GPU clocks: {e}", file=sys.stderr)	CODE
LOW	.ci/scripts/cuda_benchmark.py	244	except Exception as e:	CODE
MEDIUM	.ci/scripts/cuda_benchmark.py	245	print(f"Error: Failed to reset GPU clocks: {e}", file=sys.stderr)	CODE
MEDIUM⚡	.ci/scripts/cuda_benchmark.py	330	print(f"Error: Command timed out on run {run_num}", file=sys.stderr)	CODE
LOW⚡	.ci/scripts/cuda_benchmark.py	332	except Exception as e:	CODE
MEDIUM⚡	.ci/scripts/cuda_benchmark.py	333	print(f"Error on run {run_num}: {e}", file=sys.stderr)	CODE
LOW	.ci/scripts/test-cuda-build.sh	78	except Exception as e:	CODE
LOW	.ci/scripts/test_model_e2e.sh	515	except Exception as e:	CODE
LOW	.ci/scripts/test_backend_imports.py	122	except Exception as error:	CODE
LOW	test/end2end/test_temp_allocator_fix.py	209	except Exception as e:	CODE
MEDIUM	backends/apple/metal/tests/test_modules.py	1140	print(f"Error reading binary file {filepath}: {e}")	CODE
LOW	backends/apple/coreml/partition/coreml_partitioner.py	376	except Exception as e:	STRING
LOW	backends/nxp/tests/utils.py	38	except Exception as e:	CODE
LOW	backends/nxp/backend/edge_helper.py	397	except Exception:	CODE
LOW	backends/nxp/backend/ir/converter/tensor_utils.py	25	except Exception as e:	CODE
LOW⚡	…ends/nxp/backend/ir/converter/builder/model_builder.py	625	except Exception:	CODE
LOW⚡	…ends/nxp/backend/ir/converter/builder/model_builder.py	634	except Exception:	CODE
LOW	…ckend/ir/neutron_ir_post_processing/pattern_matcher.py	118	except Exception:	CODE
LOW	…ckend/ir/neutron_ir_post_processing/pattern_matcher.py	328	except Exception:	CODE
LOW	…nds/nxp/aten_passes/remove_nodes_with_known_outputs.py	182	except Exception:	CODE
LOW	backends/test/harness/tester.py	365	except Exception:	CODE
LOW	backends/test/suite/runner.py	129	except Exception as e:	CODE
LOW	backends/test/suite/runner.py	134	except Exception as e:	CODE
LOW⚡	backends/test/suite/runner.py	145	except Exception as e:	CODE
LOW⚡	backends/test/suite/runner.py	155	except Exception as e:	CODE
LOW⚡	backends/test/suite/runner.py	163	except Exception as e:	CODE
LOW	backends/test/suite/runner.py	202	except Exception as e:	CODE
LOW	backends/test/suite/runner.py	226	except Exception as e:	CODE
LOW	backends/test/suite/runner.py	236	except Exception:	CODE
LOW	backends/test/suite/flow.py	81	except Exception as e:	CODE
LOW	backends/test/suite/flow.py	199	except Exception as e:	CODE
LOW	backends/test/suite/generate_markdown_summary.py	243	except Exception as e:	CODE
MEDIUM	backends/test/suite/generate_markdown_summary.py	244	print(f"Error: {e}", file=sys.stderr)	CODE
263 more matches not shown…

Redundant / Tautological Comments219 hits · 329 pts

Severity	File	Line	Snippet	Context
LOW	install_executorch.py	136	# Check if the directories exist for each required submodule	COMMENT
LOW	setup.py	882	# Check if CUDA is available, and if so, enable building the CUDA	COMMENT
LOW	setup.py	920	# Check if QNN SDK is available (via QNN_SDK_ROOT env var), and if so,	COMMENT
LOW	setup.py	1043	# Set PYTHONPATH to the location of the pip package.	COMMENT
LOW	install_utils.py	118	# Check if the detected version is supported	COMMENT
LOW	…tension/llm/custom_ops/test_update_cross_attn_cache.py	17	# Check if CUDA device has compatible compute capability for Triton kernels	COMMENT
LOW	extension/llm/export/export_passes.py	51	# Check if the next node is also a transpose node	COMMENT
LOW	extension/llm/export/export_passes.py	69	# Check if the two transpose nodes undo each other	COMMENT
LOW	extension/llm/export/export_passes.py	79	# Check if the next node is also a transpose node	COMMENT
LOW	extension/llm/export/export_passes.py	91	# Check if the two permutes undo each other	COMMENT
LOW	extension/training/examples/CIFAR/train_utils.py	132	model.eval() # Set model to evaluation mode	CODE
LOW	.ci/docker/common/install_cuda_windows_cross_compile.sh	95	# Check if already downloaded and extracted	COMMENT
LOW	.ci/scripts/download_hf_hub.sh	57	# Check if script is called directly	COMMENT
LOW	.ci/scripts/compare_dirs.sh	10	# Check if dir1's files are also found in dir2 with the same	COMMENT
LOW	.ci/scripts/compare_dirs.sh	36	# Check if the corresponding file exists in the second directory	COMMENT
LOW	backends/apple/mps/operators/node_visitor.py	241	# Check if this node is a lifted parameter	COMMENT
LOW	backends/apple/metal/passes/decompose_linear_pass.py	30	# Check if this is a linear operation	COMMENT
LOW	backends/apple/metal/tests/test_metal_partitioner.py	59	# Check if we have any tagged nodes	COMMENT
LOW	backends/apple/metal/tests/test_metal_partitioner.py	63	# Check if all tagged nodes have the same tag	COMMENT
LOW	backends/apple/metal/tests/run_metal_test.sh	13	# ./run_metal_test.sh --check-build # Check if runtime is already built	COMMENT
LOW	backends/apple/metal/tests/run_metal_test.sh	55	# Check if we're on macOS	COMMENT
LOW	backends/apple/metal/tests/test_modules.py	46	# Check if MPS is available for export tests	COMMENT
LOW	backends/apple/metal/tests/test_modules.py	52	# Check if running in CI (GitHub Actions)	COMMENT
LOW	backends/apple/metal/tests/test_modules.py	68	# Check if executor_runner is built	COMMENT
LOW	backends/apple/metal/tests/test_modules.py	1222	# Check if this model/dtype combination should be skipped	COMMENT
LOW	backends/apple/metal/tests/test_modules.py	1256	# Check if this model/dtype combination should be skipped	COMMENT
LOW	backends/apple/coreml/partition/coreml_partitioner.py	291	# Check if we're being called from the deprecated to_backend workflow	STRING
LOW	backends/nxp/nxp_backend.py	249	# Check if provided intermediates_dir is a correct path (None is decoded to str)	COMMENT
LOW	backends/nxp/neutron_partitioner.py	473	# Check if the nodes support the partitioning result. Mark the problematic nodes with `NXP_DO_NOT_DELEGATE`.	COMMENT
LOW	backends/nxp/tests/graph_verifier.py	229	# Check if the node was tagged for delegation during partitioning	COMMENT
LOW⚡	backends/nxp/tests/generic_tests/test_aot_example.py	151	# Check if delegated model was created and saved.	COMMENT
LOW⚡	backends/nxp/tests/generic_tests/test_aot_example.py	157	# Check if nonempty Neutron to Edge map was created.	COMMENT
LOW⚡	backends/nxp/tests/generic_tests/test_aot_example.py	160	# Check if ETRecord was created and saved.	COMMENT
LOW	backends/nxp/tests/generic_tests/test_quantizer.py	190	# Check if QDQ pattern:	COMMENT
LOW	backends/nxp/tests/generic_tests/test_quantizer.py	201	# Check if input and output quantization is same	COMMENT
LOW	backends/nxp/tests/generic_tests/test_quantizer.py	224	# Check if QDQ pattern:	COMMENT
LOW	backends/nxp/backend/neutron_map.py	147	# Check if there is two graphs in the input dump	COMMENT
LOW	backends/nxp/backend/neutron_map.py	318	# Check if the neutron input is also the special op (Pad or Transpose)	COMMENT
LOW	backends/nxp/backend/neutron_map.py	389	# Check if there can be potential issue in mapping.	COMMENT
LOW	…ckends/nxp/backend/ir/tflite_generator/tflite_model.py	599	# Check if any of the inputs is quantized	COMMENT
LOW	…ends/nxp/backend/ir/converter/builder/model_builder.py	194	# Check if such tensor already exists	COMMENT
LOW	…ckend/ir/neutron_ir_post_processing/pattern_matcher.py	648	# Check if it is connected via the inputs.	COMMENT
LOW	…ckend/ir/neutron_ir_post_processing/pattern_matcher.py	740	# Check if it is connected via the inputs.	COMMENT
LOW	…_processing/optimizations/prune_transpose_operators.py	146	# Check if the `Transpose` is doing nothing.	COMMENT
LOW	…es/remove_additional_quantize_dequantize_nodes_pass.py	93	# Check if all quantization params are equal to ensure that QDQ cluster can be removed.	COMMENT
LOW	…kends/nxp/aten_passes/move_activation_before_concat.py	55	# Check if all cat inputs nodes are conv 2D or linear 2D type and their only user is cat.	COMMENT
LOW	…kends/nxp/aten_passes/move_activation_before_concat.py	65	# Check if following activation is supported on Neutron as fused activation.	COMMENT
LOW	…ends/nxp/aten_passes/decompose_split_to_slices_pass.py	175	# Check if split is even necessary - if not, remove it	COMMENT
LOW	backends/test/suite/runner.py	185	# Check if any undelegated ops are in the unsupported ops set.	COMMENT
LOW	backends/qualcomm/_passes/utils.py	302	# Check if pattern is valid	COMMENT
LOW	backends/qualcomm/scripts/build.sh	11	# Check if running on macOS/Darwin	COMMENT
LOW	backends/qualcomm/scripts/install_qnn_sdk.sh	11	# Check if ANDROID_NDK_ROOT is already set and valid	COMMENT
LOW	backends/qualcomm/scripts/install_qnn_sdk.sh	61	# Check if QNN_SDK_ROOT is already set and valid	COMMENT
LOW	backends/qualcomm/scripts/install_qnn_sdk.sh	132	# Check if already configured externally and valid.	COMMENT
LOW	backends/qualcomm/scripts/install_qnn_sdk.sh	212	# Check if we already have a local installation	COMMENT
LOW	backends/cuda/tests/test_cuda_partitioner.py	61	# Check if we have any tagged nodes	COMMENT
LOW	backends/cuda/tests/test_cuda_partitioner.py	65	# Check if all tagged nodes have the same tag	COMMENT
LOW	backends/cuda/tests/test_cuda_partitioner.py	181	# Check if this is a constant (param, buffer, or lifted tensor constant)	COMMENT
LOW	backends/cuda/tests/test_cuda_export.py	190	# Set module to eval mode to avoid batch norm running statistics mutations	COMMENT
LOW	backends/openvino/tests/ops/base_openvino_op_test.py	46	# Check if the number of partitions created matches the expected number of partitions	COMMENT
159 more matches not shown…

Fake / Example Data338 hits · 326 pts

Severity	File	Line	Snippet	Context
LOW	backends/apple/mps/mps_preprocess.py	160	node_visitors["placeholder"],	CODE
LOW	backends/apple/mps/mps_preprocess.py	165	"placeholder": MPSBackend.handle_placeholder,	CODE
LOW	backends/apple/mps/operators/node_visitor.py	409	node_visitors["placeholder"] = placeholder_output_visitor	CODE
LOW	backends/apple/mps/operators/node_visitor.py	431	if node.op == "placeholder" and not is_parameter(	CODE
LOW	…ckends/apple/coreml/compiler/enumerated_shape_utils.py	98	if node.op == "placeholder" and node.name in user_inputs:	CODE
LOW	backends/nxp/backend/node_format_inference.py	393	return node.op == "placeholder"	CODE
LOW	backends/nxp/backend/edge_program_converter.py	142	if node.op == "placeholder":	CODE
LOW	backends/nxp/edge_passes/remove_io_quant_ops_pass.py	31	n for n in graph.nodes if n.op == "placeholder" and n.name == user_input	CODE
LOW	…sion_passes/add_simulated_linear_bn_fusion_qat_pass.py	52	if node.op == "placeholder":	CODE
LOW	backends/nxp/quantizer/neutron_quantizer.py	383	if node.op == "placeholder" and len(node.users) > 0:	CODE
LOW	backends/test/graph_builder.py	30	if node.op != "placeholder":	CODE
LOW	backends/qualcomm/qnn_preprocess.py	105	"placeholder",	CODE
LOW	backends/qualcomm/builders/utils.py	77	return tensor.op == "placeholder" and not is_parameter(tensor, edge_program)	CODE
LOW	backends/qualcomm/builders/utils.py	88	if tensor.op == "placeholder" and is_buffer(edge_program, tensor):	CODE
LOW	backends/qualcomm/tests/test_qnn_delegate.py	6907	"placeholder",	STRING
LOW	backends/qualcomm/tests/test_qnn_delegate.py	7129	if n.op == "placeholder":	STRING
LOW	backends/qualcomm/tests/test_qnn_delegate.py	7998	"placeholder",	STRING
LOW	backends/qualcomm/utils/utils.py	132	node.op in ("placeholder", "output"),	CODE
LOW	backends/qualcomm/_passes/canonicalize_conv.py	68	if filter_arg.op == "placeholder"	CODE
LOW	backends/qualcomm/_passes/canonicalize_conv.py	128	if filter_arg.op == "placeholder"	CODE
LOW	backends/qualcomm/_passes/convert_mha_to_sha.py	563	"placeholder": _visit_placeholder,	CODE
LOW	backends/qualcomm/_passes/insert_io_qdq.py	50	n.op == "placeholder"	CODE
LOW	backends/qualcomm/_passes/recompose_rms_norm.py	23	return _is_node(node) and node.op == "placeholder"	CODE
LOW	backends/qualcomm/_passes/utils.py	210	elif decomposed_node.op == "placeholder":	CODE
LOW	backends/qualcomm/_passes/convert_linear_to_conv2d.py	123	weight_arg if weight_arg.op == "placeholder" else weight_arg.args[0]	CODE
LOW	…ds/qualcomm/_passes/lpai_partition_fallback_support.py	286	if q_node.op == "placeholder":	CODE
LOW	backends/qualcomm/_passes/i64_to_i32.py	138	elif n.op == "placeholder":	CODE
LOW	backends/qualcomm/debugger/format_outputs.py	105	if len(node.users) == 0 and node.op == "placeholder":	CODE
LOW	backends/qualcomm/debugger/format_outputs.py	143	if len(node.users) == 0 and node.op == "placeholder":	CODE
LOW	backends/qualcomm/debugger/format_outputs.py	183	if len(node.users) == 0 and node.op == "placeholder":	CODE
LOW	backends/qualcomm/partition/qnn_partitioner.py	201	if node.op != "placeholder" or len(node.users) != 0:	CODE
LOW	backends/qualcomm/quantizer/custom_annotation.py	446	if node.op == "placeholder" and "attention_kv_cache_past_" in node.name:	CODE
LOW	…backend_delegate_passes/permute_memory_formats_pass.py	46	if node.op != "placeholder"	CODE
LOW	…backend_delegate_passes/permute_memory_formats_pass.py	83	and exit_node_arg.op != "placeholder"	CODE
LOW	backends/cuda/passes/move_cond_predicate_to_cpu.py	75	pred_node.op == "placeholder"	CODE
LOW	backends/cuda/tests/test_cuda_partitioner.py	180	if node.op == "placeholder":	CODE
LOW	backends/cuda/tests/test_cuda_partitioner.py	241	placeholder = next(n for n in graph.nodes if n.op == "placeholder")	CODE
LOW	backends/cuda/tests/test_cuda_partitioner.py	288	if n.op == "placeholder" and is_buffer(exported_program, n)	CODE
LOW	backends/cuda/tests/test_cuda_partitioner.py	293	if n.op == "placeholder" and not is_buffer(exported_program, n)	CODE
LOW	backends/cuda/tests/test_cuda_partitioner.py	454	if n.op == "placeholder" and is_buffer(exported_program, n)	CODE
LOW⚡	backends/openvino/partitioner.py	172	elif node.op == "placeholder":	CODE
LOW⚡	backends/openvino/partitioner.py	173	if "placeholder" in pattern.op_types:	CODE
LOW⚡	backends/openvino/partitioner.py	182	const_node.op_types["placeholder"] = None	CODE
LOW	backends/samsung/enn_preprocess.py	95	"placeholder",	CODE
LOW	backends/samsung/utils/utils.py	28	return node.op == "placeholder" and not is_param_node(exported_program, node)	CODE
LOW	backends/samsung/_passes/annotate_scalar_parameters.py	40	if input_arg.op not in ("placeholder", "get_attr") or not is_param_node(	CODE
LOW	backends/samsung/partition/enn_partitioner.py	65	"placeholder",	CODE
LOW	backends/samsung/quantizer/annotator.py	49	if node.op == "placeholder":	CODE
LOW	backends/samsung/quantizer/annotator.py	232	if node.op == "placeholder":	CODE
LOW	backends/mediatek/partitioner.py	41	# Handle 'call_function' only cause 'placeholder' and 'output' cannot be tagged.	COMMENT
LOW	backends/mediatek/preprocess.py	34	if node.op != "placeholder":	CODE
LOW	…atek/_passes/decompose_scaled_dot_product_attention.py	18	if node.op == "placeholder":	CODE
LOW	…atek/_passes/decompose_scaled_dot_product_attention.py	47	if decomposed_node.op == "placeholder":	CODE
LOW	…atek/_passes/decompose_scaled_dot_product_attention.py	54	if decomposed_node.op == "placeholder":	CODE
LOW	backends/mediatek/quantizer/annotator.py	39	if node.op == "placeholder":	CODE
LOW	backends/transforms/convert_dtype_pass.py	50	if node.op == "placeholder":	CODE
LOW	backends/transforms/quantize_fused_convbn_bias_pass.py	35	if node.op == "placeholder"	CODE
LOW	backends/transforms/quantize_fused_convbn_bias_pass.py	65	placeholders = [n for n in graph.nodes if n.op == "placeholder"]	CODE
LOW	backends/transforms/quantize_fused_convbn_bias_pass.py	131	if n.op == "placeholder" and "val" in n.meta	CODE
LOW	backends/transforms/quantize_fused_convbn_bias_pass.py	151	if n.op == "placeholder" and "val" in n.meta	CODE
278 more matches not shown…

Hallucination Indicators28 hits · 315 pts

Severity	File	Line	Snippet	Context
CRITICAL	kernels/prim_ops/et_copy_index.cpp	59	// torch.ops.executorch.prim.add.int(iteration_index, 1, iteration_index)	COMMENT
CRITICAL	kernels/prim_ops/et_copy_index.cpp	60	// done_bool = torch.ops.executorch.prim.eq.int(iteration_index,	COMMENT
CRITICAL	backends/apple/coreml/test/tester.py	103	executorch.backends.test.harness.Tester.default_stage_classes()	CODE
CRITICAL	backends/apple/coreml/partition/coreml_partitioner.py	194	is_supported = ct.converters.mil.frontend.torch.is_torch_fx_node_supported(	CODE
CRITICAL	backends/test/suite/__init__.py	52	for name, f in executorch.backends.test.suite.flow.all_flows().items()	CODE
CRITICAL	backends/qualcomm/tests/tester.py	107	stage_classes = executorch.backends.test.harness.Tester.default_stage_classes() \| {	CODE
CRITICAL	backends/qualcomm/quantizer/rules.py	194	quantization_config.weight.observer_or_fake_quant_ctr.p.keywords.update(	CODE
CRITICAL	backends/qualcomm/quantizer/README.md	173	quantization_config.weight.observer_or_fake_quant_ctr.p.keywords.update(	CODE
CRITICAL	backends/qualcomm/quantizer/annotators/lpai_rules.py	535	quantization_config.weight.observer_or_fake_quant_ctr.p.keywords.update(	CODE
CRITICAL	backends/qualcomm/quantizer/annotators/htp_rules.py	965	quantization_config.weight.observer_or_fake_quant_ctr.p.keywords.update(	CODE
CRITICAL	backends/webgpu/test/tester.py	79	executorch.backends.test.harness.Tester.default_stage_classes()	CODE
CRITICAL	backends/openvino/test/tester/tester.py	76	executorch.backends.test.harness.Tester.default_stage_classes()	CODE
CRITICAL	backends/openvino/quantizer/quantizer.py	129	self._algo = nncf.quantization.algorithms.weight_compression.algorithm.WeightCompression(	CODE
CRITICAL	backends/vulkan/test/tester.py	79	executorch.backends.test.harness.Tester.default_stage_classes()	CODE
CRITICAL	backends/xnnpack/test/tester/tester.py	116	executorch.backends.test.harness.Tester.default_stage_classes()	CODE
CRITICAL	…kends/xnnpack/test/quantizer/test_pt2e_quantization.py	215	torch.ao.quantization.observer.PlaceholderObserver.with_args(	CODE
CRITICAL	…kends/xnnpack/test/quantizer/test_pt2e_quantization.py	306	torch.ao.quantization.observer.PlaceholderObserver.with_args(	CODE
CRITICAL	backends/mlx/test/tester.py	62	executorch.backends.test.harness.Tester.default_stage_classes()	CODE
CRITICAL	exir/tests/test_arg_validator.py	63	ops.edge.aten.add.Tensor.name(),	CODE
CRITICAL⚡	exir/tests/common.py	90	torch.ops.aten.max.default.tags.append(canonical)	CODE
CRITICAL⚡	exir/tests/common.py	91	torch.ops.aten.sum.default.tags.append(canonical)	CODE
CRITICAL⚡	exir/tests/common.py	92	torch.ops.aten.searchsorted.Tensor.tags.append(canonical)	CODE
CRITICAL⚡	exir/tests/common.py	93	torch.ops.aten.ones_like.default.tags.append(canonical)	CODE
CRITICAL⚡	exir/tests/common.py	94	torch.ops.aten.upsample_nearest2d.default.tags.append(canonical)	CODE
CRITICAL⚡	exir/tests/common.py	95	torch.ops.aten.index.Tensor.tags.append(canonical)	CODE
CRITICAL⚡	exir/tests/common.py	96	torch.ops.aten.addbmm.default.tags.append(canonical)	CODE
CRITICAL	exir/emit/_emitter.py	795	iteration_index) iteration_index = torch.ops.executorch.prim.add.int(iteration_index, 1,	STRING
CRITICAL	exir/emit/_emitter.py	796	iteration_index) done_bool = torch.ops.executorch.prim.eq.int(iteration_index, sym_size,	STRING

Modern Structural Boilerplate286 hits · 288 pts

Severity	File	Line	Snippet	Context
LOW	extension/pytree/__init__.py	12	logger = logging.getLogger(__name__)	CODE
LOW	extension/llm/runner/__init__.py	226	__all__ = [	CODE
LOW	extension/llm/export/__init__.py	9	__all__ = [	CODE
LOW	extension/llm/modules/attention.py	19	logger = logging.getLogger(__name__)	CODE
LOW	extension/llm/modules/__init__.py	16	__all__ = [	CODE
LOW	extension/llm/modules/turboquant/__init__.py	9	__all__ = [	CODE
LOW	extension/training/__init__.py	17	__all__ = [	CODE
LOW	extension/pybindings/portable_lib.py	34	logger = logging.getLogger(__name__)	CODE
LOW	extension/export_util/__init__.py	9	__all__ = [	CODE
LOW	.ci/scripts/gather_benchmark_configs.py	274	def set_output(name: str, val: Any) -> None:	CODE
LOW	.ci/scripts/gather_test_models.py	78	def set_output(name: str, val: Any) -> None:	CODE
LOW	backends/apple/mps/__init__.py	17	__all__ = [	CODE
LOW	backends/apple/mps/operators/__init__.py	42	__all__ = [	CODE
LOW	backends/apple/mps/partition/__init__.py	18	__all__ = [	CODE
LOW	backends/apple/metal/passes/__init__.py	11	__all__ = ["DecomposeLinearPass"]	CODE
LOW	backends/apple/coreml/recipes/__init__.py	14	__all__ = [	CODE
LOW	backends/apple/coreml/partition/__init__.py	7	__all__ = [	CODE
LOW	backends/apple/coreml/partition/coreml_partitioner.py	32	logger = logging.getLogger(__name__)	CODE
LOW	backends/apple/coreml/quantizer/__init__.py	7	__all__ = [	CODE
LOW	backends/apple/coreml/compiler/coreml_preprocess.py	35	logger = logging.getLogger(__name__)	CODE
LOW	backends/apple/coreml/compiler/__init__.py	7	__all__ = [	CODE
LOW	backends/nxp/tests/config_importer.py	8	logger = logging.getLogger(__name__)	CODE
LOW	backends/nxp/tests/nsys_testing.py	55	logger = logging.getLogger(__name__)	CODE
LOW	backends/nxp/tests/outputs_dir_importer.py	8	logger = logging.getLogger(__name__)	CODE
LOW	backends/nxp/backend/node_format_inference.py	41	logger = logging.getLogger(__name__)	CODE
LOW	backends/nxp/backend/neutron_map.py	227	def _update_neutron_subgraphs_info(self, extracted_graph: str) -> None:	CODE
LOW	…ckends/nxp/backend/ir/tflite_generator/tflite_model.py	31	logger = logging.getLogger(__name__)	CODE
LOW	backends/nxp/backend/ir/tflite_generator/meta/meta.py	21	logger = logging.getLogger(__name__)	CODE
LOW	…r/converter/node_converters/ops_converters/__init__.py	111	__all__ = [	CODE
LOW	…n_passes/simulated_linear_bn_fusion_passes/__init__.py	8	__all__ = [	CODE
LOW	backends/test/harness/__init__.py	3	__all__ = ["Tester"]	CODE
LOW	backends/test/harness/stages/serialize.py	11	logger = logging.getLogger(__name__)	CODE
LOW	backends/test/harness/stages/__init__.py	11	__all__ = [	CODE
LOW	backends/test/suite/__init__.py	19	logger = logging.getLogger(__name__)	CODE
LOW	backends/test/suite/flow.py	15	logger = logging.getLogger(__name__)	CODE
LOW	backends/test/suite/flows/portable.py	6	logger = logging.getLogger(__name__)	CODE
LOW	backends/test/suite/flows/xnnpack.py	14	logger = logging.getLogger(__name__)	CODE
LOW	backends/qualcomm/qnn_preprocess.py	44	logger = logging.getLogger(__name__)	CODE
LOW	backends/qualcomm/builders/__init__.py	122	__all__ = [	CODE
LOW	backends/qualcomm/_passes/__init__.py	71	__all__ = [	CODE
LOW	…ends/qualcomm/_passes/lift_constant_scalar_operands.py	126	def _update_node(self, node: fx.Node, tensor_args: Dict) -> None:	CODE
LOW	backends/qualcomm/_passes/i64_to_i32.py	112	def _update_meta(self, node: torch.fx.node) -> None:	CODE
LOW	backends/qualcomm/_passes/backends/htp/__init__.py	9	__all__ = [	CODE
LOW	backends/qualcomm/_passes/backends/gpu/__init__.py	9	__all__ = [	CODE
LOW	backends/qualcomm/_passes/backends/lpai/__init__.py	9	__all__ = [	CODE
LOW	backends/qualcomm/genai_pipeline/__init__.py	32	__all__ = [	CODE
LOW	backends/qualcomm/genai_pipeline/configs/__init__.py	32	__all__ = [	CODE
LOW	backends/qualcomm/recipes/__init__.py	16	__all__ = ["QNNRecipeProvider", "QNNRecipeType"]	CODE
LOW	backends/qualcomm/scripts/download_qnn_sdk.py	21	logger = logging.getLogger(__name__)	CODE
LOW	…ds/qualcomm/debugger/qcom_numerical_comparator_base.py	28	logger = logging.getLogger(__name__)	CODE
LOW	backends/qualcomm/partition/qnn_partitioner.py	48	logger = logging.getLogger(__name__)	CODE
LOW	backends/qualcomm/custom_op/annotator.py	20	logger = logging.getLogger(__name__)	CODE
LOW	backends/qualcomm/quantizer/quantizer.py	75	__all__ = [	CODE
LOW	backends/qualcomm/quantizer/quantizer.py	674	def set_block_size_map(self, block_size_map: Dict[str, Tuple]) -> None:	CODE
LOW	backends/cuda/coalesced_int4_tensor.py	59	__all__ = [	CODE
LOW	backends/cuda/dp4a_planar_int6_tensor.py	60	__all__ = [	CODE
LOW	backends/cuda/dp4a_planar_int5_tensor.py	58	__all__ = [	CODE
LOW	backends/cuda/passes/replace_int64_floordiv.py	23	logger = logging.getLogger(__name__)	CODE
LOW	backends/cuda/triton/__init__.py	15	__all__ = [	CODE
LOW	backends/cuda/triton/replacement_pass.py	21	logger = logging.getLogger(__name__)	CODE
226 more matches not shown…

Verbosity Indicators156 hits · 286 pts

Severity	File	Line	Snippet	Context
LOW⚡	install_executorch.py	225	# Step 1: Install core dependencies first	COMMENT
LOW⚡	install_executorch.py	228	# Step 2: Install core package	COMMENT
LOW	install_executorch.py	249	# Step 3: Extra (optional) packages that is only useful for running examples.	COMMENT
LOW	extension/module/test/resources/gen_bundled_program.py	14	# Step 1: ExecuTorch Program Export	COMMENT
LOW	extension/module/test/resources/gen_bundled_program.py	54	# Step 2: Construct MethodTestSuite for Each Method	COMMENT
LOW⚡	extension/module/test/resources/gen_bundled_program.py	86	# Step 3: Generate BundledProgram	COMMENT
LOW⚡	extension/module/test/resources/gen_bundled_program.py	89	# Step 4: Serialize BundledProgram to flatbuffer.	COMMENT
LOW⚡	extension/gguf_util/load_gguf.py	96	# Step 1: Build GGUFModelArgs	COMMENT
LOW⚡	extension/gguf_util/load_gguf.py	100	# Step 2: Build GGUFWeights	COMMENT
LOW⚡	extension/gguf_util/convert_main.py	32	# Step 1: Load the GGUF file	COMMENT
LOW⚡	extension/gguf_util/convert_main.py	35	# Step 2: Convert the GGUF model to PTE	COMMENT
LOW	extension/gguf_util/convert_main.py	47	# Step 3: Save the PTE program so that	COMMENT
LOW⚡	extension/gguf_util/converters/llama_converter.py	110	# Step 1: Create the PyTorch model	COMMENT
LOW⚡	extension/gguf_util/converters/llama_converter.py	116	# Step 2: Load the weights into the PyTorch model	COMMENT
LOW⚡	extension/gguf_util/converters/llama_converter.py	120	# Step 3: Export to ExecuTorch	COMMENT
LOW	extension/image/image_processor.cpp	323	// Step 1: orientation (orient -> ROI -> resize). Produce an oriented copy of	COMMENT
LOW	extension/image/image_processor.cpp	347	// Step 2: ROI crop (pointer arithmetic).	COMMENT
LOW	extension/image/image_processor.cpp	368	// Step 3: Swizzle BGRA/RGBA → RGB (alpha discarded).	COMMENT
LOW	extension/image/image_processor.cpp	382	// Step 4: Resize.	COMMENT
LOW	extension/image/image_processor.cpp	403	// Step 5: Normalize + layout into the caller's CHW output (padded).	COMMENT
LOW	extension/image/image_processor_apple.cpp	694	// Step 2: orientation. Rotate the BGRA buffer (vImage) so ROI/resize run in	COMMENT
LOW⚡	extension/image/image_processor_apple.cpp	720	// Step 3: ROI crop (pointer arithmetic on BGRA data).	COMMENT
LOW⚡	extension/image/image_processor_apple.cpp	723	// Step 4: resize. Letterbox padding is applied during normalization.	COMMENT
LOW	extension/image/image_processor_apple.cpp	752	// Step 5: normalize BGRA → CHW float buffer.	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	61	# Step 1: Checkout old version of ExecuTorch	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	66	# Step 2: Install the old version	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	75	# Step 3: Export model	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	124	# Step 4: Checkout current version (main)	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	130	# Step 5: Install current version	COMMENT
LOW⚡	.ci/scripts/test_coreml_bc.sh	135	# Step 6: Run the old pte file	COMMENT
LOW	backends/qualcomm/_passes/decompose_diagonal.py	95	# Step 1: Permute dim1, dim2 to last positions (if needed)	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_diagonal.py	107	# Step 2: Reshape [..., M, N] -> [..., M*N]	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_diagonal.py	117	# Step 3: Compute flat diagonal indices	COMMENT
LOW	backends/qualcomm/_passes/decompose_diagonal.py	142	# Step 4: index_select on last dim	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_cdist.py	19	# Step 1: Compute differences	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_cdist.py	22	# Step 2: Square differences	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_cdist.py	25	# Step 3: Sum of squares	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_cdist.py	28	# Step 4: Square root	COMMENT
LOW	backends/qualcomm/_passes/decompose_var.py	106	# Step 1: mean_x = mean(x, dim, keepdim=True)	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_var.py	118	# Step 2: diff = x - mean_x	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_var.py	126	# Step 3: sq = diff * diff (more efficient than pow(diff, 2))	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_var.py	134	# Step 4: var = mean(sq, dim, keepdim)	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_var.py	140	# Step 5: Apply correction factor if needed	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_div_mode.py	55	# Step 1: div_result = div(x, y)	COMMENT
LOW⚡	backends/qualcomm/_passes/decompose_div_mode.py	61	# Step 2: Apply rounding mode if needed	COMMENT
LOW	backends/qualcomm/_passes/convert_linear_to_conv2d.py	134	# Step 1: reshape input	COMMENT
LOW	backends/qualcomm/_passes/convert_linear_to_conv2d.py	168	# Step 2: reshape weight	COMMENT
LOW	backends/qualcomm/_passes/convert_linear_to_conv2d.py	202	# Step 3: restore shape	COMMENT
LOW	backends/cuda/runtime/cuda_backend.cpp	1028	// Step 1: Enumerate constants and partition into cached/uncached	COMMENT
LOW	backends/cuda/runtime/cuda_backend.cpp	1071	// Step 2: Load uncached constants from blob (if any).	COMMENT
LOW	backends/cuda/runtime/cuda_backend.cpp	1157	// Step 3: Point the container to cached tensors via user_managed pairs	COMMENT
LOW	backends/vulkan/test/utils.py	892	# Step 1: Export model to get edge_program and extract operators	COMMENT
LOW	backends/vulkan/test/utils.py	907	# Step 2: Scan edge_program.graph_module to obtain unique operators and their frequencies	COMMENT
LOW	…nds/vulkan/runtime/graph/ops/impl/Q8taConv2dIm2Col.cpp	300	// Step 1: Perform im2col transformation	COMMENT
LOW	…nds/vulkan/runtime/graph/ops/impl/Q8taConv2dIm2Col.cpp	313	// Step 2: Perform pointwise convolution on the im2col result	COMMENT
LOW⚡	backends/arm/cmsis_pack/scripts/build_pack.sh	65	# Step 1: Copy sources from the repo / build tree into the pack layout	COMMENT
LOW⚡	backends/arm/cmsis_pack/scripts/build_pack.sh	72	# Step 2: Generate RegisterAllKernels.cpp with #ifdef-guarded registrations	COMMENT
LOW⚡	backends/arm/cmsis_pack/scripts/build_pack.sh	78	# Step 3: Generate PDSC from template	COMMENT
LOW⚡	backends/arm/cmsis_pack/scripts/build_pack.sh	91	# Step 4: Copy static files (LICENSE, docs)	COMMENT
LOW⚡	backends/arm/cmsis_pack/scripts/build_pack.sh	101	# Step 5: Create .pack archive (a zip file)	COMMENT
96 more matches not shown…

Docstring Block Structure48 hits · 240 pts

Severity	File	Line	Snippet	Context
HIGH	extension/llm/modules/attention.py	209	Args: x (torch.Tensor): input tensor with shape [b x s_x x d] for the query y (Optional	STRING
HIGH	extension/llm/modules/kv_cache.py	69	Update KV cache with the new ``k_val``, ``v_val`` and return the updated cache. Note: When updating	STRING
HIGH	extension/flat_tensor/serialize/serialize.py	146	Tries to read an flat_tensor header from the provided data. Does not validate that the header is well-formed. C	STRING
HIGH	.ci/scripts/gather_benchmark_configs.py	164	Generate a list of compatible benchmark configurations for a given model name and target OS. Args: mod	STRING
HIGH	.ci/scripts/test_huggingface_optimum_model.py	233	Determine the tokenizer path based on files saved by tokenizer.save_pretrained(). Args: model_dir: The	STRING
HIGH	backends/qualcomm/utils/utils.py	1157	Helper function generating compiler specs for Qualcomm AI Engine Direct Args: soc_model: The SoC you p	STRING
HIGH	backends/cuda/triton/kernels/sdpa.py	115	Validate dimensions and return shape info. Args: query: Query tensor [B, H_q, L_q, D] key: Key	STRING
HIGH	backends/transforms/utils.py	32	Helper function to create a fake tensor using the fake_mode from existing nodes in the graph. Args: gr	STRING
HIGH	…kends/xnnpack/serialization/xnnpack_graph_serialize.py	134	Converts the given bytes into an XNNHeader object. We check that the magic and length is valid, but do	STRING
HIGH	…kends/xnnpack/serialization/xnnpack_graph_serialize.py	263	Returns the input followed by enough zero bytes to become the requested length. Args: data: The data to pad	STRING
HIGH	backends/arm/test/runner_utils.py	200	Get output QuantizationParams from a program. Args: output_nodes (list(Node)): A list of output nodes to ge	STRING
HIGH	backends/arm/operators/op_tosa_rescale.py	28	Derive integer multipliers and shifts from floating-point scales. TOSA uses the RESCALE operation to scale between	STRING
HIGH	backends/arm/operators/node_visitor.py	185	Define a TOSA operator node. Args: node (torch.fx.Node): FX node being lowered. tosa_gr	STRING
HIGH	backends/arm/tosa/backend.py	186	Lower an exported program to a TOSA flatbuffer. Apply Arm transformation passes to ``edge_program``, then walk	STRING
HIGH	backends/arm/tosa/utils.py	26	Determine whether the fake tensors share a broadcastable shape. Args: fake_tensors (list[FakeTensor]): Fake	STRING
HIGH	backends/arm/tosa/utils.py	65	Broadcast the FX nodes to a shared shape inside the TOSA graph. This emits the reshape and tile operators needed to	STRING
HIGH	backends/arm/tosa/specification.py	192	Create a specification from a standard string format. Example: ``TOSA-1.00.0+INT+FP+int4+cf``. Args:	STRING
HIGH	backends/arm/tosa/mapping.py	84	Map a ``torch.dtype`` to a ``ts.DType``. Args: data_type (torch.dtype): PyTorch dtype to convert. Retu	STRING
HIGH	backends/arm/tosa/mapping.py	129	Extract dtype, shape, and dimension order from FX metadata. Args: meta (dict): FX node ``meta`` containing	STRING
HIGH	backends/arm/_passes/rewrite_conv_pass.py	71	Adjust padding to satisfy TOSA's integer output-size requirement. Torch ``Conv2d`` does not require the result	STRING
HIGH	backends/arm/_passes/rewrite_upsample.py	50	Compute resize coefficients for a single spatial dimension. Args: input_size (int \| torch.SymInt):	STRING
HIGH	backends/arm/scripts/aot_arm_compiler.py	209	Resolve a model name into a model instance and example inputs. Args: model_name: Identifier for the model.	STRING
HIGH	backends/arm/quantizer/quantization_config.py	140	Get the derived or validated bias spec. For conv/linear ops, derive bias qparams from the input/weight observer	STRING
HIGH	backends/arm/quantizer/quantization_config.py	165	Compute bias scale/zero-point from activation/weight observers. Expect two observers or fake-quantize modul	STRING
HIGH	backends/arm/vgf/compile_spec.py	74	Run VGF environment preflight checks. By default this validates only AoT/export prerequisites. Runtime and	STRING
HIGH	backends/mlx/partitioner.py	263	Partition the edge program for MLX delegation. Args: edge_program: The ExportedProgram to	STRING
HIGH	backends/mlx/pattern_utils.py	106	One step in a backward walk through the graph. Used with walk_back() to define pattern chains. Supports both e	STRING
HIGH	backends/mlx/pattern_utils.py	201	Walk backwards through a chain of ops, matching against a pattern. Starting from node, try to match each ste	STRING
HIGH	devtools/inspector/_inspector.py	1178	Resolve the reference graph module to use for AOT operations. This method centralizes the logic for de	STRING
HIGH	…ctor/numerical_comparator/numerical_comparator_base.py	166	Compare two outputs, handling both sequence and non-sequence cases. Args: a: The first intermediate	STRING
HIGH	examples/qualcomm/oss_scripts/llama/tokenizer.py	143	Validate and normalize a multi-turn prompt sequence, then prepare it into a message list. This	STRING
HIGH	…/aot_utils/llm_utils/tokenizers_/tokenization_utils.py	408	Add a list of new tokens to the tokenizer class. If the new tokens are not in the vocabulary, they are added to	STRING
HIGH	…utils/llm_utils/tokenizers_/tokenization_utils_base.py	61	Import the DecodeError from protobuf if available. Args: _ (Any): Placeholder argument. Returns:	STRING
HIGH	…utils/llm_utils/tokenizers_/tokenization_utils_base.py	913	Adds special tokens. Add a dictionary of special tokens (eos, pad, cls, etc.) to the encoder and link them to c	STRING
HIGH	…utils/llm_utils/tokenizers_/tokenization_utils_base.py	1038	Add tokens. Add a list of new tokens to the tokenizer class. If the new tokens are not in the vocabulary, they	STRING
HIGH	…ples/mediatek/aot_utils/llm_utils/tokenizers_/utils.py	374	Tries to locate a file in a local folder and repo, downloads and cache it if necessary. Args: path_or_repo_	STRING
HIGH	examples/models/model_factory.py	23	Create an instance of a model class that implements EagerModelBase and retrieve related data. Args:	STRING
HIGH	scripts/pick_doc_commits.py	46	Runs a git command and returns its stdout as a list of lines. Prints the command and its output to debug_log() if v	STRING
HIGH	exir/backend/backend_api.py	74	Add overloaded implementations for to_backend: :: def to_backend( backend_id: str, edg	STRING
HIGH	exir/backend/test/backend_with_compiler_demo.py	30	An example implementation to lower a module. Currently this example only supports the sin operator. The exa	STRING
HIGH	exir/_serialize/_program.py	92	Inserts a header just after the magic string of the provided flatbuffer data. Args: flatbuffer_data: The in	STRING
HIGH	exir/_serialize/_program.py	190	Tries to read an extended header from the provided data. Does not validate that the header is well-formed. Call	STRING
HIGH	exir/_serialize/padding.py	11	Returns the input followed by enough zero bytes to become the requested length. Args: data: The data to pad	STRING
HIGH	export/export.py	477	Get the ExportedProgram for a specific method after torch export. Args: method_name: Name	STRING
HIGH	export/export.py	592	Get the example input for a specific method. Args: method_name: Name of the method to get	STRING
HIGH	export/export.py	621	Run a specific method with the given inputs. Args: method_name: Name of the method to run,	STRING
HIGH	export/target_recipes.py	63	Get iOS-optimized recipe for specified hardware configuration. Supported configurations: - 'ios-arm64-core	STRING
HIGH	export/target_recipes.py	122	Get Android-optimized recipe for specified hardware configuration. Supported configurations: - 'android-ar	STRING

AI Structural Patterns247 hits · 215 pts

Severity	File	Line	Context
LOW	extension/llm/custom_ops/custom_ops.py	361	CODE
LOW	extension/llm/custom_ops/test_sdpa_with_kv_cache.py	45	CODE
LOW	extension/llm/custom_ops/preprocess_custom_ops.py	56	CODE
LOW	extension/llm/custom_ops/test_quantized_sdpa.py	155	CODE
LOW	extension/llm/custom_ops/test_quantized_sdpa.py	220	CODE
LOW	extension/llm/export/builder.py	79	CODE
LOW	extension/audio/mel_spectrogram.py	44	CODE
LOW	…ripts/benchmark_tooling/get_benchmark_analysis_data.py	603	CODE
LOW	test/end2end/test_end2end.py	457	CODE
LOW	backends/apple/mps/test/test_mps.py	719	CODE
LOW	backends/apple/mps/test/test_mps.py	735	CODE
LOW	backends/apple/mps/test/test_mps.py	751	CODE
LOW	backends/apple/mps/test/test_mps.py	768	CODE
LOW	backends/apple/mps/test/test_mps.py	868	CODE
LOW	backends/apple/mps/test/test_mps.py	887	CODE
LOW	backends/apple/mps/test/test_mps.py	904	CODE
LOW	backends/apple/mps/test/test_mps.py	920	CODE
LOW	backends/apple/mps/test/test_mps.py	939	CODE
LOW	backends/apple/mps/test/test_mps.py	956	CODE
LOW	backends/apple/mps/test/test_mps.py	972	CODE
LOW	backends/apple/mps/test/test_mps.py	988	CODE
LOW	backends/apple/mps/test/test_mps.py	1006	CODE
LOW	backends/apple/mps/test/test_mps.py	1022	CODE
LOW	backends/apple/mps/test/test_mps.py	1038	CODE
LOW	backends/apple/mps/test/test_mps.py	1055	CODE
LOW	backends/apple/mps/test/test_mps.py	1277	CODE
LOW	backends/apple/mps/test/test_mps.py	1293	CODE
LOW	backends/apple/mps/test/test_mps.py	1310	CODE
LOW	backends/apple/mps/test/test_mps.py	1326	CODE
LOW	backends/apple/mps/test/test_mps.py	1342	CODE
LOW	backends/apple/mps/test/test_mps.py	1359	CODE
LOW	backends/apple/mps/test/test_mps.py	1497	CODE
LOW	backends/apple/mps/test/test_mps_binary_ops.py	137	CODE
LOW	backends/apple/metal/tests/test_modules.py	450	CODE
LOW	backends/nxp/nxp_backend.py	155	CODE
LOW	backends/nxp/tests/models.py	15	CODE
LOW	backends/nxp/tests/models.py	44	CODE
LOW	backends/nxp/tests/models.py	75	CODE
LOW	backends/nxp/tests/models.py	104	CODE
LOW	backends/nxp/tests/executorch_pipeline.py	177	CODE
LOW	backends/nxp/tests/executorch_pipeline.py	271	CODE
LOW	backends/nxp/tests/executors.py	310	CODE
LOW	backends/nxp/tests/nsys_testing.py	401	CODE
LOW	…ir/tflite_generator/builtin_options/conv_3d_options.py	29	CODE
LOW	backends/test/harness/tester.py	228	CODE
LOW	backends/test/harness/tester.py	316	CODE
LOW	backends/test/harness/tester.py	538	CODE
LOW	backends/test/suite/operators/test_conv1d.py	21	CODE
LOW	backends/test/suite/operators/test_conv3d.py	22	CODE
LOW	backends/test/suite/operators/test_conv2d.py	22	CODE
LOW	backends/test/suite/operators/test_convtranspose2d.py	22	CODE
LOW	backends/test/suite/operators/test_convtranspose3d.py	22	CODE
LOW	backends/test/suite/operators/test_convtranspose1d.py	22	CODE
LOW	backends/test/suite/operators/test_lstm.py	27	CODE
LOW	backends/qualcomm/export_utils.py	502	CODE
LOW	backends/qualcomm/export_utils.py	652	CODE
LOW	backends/qualcomm/tests/models.py	963	CODE
LOW	backends/qualcomm/tests/utils.py	350	CODE
LOW	backends/qualcomm/tests/utils.py	664	CODE
LOW	backends/qualcomm/tests/utils.py	755	CODE
187 more matches not shown…

AI Slop Vocabulary47 hits · 107 pts

Severity	File	Line	Snippet	Context
MEDIUM	kernels/portable/test/op_upsample_bilinear2d_aa_test.py	11	# The comprehensive C++ test suite in op_upsample_bilinear2d_aa_test.cpp provides	COMMENT
MEDIUM	kernels/portable/cpu/util/broadcast_util.cpp	150	// Essentially, we can broadcast broadcast_from if it meets three conditions	COMMENT
MEDIUM	kernels/optimized/utils/llvmMathExtras.h	697	/// Essentially, it is a floor operation across the domain of powers of two.	COMMENT
MEDIUM	kernels/optimized/utils/llvmMathExtras.h	705	/// Essentially, it is a ceil operation across the domain of powers of two.	COMMENT
LOW	extension/llm/custom_ops/custom_ops.py	252	# workaround. Should we just return cache instead? But I am afraid that	COMMENT
LOW	extension/llm/custom_ops/custom_ops.py	272	# workaround. Should we just return cache instead? But I am afraid that	COMMENT
LOW	extension/llm/custom_ops/custom_ops.py	404	# no copies, just pass-through	COMMENT
LOW	extension/llm/custom_ops/custom_ops.py	421	# Alias is logically a no-op – just pass them through.	COMMENT
LOW	extension/llm/export/export_passes.py	135	# Ignore specified mask simply set the is_causal flag.	COMMENT
MEDIUM	.ci/scripts/gather_test_models.py	115	# comprehensive E2E test of llava.	COMMENT
MEDIUM	…ripts/benchmark_tooling/analyze_benchmark_stability.py	1245	"""Generate a comprehensive summary report"""	STRING
LOW	backends/vulkan/utils.py	1051	# or non tensor argument. In this case, just return default storage and	COMMENT
MEDIUM	…vulkan/runtime/graph/ops/impl/QuantizedLinearQCSNW.cpp	410	// TODO(ssjia): Determine a more robust heuristic to determine when the coop	COMMENT
MEDIUM	backends/vulkan/runtime/graph/ops/impl/Q4gswLinear.cpp	127	// shape, but (8,8) is within 3% there and is robust across	COMMENT
LOW	backends/vulkan/_passes/tag_memory_meta_pass.py	75	# Case where the arg node has not been touched yet; in this case, simply set it and	COMMENT
LOW	backends/arm/process_node.py	162	# Get item nodes just add tensors, no node visitor is needed.	COMMENT
MEDIUM	backends/arm/test/runner_utils.py	392	# This is a hack since Q/DQ ops does not handle channels last input correctly: the simplest and most robust	COMMENT
LOW	backends/arm/operators/op_while.py	76	# If we have more inputs than outputs, we can just add missing output tensors.	COMMENT
LOW	backends/arm/tosa/utils.py	94	# Already in the right shape we can just add it to the list.	COMMENT
LOW	backends/arm/operator_support/ethos_u55_support.py	106	# If we couldn't determine dtype, just return ok.	COMMENT
MEDIUM	backends/arm/scripts/aot_arm_compiler.py	779	# To create a bundled program, we first create every test cases from input. We leverage eager model	COMMENT
LOW	backends/cadence/aot/compiler_utils.py	197	# If x was quantized per-tensor, simply create a tensor out of the scalar	COMMENT
LOW	backends/mlx/builder/op_helpers.py	153	# Both static - just return the sum	COMMENT
MEDIUM	backends/aoti/utils.h	33	// Map int32_t dtype to ExecuTorch ScalarType (robust version of hardcoded	COMMENT
MEDIUM	…ntime/core/portable_type/c10/c10/util/llvmMathExtras.h	691	/// Essentially, it is a floor operation across the domain of powers of two.	COMMENT
MEDIUM	…ntime/core/portable_type/c10/c10/util/llvmMathExtras.h	699	/// Essentially, it is a ceil operation across the domain of powers of two.	COMMENT
MEDIUM	…source/tutorials_source/devtools-debugging-tutorial.py	18	# This tutorial will show a full end-to-end flow of how to utilize the Developer Tools to debug a model	COMMENT
MEDIUM⚡	…urce/tutorials_source/devtools-integration-tutorial.py	20	# This tutorial will show a full end-to-end flow of how to utilize the Developer Tools to profile a model.	COMMENT
LOW	devtools/inspector/_inspector.py	585	# If it's not a delegated op then we can just use the raw time values	COMMENT
MEDIUM	examples/qualcomm/qaihub_scripts/utils/export.py	74	# leverage QNN pybind interface to retrieve tensor encodings	COMMENT
MEDIUM	examples/qualcomm/qaihub_scripts/utils/export.py	169	# leverage SimpleADB for model library conversion	COMMENT
MEDIUM	examples/qualcomm/qaihub_scripts/utils/export.py	315	# leverage SimpleADB for e2e inference	COMMENT
MEDIUM	examples/qualcomm/util_scripts/cli.py	111	# leverage QNN pybind interface to retrieve tensor encodings	COMMENT
MEDIUM	examples/qualcomm/util_scripts/cli.py	363	# leverage SimpleADB for e2e inference	COMMENT
MEDIUM	examples/models/llama/install_requirements.sh	12	# Install lm-eval for Model Evaluation with lm-evalution-harness.	COMMENT
MEDIUM	examples/models/llama/export_llama_lib.py	96	# All models that leverage the transformer architecture defined in llama_transformer.py.	COMMENT
MEDIUM	examples/models/llama/eval_llama_lib.py	363	# See https://github.com/EleutherAI/lm-evaluation-harness/pull/1998/files	COMMENT
LOW	…/models/llama/source_transformation/custom_kv_cache.py	185	# When returning float values we just use the last value	COMMENT
MEDIUM	examples/devtools/scripts/export_bundled_program.py	94	# To create a bundled program, we first create every test cases from input. We leverage eager model	COMMENT
LOW⚡	.github/scripts/propose_ghstack_orig_pr.py	91	# If there's more than one Differential Revision, let's just return empty	COMMENT
LOW	.github/scripts/trymerge.py	737	# so let's just use main instead	STRING
LOW	exir/lowered_backend_module.py	698	# the delegate will just return the result of the	COMMENT
LOW	exir/backend/backend_api.py	789	# this edge program wasn't partitioned, so we can just return it as is	COMMENT
LOW	exir/serde/serialize.py	131	# Replace the edge op with the original ATen op so that we can just call into	COMMENT
LOW	exir/serde/serialize.py	490	# Replace the edge op with the original ATen op so that we can just call into	COMMENT
LOW	exir/serde/schema.py	217	# (ex. cond, map) where single tensor returns will just return a single	COMMENT
MEDIUM	export/tests/test_export_session.py	468	# Create comprehensive recipes	COMMENT

Modern AI Meta-Vocabulary22 hits · 64 pts

Severity	File	Line	Snippet	Context
MEDIUM⚡	…/cuda/runtime/shims/tests/gen_plain_mm_test_vectors.py	349	# Shared emit / check scaffolding.	COMMENT
MEDIUM	backends/vulkan/_passes/squeeze_unsqueeze_inputs.py	45	# Prefer not to introduce additional orchestration ops by default	COMMENT
MEDIUM	codegen/gen.py	137	# and the scaffolding to call into the dispatcher from these functions.	STRING
MEDIUM	…es/qualcomm/oss_scripts/llama/wrappers/llm_wrappers.py	347	# get embedding model	COMMENT
MEDIUM	…utils/llm_utils/tokenizers_/tokenization_utils_base.py	1643	see the RAG section of the [chat templating guide](https://huggingface.co/docs/transformers/main/en/chat	STRING
MEDIUM	examples/models/qwen3_5_moe/export.py	1015	# The wrapper.cpp is pure kernel launch orchestration — no heavy compute.	COMMENT
MEDIUM	examples/llm_server/python/openai_transcript.py	35	# The assistant header that precedes a turn's generation scaffold + content.	COMMENT
MEDIUM	examples/llm_server/python/openai_transcript.py	37	# A scaffold region is exactly empty (history strips it before the last user) or	COMMENT
MEDIUM	examples/llm_server/python/openai_transcript.py	130	# No scaffold for this turn's mode/template: nothing to reproduce, so	COMMENT
MEDIUM	examples/llm_server/python/openai_transcript.py	238	# turn (few-shot exemplar, pre-seeded turn, reused session) shifts that	COMMENT
MEDIUM	examples/llm_server/python/openai_transcript.py	298	# Splice ids and normalize each turn's scaffold; None => ambiguous region.	COMMENT
MEDIUM	examples/llm_server/python/chat_template.py	117	# Cache of the (deterministic) generation scaffold per resolved mode, so	COMMENT
MEDIUM	examples/llm_server/python/serving_chat.py	306	# possibly a clamp to the context window) — trust it over the token-count	COMMENT
MEDIUM	examples/llm_server/python/serving_chat.py	447	# The generation scaffold the worker will prefill ahead of this turn's	COMMENT
MEDIUM	examples/llm_server/python/serving_chat.py	450	# reproduces the exact resident scaffold even if the mode changes between	COMMENT
MEDIUM	examples/llm_server/python/worker_client.py	46	# stop) or "length" (ran to max_new, possibly clamped to the context window).	COMMENT
MEDIUM	examples/llm_server/python/tests/test_sessions.py	398	# A client-injected assistant turn we never generated (few-shot exemplar /	COMMENT
MEDIUM	examples/llm_server/python/tests/test_contract.py	323	# context window, so the token-count heuristic can't tell a real stop from a	COMMENT
MEDIUM	…s/llm_server/python/tests/test_warm_resume_scaffold.py	160	# History stripped the scaffold; the fix inserts exactly one copy.	COMMENT
MEDIUM	…s/llm_server/python/tests/test_warm_resume_scaffold.py	192	# (wrong scaffold) and not append a second one (double-insert).	COMMENT
MEDIUM	…s/llm_server/python/tests/test_warm_resume_scaffold.py	240	# Regression: a no-scaffold template whose assistant header isn't the	COMMENT
MEDIUM	…s/llm_server/python/tests/test_warm_resume_scaffold.py	722	# on them, so a template whose post-header scaffold depends on tools gets the	COMMENT

AI Response Leakage4 hits · 30 pts

Severity	File	Line	Snippet	Context
HIGH	backends/vulkan/runtime/graph/ops/impl/BatchNorm.cpp	48	// In this implementation, we pack the weights along the x dimension, and	COMMENT
HIGH	…urce/examples-end-to-end-to-lower-model-to-delegate.md	53	# The submodule runs in a specific backend. In this example, `BackendWithCompilerDemo` backend	COMMENT
HIGH	examples/devtools/CMakeLists.txt	9	# support. In this example we link devtools and bundled_program libraries into	COMMENT
HIGH	exir/backend/test/test_backends_lifted.py	515	# the submodule runs in a specific backend. In this example, `BackendWithCompilerDemo` backend	COMMENT

Example Usage Blocks16 hits · 25 pts

Severity	File	Line	Snippet	Context
LOW	extension/kernel_util/make_boxed_from_unboxed_functor.h	13	/// Example usage:	COMMENT
LOW	.ci/scripts/build-cadence-xtensa.sh	16	# Usage:	COMMENT
LOW	.ci/scripts/setup-xtensa-tools.sh	17	# Usage:	COMMENT
LOW	backends/apple/metal/tests/run_metal_test.sh	9	# Usage:	COMMENT
LOW	backends/xnnpack/utils/gen_xnnpack_constants.sh	8	# Usage:	COMMENT
LOW	backends/arm/cmsis_pack/scripts/build_pack.sh	14	# Usage:	COMMENT
LOW⚡	backends/arm/cmsis_pack/scripts/copy_sources.sh	14	# Usage:	COMMENT
LOW	backends/arm/scripts/utils.sh	17	# Usage:	COMMENT
LOW	backends/aoti/slim/factory/from_etensor.h	34	/// Example usage:	COMMENT
LOW	devtools/scripts/profile_model.sh	15	# Usage:	COMMENT
LOW	devtools/scripts/benchmark_android.sh	13	# Usage:	COMMENT
LOW	examples/models/voxtral_tts/run_cuda_e2e.sh	7	# Usage:	COMMENT
LOW	examples/arduino/build_arduino_library.sh	11	# Usage:	COMMENT
LOW	examples/espressif/build.sh	15	# Usage:	COMMENT
LOW	scripts/test_ios.sh	8	# Usage:	COMMENT
LOW	.github/scripts/trigger_cuda_perf.sh	9	# Usage:	COMMENT

Slop Phrases8 hits · 10 pts

Severity	File	Line	Snippet	Context
LOW	extension/llm/export/builder.py	127	# make sure to re-export this graph module to persist any changes. See	COMMENT
LOW	runtime/core/exec_aten/util/scalar_type_util.h	510	// Don't forget to extend this when adding new QInt types	COMMENT
LOW	runtime/core/exec_aten/util/scalar_type_util.h	524	// Don't forget to extend this when adding new QInt types	COMMENT
LOW	runtime/core/exec_aten/util/scalar_type_util.h	538	// Don't forget to extend this when adding new QInt types	COMMENT
MEDIUM	devtools/inspector/tests/inspector_test.py	1728	"""A simple test model for demonstration purposes."""	STRING
MEDIUM	…es/qualcomm/oss_scripts/llama/assets/samples/text.json	10	"content": "Example: \"Guessing Game\"\n\nObjective: To create a simple program that asks the user to gu	CODE
LOW	…utils/llm_utils/tokenizers_/tokenization_utils_fast.py	715	# make sure to be foward compatible	STRING
LOW	exir/backend/backend_api.py	431	# we make sure to remove them after we finished partition_and_lower	COMMENT

TODO Padding8 hits · 9 pts

Severity	File	Line	Snippet	Context
LOW	extension/data_loader/mman_windows.cpp	97	// TODO: implement	COMMENT
LOW	backends/xnnpack/quantizer/xnnpack_quantizer.py	398	# TODO: implement the support for None to be canceling out previous annotations	COMMENT
LOW	backends/arm/quantizer/arm_quantizer.py	1008	# TODO: implement the support for None to be canceling out previous annotations	COMMENT
LOW	backends/cortex_m/ops/operators.py	1285	# TODO: implement dilation != 1.	COMMENT
LOW⚡	devtools/inspector/_inspector.py	1538	# TODO: implement	STRING
LOW⚡	devtools/inspector/_inspector.py	1545	# TODO: implement	STRING
LOW	examples/models/llava/model.py	217	# TODO: implement _upsample_bicubic_aa.out in portable kernel library.	COMMENT
LOW	exir/tests/test_quantization.py	87	# TODO: implement torch.ops.quantized_decomposed.add_relu.out	COMMENT

Dead Code2 hits · 4 pts

Severity	File	Line	Snippet	Context
MEDIUM	backends/cadence/aot/fuse_ops.py	429		CODE
MEDIUM	backends/cadence/aot/fuse_ops.py	431		CODE

Overly Generic Function Names1 hit · 1 pts

Severity	File	Line	Snippet	Context
LOW	examples/models/llama/static_attention.py	1583	def helper(m):	CODE

Analysis Overview

What These Metrics Mean

Score History

Severity Breakdown

Directory Score Breakdown

Pattern Findings