apache/tvm

17.2

Adjusted Score

17.2

Raw Score

100%

Time Factor

2026-07-13

Last Push

13.6K

Stars

Python

Language

702.9K

Lines of Code

2.2K

Files

11.1K

Pattern Hits

2026-07-14

Scan Date

0.05

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 35HIGH 74MEDIUM 621LOW 10410

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 11140 distinct pattern matches across 22 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 Identifiers4151 hits · 3832 pts

Severity	File	Line	Snippet	Context
LOW	python/tvm/topi/scatter.py	28	def _verify_scatter_nd_inputs(data, indices, updates):	CODE
LOW	python/tvm/topi/math.py	689	def fixed_point_multiply_per_axis(	CODE
LOW	python/tvm/topi/gpu/scan.py	219	def get_reduction_from_exclusive_scan(data, ex_scan_output, binop=operator.add):	CODE
LOW	python/tvm/topi/nn/winograd_util.py	164	def winograd_transform_matrices(tile_size, kernel_size, out_dtype):	CODE
LOW	python/tvm/topi/vision/nms_util.py	245	def collect_selected_indices_and_scores(	CODE
LOW	python/tvm/topi/vision/nms.py	852	def _collect_selected_indices_ir(	CODE
LOW	python/tvm/topi/vision/nms.py	898	def _collect_selected_indices_and_scores_ir(	CODE
LOW	python/tvm/topi/vision/nms.py	944	def all_class_non_max_suppression(	CODE
LOW	python/tvm/topi/image/resize.py	28	def can_convert_multiply_to_intdiv(origin_size, scaled_size):	CODE
LOW	python/tvm/topi/testing/adaptive_pool_python.py	77	def adaptive_pool_channel_first(np_data, out_size, pool_op, np_op):	CODE
LOW	python/tvm/topi/testing/adaptive_pool_python.py	91	def adaptive_pool_channel_last(np_data, out_size, pool_op, np_op):	CODE
LOW	python/tvm/topi/testing/nms_python.py	37	def non_max_suppression_python(	CODE
LOW	python/tvm/ir/instrument.py	154	def _wrap_class_pass_instrument(pi_cls):	CODE
LOW	python/tvm/target/detect_target.py	52	def register_device_target_detector(device_type: str, detector: Callable[[Device], Target]) -> None:	CODE
LOW	python/tvm/target/detect_target.py	57	def detect_target_from_device(dev: str \| Device) -> Target:	CODE
LOW	python/tvm/runtime/params.py	98	def load_param_dict_from_file(path):	CODE
LOW	python/tvm/runtime/device.py	87	def max_shared_memory_per_block(self):	CODE
LOW	python/tvm/runtime/module.py	112	def _collect_from_import_tree(self, filter_func):	CODE
LOW	python/tvm/runtime/module.py	235	def get_source_format_from_module(module):	CODE
LOW	python/tvm/runtime/disco/session.py	600	def _create_socket_session_local_workers(num_workers) -> Session:	CODE
LOW	python/tvm/s_tir/schedule/analysis.py	69	def get_tensorize_loop_mapping(	CODE
LOW	python/tvm/s_tir/schedule/analysis.py	97	def get_auto_tensorize_mapping_info(	CODE
LOW	python/tvm/s_tir/schedule/schedule.py	79	def _parse_error_render_level(error_render_level: str) -> int:	CODE
LOW	python/tvm/s_tir/schedule/schedule.py	104	def _get_sblock_default_dtype(block: SBlock) -> str:	CODE
LOW	python/tvm/s_tir/schedule/schedule.py	1050	def matmul_after_reorder_block_iter_var(	STRING
LOW	python/tvm/s_tir/schedule/schedule.py	1588	def before_reindex_cache_read(a: T.handle, b: T.handle) -> None:	STRING
LOW	python/tvm/s_tir/schedule/schedule.py	1688	def before_reindex_cache_write(a: T.handle, b: T.handle) -> None:	CODE
LOW	python/tvm/s_tir/schedule/schedule.py	2174	def before_reverse_compute_at(a: T.handle, c: T.handle) -> None:	CODE
LOW	python/tvm/s_tir/schedule/schedule.py	3428	def two_elementwise_transformed_intermediate_buffer(a: T.handle, c: T.handle) -> None:	CODE
LOW	python/tvm/s_tir/schedule/schedule.py	3517	def before_transform_block_layout(	CODE
LOW	python/tvm/s_tir/schedule/schedule.py	3539	def after_transform_block_layout(	CODE
LOW	python/tvm/s_tir/schedule/schedule.py	3603	def before_set_axis_separator(	CODE
LOW	python/tvm/s_tir/schedule/schedule.py	3631	def after_set_axis_separators(	CODE
LOW	python/tvm/s_tir/schedule/schedule.py	3944	def unsafe_hide_buffer_access(	CODE
LOW	python/tvm/s_tir/schedule/schedule.py	4004	def before_annotate_buffer_access(	STRING
LOW	python/tvm/s_tir/schedule/schedule.py	4033	def after_annotate_buffer_access(	STRING
LOW	python/tvm/s_tir/schedule/testing.py	29	def assert_structural_equal_ignore_global_symbol(	CODE
LOW	python/tvm/s_tir/analysis/__init__.py	58	def get_sblock_read_write_region(	CODE
LOW	python/tvm/s_tir/analysis/__init__.py	145	def calculate_allocated_bytes(	CODE
LOW	python/tvm/s_tir/analysis/__init__.py	197	def get_vtcm_compaction_passes() -> list[tvm.transform.Pass]:	CODE
LOW	python/tvm/s_tir/meta_schedule/logging.py	179	def _batch_parameterize_config(	CODE
LOW	python/tvm/s_tir/meta_schedule/logging.py	239	def get_loggers_from_work_dir(	CODE
LOW	python/tvm/s_tir/meta_schedule/trace_apply.py	25	def schedule_using_anchor_trace(sch: Schedule, anchor_trace: Trace, target: Target) -> None:	CODE
LOW	python/tvm/s_tir/meta_schedule/relax_integration.py	105	def extracted_tasks_to_tune_contexts(	CODE
LOW	python/tvm/s_tir/meta_schedule/utils.py	134	def get_global_func_with_default_on_worker(	CODE
LOW	python/tvm/s_tir/meta_schedule/utils.py	171	def get_global_func_on_rpc_session(	CODE
LOW	python/tvm/s_tir/meta_schedule/tune_context.py	248	def generate_measure_candidates(self) -> list["MeasureCandidate"] \| None:	CODE
LOW	…/s_tir/meta_schedule/schedule/cuda/layout_transform.py	135	def pad_dimension_to_at_least_number(loop: LoopRV, requested_size: int):	CODE
LOW	…/s_tir/meta_schedule/schedule/cuda/layout_transform.py	140	def pad_dimension_to_factor_of_tile_size(	CODE
LOW	…/s_tir/meta_schedule/schedule/cuda/layout_transform.py	239	def get_high_level_loop_structure(	CODE
LOW	…/s_tir/meta_schedule/schedule/cuda/layout_transform.py	503	def cuda_layout_transform_schedule_rule(	CODE
LOW	…s_tir/meta_schedule/space_generator/space_generator.py	65	def _initialize_with_tune_context(self, context: "TuneContext") -> None:	CODE
LOW	…s_tir/meta_schedule/space_generator/space_generator.py	231	def _initialize_with_tune_context(self, context: "TuneContext") -> None:	CODE
LOW	python/tvm/s_tir/meta_schedule/cost_model/xgb_model.py	684	def optional_xgboost_callback(cls):	CODE
LOW	python/tvm/s_tir/meta_schedule/postproc/postproc.py	43	def _initialize_with_tune_context(self, context: "TuneContext") -> None:	CODE
LOW	python/tvm/s_tir/meta_schedule/postproc/postproc.py	149	def _initialize_with_tune_context(self, context: "TuneContext") -> None:	CODE
LOW	python/tvm/s_tir/meta_schedule/testing/dummy_object.py	56	def _initialize_with_tune_context(self, context: "TuneContext") -> None:	CODE
LOW	…n/tvm/s_tir/meta_schedule/testing/validate_database.py	422	def f_with_args_alloc_argument_common(	CODE
LOW	…n/tvm/s_tir/meta_schedule/testing/validate_database.py	449	def f_with_args_run_evaluator_common(	CODE
LOW	…n/tvm/s_tir/meta_schedule/testing/validate_database.py	516	def f_with_args_alloc_argument_rpc(	CODE
4091 more matches not shown…

Over-Commented Block1946 hits · 1890 pts

Severity	File	Line	Snippet	Context
LOW	.asf.yaml	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	.asf.yaml	41		COMMENT
LOW	.markdownlint-cli2.yaml	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	CMakeLists.txt	601	set_property(TARGET tvm_compiler APPEND PROPERTY LINK_OPTIONS "${TVM_NO_UNDEFINED_SYMBOLS}")	COMMENT
LOW	CMakeLists.txt	861	# libtvm_runtime_extra assembly block above.	COMMENT
LOW	.pre-commit-config.yaml	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	pyproject.toml	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	.yamllint.yaml	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	docker/dev_common.sh	1	#!/bin/bash -e	COMMENT
LOW	docker/build.sh	1	#!/usr/bin/env bash	COMMENT
LOW	docker/build.sh	21	# Build a docker container and optionally execute command within a it	COMMENT
LOW	docker/build.sh	41	# DOCKERFILE_PATH: (Optional) Path to the Dockerfile used for docker build. If	COMMENT
LOW	docker/bash.sh	1	#!/usr/bin/env bash	COMMENT
LOW	docker/bash.sh	21	# Start a bash, mount REPO_MOUNT_POINT to be current directory.	COMMENT
LOW	docker/clear-stale-images.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_core.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_libtorch.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_dgl.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_setup_tz.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_nodejs.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_tflite.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_googletest.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_vulkan.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_onnx.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_dnnl.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_aprofile_aem.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_xgboost.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_java.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_jax.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_sphinx.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_rocm.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_nnef.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_universal.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_cmake.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_tensorflow_aarch64.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_coreml.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_emscripten.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_opencl.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_python_package.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_python.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_sccache.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_tensorflow.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_llvm.sh	1	#!/bin/bash	COMMENT
LOW	docker/install/ubuntu_install_cuda_python.sh	1	#!/bin/bash	COMMENT
LOW	docker/python/bootstrap/generate.sh	1	#!/bin/env bash	COMMENT
LOW	docker/utils/apt-install-and-clear.sh	1	#!/bin/bash -e	COMMENT
LOW	docker/utils/download-and-verify.sh	1	#!/bin/bash	COMMENT
LOW	python/tvm/error.py	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	python/tvm/__init__.py	1	# isort: skip_file	COMMENT
LOW	python/tvm/libinfo.py	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	python/tvm/base.py	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	python/tvm/topi/signal.py	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	python/tvm/topi/slice_scatter.py	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	python/tvm/topi/reduction.py	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	python/tvm/topi/sort.py	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	python/tvm/topi/einsum.py	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	python/tvm/topi/unique.py	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	python/tvm/topi/_te_tensor_overload.py	1	# Licensed to the Apache Software Foundation (ASF) under one	COMMENT
LOW	python/tvm/topi/__init__.py	1	# isort: skip_file	COMMENT
LOW	python/tvm/topi/__init__.py	61	# not import testing by default	COMMENT
1886 more matches not shown…

Unused Imports2174 hits · 1736 pts

Severity	File	Line	Context
LOW	python/tvm/__init__.py	26	CODE
LOW	python/tvm/__init__.py	26	CODE
LOW	python/tvm/__init__.py	26	CODE
LOW	python/tvm/__init__.py	29	CODE
LOW	python/tvm/__init__.py	33	CODE
LOW	python/tvm/__init__.py	34	CODE
LOW	python/tvm/__init__.py	34	CODE
LOW	python/tvm/__init__.py	34	CODE
LOW	python/tvm/__init__.py	34	CODE
LOW	python/tvm/__init__.py	34	CODE
LOW	python/tvm/__init__.py	34	CODE
LOW	python/tvm/__init__.py	35	CODE
LOW	python/tvm/__init__.py	35	CODE
LOW	python/tvm/__init__.py	35	CODE
LOW	python/tvm/__init__.py	35	CODE
LOW	python/tvm/__init__.py	36	CODE
LOW	python/tvm/__init__.py	36	CODE
LOW	python/tvm/__init__.py	42	CODE
LOW	python/tvm/__init__.py	43	CODE
LOW	python/tvm/__init__.py	44	CODE
LOW	python/tvm/__init__.py	45	CODE
LOW	python/tvm/__init__.py	49	CODE
LOW	python/tvm/__init__.py	52	CODE
LOW	python/tvm/__init__.py	55	CODE
LOW	python/tvm/__init__.py	58	CODE
LOW	python/tvm/__init__.py	61	CODE
LOW	python/tvm/__init__.py	64	CODE
LOW	python/tvm/__init__.py	64	CODE
LOW	python/tvm/__init__.py	67	CODE
LOW	python/tvm/__init__.py	70	CODE
LOW	python/tvm/__init__.py	73	CODE
LOW	python/tvm/__init__.py	73	CODE
LOW	python/tvm/__init__.py	79	CODE
LOW	python/tvm/libinfo.py	19	CODE
LOW	python/tvm/topi/__init__.py	29	CODE
LOW	python/tvm/topi/__init__.py	33	CODE
LOW	python/tvm/topi/__init__.py	35	CODE
LOW	python/tvm/topi/__init__.py	36	CODE
LOW	python/tvm/topi/__init__.py	37	CODE
LOW	python/tvm/topi/__init__.py	38	CODE
LOW	python/tvm/topi/__init__.py	39	CODE
LOW	python/tvm/topi/__init__.py	40	CODE
LOW	python/tvm/topi/__init__.py	41	CODE
LOW	python/tvm/topi/__init__.py	42	CODE
LOW	python/tvm/topi/__init__.py	43	CODE
LOW	python/tvm/topi/__init__.py	44	CODE
LOW	python/tvm/topi/__init__.py	45	CODE
LOW	python/tvm/topi/__init__.py	46	CODE
LOW	python/tvm/topi/__init__.py	47	CODE
LOW	python/tvm/topi/__init__.py	48	CODE
LOW	python/tvm/topi/__init__.py	49	CODE
LOW	python/tvm/topi/__init__.py	50	CODE
LOW	python/tvm/topi/__init__.py	51	CODE
LOW	python/tvm/topi/__init__.py	52	CODE
LOW	python/tvm/topi/__init__.py	53	CODE
LOW	python/tvm/topi/__init__.py	54	CODE
LOW	python/tvm/topi/__init__.py	55	CODE
LOW	python/tvm/topi/__init__.py	56	CODE
LOW	python/tvm/topi/__init__.py	59	CODE
LOW	python/tvm/topi/gpu/__init__.py	22	CODE
2114 more matches not shown…

Decorative Section Separators537 hits · 1666 pts

Severity	File	Line	Snippet	Context
MEDIUM	python/tvm/base.py	27	# ----------------------------	COMMENT
MEDIUM	python/tvm/base.py	29	# ----------------------------	COMMENT
MEDIUM	…d/trn/operator/tile_primitive/unary/with_bias_scale.py	78	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…d/trn/operator/tile_primitive/unary/with_bias_scale.py	80	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…m/backend/trn/operator/tile_primitive/unary/default.py	80	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…m/backend/trn/operator/tile_primitive/unary/default.py	82	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…/backend/trn/operator/tile_primitive/binary/default.py	109	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	…/backend/trn/operator/tile_primitive/binary/default.py	111	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	122	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	124	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	133	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	135	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	171	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	173	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	182	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	184	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	187	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	189	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	192	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	194	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	203	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	205	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	212	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	214	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	360	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	362	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	368	# =====================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/target_tags.py	370	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	56	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	58	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	83	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	85	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	147	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	149	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	225	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	227	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	252	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	254	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	266	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	268	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	283	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	285	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	305	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	307	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	338	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/target_tags.py	340	# =====================================================================	COMMENT
MEDIUM	python/tvm/backend/cuda/lang/alloc_pool.py	27	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	python/tvm/backend/cuda/lang/alloc_pool.py	29	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	python/tvm/backend/cuda/lang/alloc_pool.py	49	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	python/tvm/backend/cuda/lang/alloc_pool.py	51	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	python/tvm/backend/cuda/lang/alloc_pool.py	176	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	python/tvm/backend/cuda/lang/alloc_pool.py	178	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	python/tvm/backend/cuda/lang/alloc_pool.py	228	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	python/tvm/backend/cuda/lang/alloc_pool.py	230	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	python/tvm/backend/cuda/lang/alloc_pool.py	412	# ---------------------------------------------------------------------------	COMMENT
MEDIUM	python/tvm/backend/cuda/lang/alloc_pool.py	414	# ---------------------------------------------------------------------------	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/operator/intrinsics/sync.py	112	# =============================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/operator/intrinsics/sync.py	114	# =============================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/operator/intrinsics/sync.py	121	# =============================================================================	COMMENT
MEDIUM⚡	python/tvm/backend/cuda/operator/intrinsics/sync.py	123	# =============================================================================	COMMENT
477 more matches not shown…

Deep Nesting1492 hits · 1275 pts

Severity	File	Line	Context
LOW	python/tvm/topi/signal.py	27	CODE
LOW	python/tvm/topi/signal.py	131	CODE
LOW	python/tvm/topi/signal.py	71	CODE
LOW	python/tvm/topi/signal.py	160	CODE
LOW	python/tvm/topi/unique.py	25	CODE
LOW	python/tvm/topi/sparse_reshape.py	26	CODE
LOW	python/tvm/topi/sparse_reshape.py	74	CODE
LOW	python/tvm/topi/utils.py	176	CODE
LOW	python/tvm/topi/scatter.py	57	CODE
LOW	python/tvm/topi/scatter.py	98	CODE
LOW	python/tvm/topi/scan.py	33	CODE
LOW	python/tvm/topi/scan.py	118	CODE
LOW	python/tvm/topi/scatter_elements.py	27	CODE
LOW	python/tvm/topi/scatter_elements.py	99	CODE
LOW	python/tvm/topi/index_put.py	27	CODE
LOW	python/tvm/topi/index_put.py	91	CODE
LOW	python/tvm/topi/gpu/sort.py	39	CODE
LOW	python/tvm/topi/gpu/sort.py	85	CODE
LOW	python/tvm/topi/gpu/sort.py	185	CODE
LOW	python/tvm/topi/gpu/sort.py	1084	CODE
LOW	python/tvm/topi/gpu/sort.py	259	CODE
LOW	python/tvm/topi/gpu/sort.py	1122	CODE
LOW	python/tvm/topi/gpu/scan.py	57	CODE
LOW	python/tvm/topi/gpu/scan.py	219	CODE
LOW	python/tvm/topi/gpu/scan.py	246	CODE
LOW	python/tvm/topi/gpu/scatter_elements.py	30	CODE
LOW	python/tvm/topi/gpu/scatter_elements.py	62	CODE
LOW	python/tvm/topi/gpu/scatter_nd.py	31	CODE
LOW	python/tvm/topi/gpu/scatter_nd.py	35	CODE
LOW	python/tvm/topi/nn/fifo_buffer.py	29	CODE
LOW	python/tvm/topi/nn/bitserial_util.py	29	CODE
LOW	python/tvm/topi/nn/utils.py	118	CODE
LOW	python/tvm/topi/nn/utils.py	167	CODE
LOW	python/tvm/topi/nn/utils.py	215	CODE
LOW	python/tvm/topi/nn/utils.py	274	CODE
LOW	python/tvm/topi/vision/nms_util.py	87	CODE
LOW	python/tvm/topi/vision/nms_util.py	126	CODE
LOW	python/tvm/topi/vision/nms_util.py	288	CODE
LOW	python/tvm/topi/vision/nms.py	39	CODE
LOW	python/tvm/topi/vision/nms.py	182	CODE
LOW	python/tvm/topi/vision/nms.py	664	CODE
LOW	python/tvm/topi/vision/nms.py	706	CODE
LOW	python/tvm/topi/vision/nms.py	778	CODE
LOW	python/tvm/topi/vision/nms.py	852	CODE
LOW	python/tvm/topi/vision/nms.py	898	CODE
LOW	python/tvm/topi/vision/nms.py	944	CODE
LOW	python/tvm/topi/vision/nms.py	670	CODE
LOW	python/tvm/topi/vision/nms.py	724	CODE
LOW	python/tvm/topi/vision/nms.py	781	CODE
LOW	python/tvm/topi/vision/nms.py	824	CODE
LOW	python/tvm/topi/image/resize.py	140	CODE
LOW	python/tvm/topi/image/resize.py	178	CODE
LOW	python/tvm/topi/image/resize.py	224	CODE
LOW	python/tvm/topi/image/resize.py	373	CODE
LOW	python/tvm/topi/image/resize.py	495	CODE
LOW	python/tvm/topi/image/resize.py	746	CODE
LOW	python/tvm/topi/image/resize.py	959	CODE
LOW	python/tvm/topi/testing/adaptive_pool_python.py	91	CODE
LOW	python/tvm/topi/testing/nms_python.py	37	CODE
LOW	python/tvm/topi/testing/pool_grad_python.py	23	CODE
1432 more matches not shown…

Hallucination Indicators35 hits · 385 pts

Severity	File	Line	Snippet	Context
CRITICAL	…backend/cuda/operator/tile_primitive/copy_async/tma.py	1200	T.ptx.cp_async.bulk.tensor.g2c(	CODE
CRITICAL	…backend/cuda/operator/tile_primitive/copy_async/tma.py	1212	T.ptx.cp_async.bulk.tensor.s2g(	CODE
CRITICAL	…backend/cuda/operator/tile_primitive/copy_async/tma.py	1220	T.ptx.cp_async.bulk.tensor.s2g_reduce(	CODE
CRITICAL	…end/adreno/test_transform_fold_vdevice_scope_change.py	29	mod = tvm.relax.backend.adreno.transform.FoldVDeviceScopeChange()(input)	CODE
CRITICAL	…backend/adreno/test_transform_annotate_custom_scope.py	93	mod = tvm.relax.backend.adreno.transform.AnnotateCustomMemoryScope(tgt)(mod)	CODE
CRITICAL	…backend/adreno/test_transform_annotate_custom_scope.py	105	mod = tvm.relax.backend.adreno.transform.FoldVDeviceScopeChange()(mod)	CODE
CRITICAL	tests/python/tirx/test_printer_tir_namespaces.py	454	'T.ptx.cp_async.bulk.tensor.g2c(2, tm, 0, tm, 0, 1, T.uint64(0), 0, 0, 1, "")',	CODE
CRITICAL⚡	tests/python/tirx/test_printer_tir_namespaces.py	466	'tm = T.handle()\nT.ptx.cp_async.bulk.tensor.g2c_prefetch(2, tm, T.uint64(0), 0, 0, 0, "")',	CODE
CRITICAL⚡	tests/python/tirx/test_printer_tir_namespaces.py	470	'tm = T.handle()\nT.ptx.cp_async.bulk.tensor.s2g(2, 0, tm, T.uint64(0), 0, 0, 0, "")',	CODE
CRITICAL	tests/python/tirx/test_parser_printer.py	1812	T.ptx.cp_async.bulk.tensor.g2c(	CODE
CRITICAL	tests/python/tirx/test_parser_printer.py	1833	T.ptx.cp_async.bulk.tensor.s2g(	CODE
CRITICAL	tests/python/tirx/test_parser_printer.py	1853	T.ptx.cp_async.bulk.tensor.g2c_prefetch(	CODE
CRITICAL	tests/python/tirx/test_parser_printer.py	1874	T.ptx.cp_async.bulk.tensor.s2g_reduce(	CODE
CRITICAL	tests/python/tirx/test_verifier.py	215	T.ptx.cp_async.bulk.tensor.g2c(2, A_smem.data, bar.data, T.address_of(A_map), 0, 1, "", 0, 0) # noq	CODE
CRITICAL⚡	tests/python/tirx/codegen/test_codegen_hopper.py	777	T.ptx.cp_async.bulk.tensor.g2c(len(shape), A_smem.access_ptr(Buffer.WRITE, offset=A_smem	CODE
CRITICAL⚡	tests/python/tirx/codegen/test_codegen_hopper.py	780	T.ptx.cp_async.bulk.tensor.g2c(len(shape), A_smem.access_ptr(Buffer.WRITE, offset=A_smem	CODE
CRITICAL⚡	tests/python/tirx/codegen/test_codegen_hopper.py	783	T.ptx.cp_async.bulk.tensor.g2c(len(shape), A_smem.access_ptr(Buffer.WRITE, offset=A_smem	CODE
CRITICAL⚡	tests/python/tirx/codegen/test_codegen_hopper.py	786	T.ptx.cp_async.bulk.tensor.g2c(len(shape), A_smem.access_ptr(Buffer.WRITE, offset=A_smem	CODE
CRITICAL⚡	tests/python/tirx/codegen/test_codegen_hopper.py	795	T.ptx.cp_async.bulk.tensor.s2g(len(shape), A_smem.access_ptr("r", offset=0), T.address_o	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_hopper.py	432	T.ptx.cp_async.bulk.tensor.g2c(len(shape), A_smem.data, T.address_of(bar), T.address_of(A_map),	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_hopper.py	441	T.ptx.cp_async.bulk.tensor.s2g(len(shape), A_smem.access_ptr("r", offset=0), T.address_of(B_map)	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_hopper.py	599	T.ptx.cp_async.bulk.tensor.g2c(len(shape), A_smem.data, T.address_of(bar), T.address_of(A_map),	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_hopper.py	608	T.ptx.cp_async.bulk.tensor.s2g(len(shape), A_smem.access_ptr("r", offset=0), T.address_of(B_map)	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_hopper.py	692	T.ptx.cp_async.bulk.tensor.g2c(len(shape), A_smem.data, T.address_of(bar), T.address_of(	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_hopper.py	701	T.ptx.cp_async.bulk.tensor.s2g(len(shape), A_smem.access_ptr("r", offset=0), T.address_o	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_hopper.py	860	T.ptx.cp_async.bulk.tensor.s2g(len(shape), A_smem.access_ptr("r", offset=0), T.address_of(A_map), "", *c	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_hopper.py	951	T.ptx.cp_async.bulk.tensor.g2c(len(shapeA), A_smem.data, T.address_of(bar), T.address_of(A_map), 0, 1, "	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_hopper.py	952	T.ptx.cp_async.bulk.tensor.g2c(len(shapeB), B_smem.data, T.address_of(bar), T.address_of(B_map), 0, 1, "	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_hopper.py	1111	T.ptx.cp_async.bulk.tensor.g2c(len(shapeB), B_smem.data, T.address_of(bar), T.address_of(B_map), 0, 1, "	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_cuda.py	324	T.ptx.cp_async.bulk.tensor.g2c(	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_cuda.py	336	T.ptx.cp_async.bulk.tensor.g2c(	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_cuda.py	348	T.ptx.cp_async.bulk.tensor.s2g(	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_cuda.py	352	T.ptx.cp_async.bulk.tensor.g2c_bar_addr(	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_cuda.py	365	T.ptx.cp_async.bulk.tensor.g2c_bar_addr(	CODE
CRITICAL	tests/python/tirx/codegen/test_codegen_cuda.py	378	T.ptx.cp_async.bulk.tensor.g2c_bar_addr(	CODE

Cross-File Repetition58 hits · 290 pts

Severity	File	Snippet	Context
HIGH	python/tvm/topi/math.py	take asinh of input x. parameters ---------- x : expr input argument. returns ------- y : expr the result.	STRING
HIGH	python/tvm/topi/nn/elemwise.py	take asinh of input x. parameters ---------- x : expr input argument. returns ------- y : expr the result.	STRING
HIGH	python/tvm/tirx/op.py	take asinh of input x. parameters ---------- x : expr input argument. returns ------- y : expr the result.	STRING
HIGH	python/tvm/topi/testing/rms_norm_python.py	layer normalization operator in python. parameters ---------- data : numpy.ndarray n-d with shape (d_0, d_1, ..., d_{n-1	STRING
HIGH	python/tvm/topi/testing/instance_norm_python.py	layer normalization operator in python. parameters ---------- data : numpy.ndarray n-d with shape (d_0, d_1, ..., d_{n-1	STRING
HIGH	python/tvm/topi/testing/layer_norm_python.py	layer normalization operator in python. parameters ---------- data : numpy.ndarray n-d with shape (d_0, d_1, ..., d_{n-1	STRING
HIGH	python/tvm/s_tir/meta_schedule/database/database.py	an abstract schedule rule with customized methods on the python-side. this is the user facing class for function overloa	STRING
HIGH	…m/s_tir/meta_schedule/task_scheduler/task_scheduler.py	an abstract schedule rule with customized methods on the python-side. this is the user facing class for function overloa	STRING
HIGH	…s_tir/meta_schedule/space_generator/space_generator.py	an abstract schedule rule with customized methods on the python-side. this is the user facing class for function overloa	STRING
HIGH	…tir/meta_schedule/measure_callback/measure_callback.py	an abstract schedule rule with customized methods on the python-side. this is the user facing class for function overloa	STRING
HIGH	python/tvm/s_tir/meta_schedule/cost_model/cost_model.py	an abstract schedule rule with customized methods on the python-side. this is the user facing class for function overloa	STRING
HIGH	python/tvm/s_tir/meta_schedule/postproc/postproc.py	an abstract schedule rule with customized methods on the python-side. this is the user facing class for function overloa	STRING
HIGH	…s_tir/meta_schedule/search_strategy/search_strategy.py	an abstract schedule rule with customized methods on the python-side. this is the user facing class for function overloa	STRING
HIGH	python/tvm/s_tir/meta_schedule/mutator/mutator.py	an abstract schedule rule with customized methods on the python-side. this is the user facing class for function overloa	STRING
HIGH	…r/meta_schedule/feature_extractor/feature_extractor.py	an abstract schedule rule with customized methods on the python-side. this is the user facing class for function overloa	STRING
HIGH	…tvm/s_tir/meta_schedule/schedule_rule/schedule_rule.py	an abstract schedule rule with customized methods on the python-side. this is the user facing class for function overloa	STRING
HIGH	python/tvm/s_tir/meta_schedule/runner/local_runner.py	default function to run the evaluator parameters ---------- session: rpcsession the session to run the evaluator rt_mod:	STRING
HIGH	python/tvm/s_tir/meta_schedule/runner/utils.py	default function to run the evaluator parameters ---------- session: rpcsession the session to run the evaluator rt_mod:	STRING
HIGH	python/tvm/s_tir/meta_schedule/runner/rpc_runner.py	default function to run the evaluator parameters ---------- session: rpcsession the session to run the evaluator rt_mod:	STRING
HIGH	python/tvm/s_tir/meta_schedule/cost_model/xgb_model.py	update the cost model given running results. parameters ---------- context : tunecontext, the tuning context. candidates	STRING
HIGH	…hon/tvm/s_tir/meta_schedule/cost_model/random_model.py	update the cost model given running results. parameters ---------- context : tunecontext, the tuning context. candidates	STRING
HIGH	python/tvm/s_tir/meta_schedule/cost_model/cost_model.py	update the cost model given running results. parameters ---------- context : tunecontext, the tuning context. candidates	STRING
HIGH	…hon/tvm/s_tir/meta_schedule/cost_model/random_model.py	save the cost model and data to given file location. parameters ---------- path : str the file path.	STRING
HIGH	python/tvm/s_tir/meta_schedule/cost_model/cost_model.py	save the cost model and data to given file location. parameters ---------- path : str the file path.	STRING
HIGH	python/tvm/s_tir/meta_schedule/cost_model/mlp_model.py	save the cost model and data to given file location. parameters ---------- path : str the file path.	STRING
HIGH	python/tvm/s_tir/dlight/adreno/base.py	check whether the target is available for gpu rule. parameters ---------- target : target the compilation target to chec	STRING
HIGH	python/tvm/s_tir/dlight/gpu/base.py	check whether the target is available for gpu rule. parameters ---------- target : target the compilation target to chec	STRING
HIGH	python/tvm/s_tir/dlight/cpu/base.py	check whether the target is available for gpu rule. parameters ---------- target : target the compilation target to chec	STRING
HIGH	…hon/tvm/backend/cuda/operator/tile_primitive/common.py	return the smallest power of 2 greater than or equal to x.	STRING
HIGH	…ator/tile_primitive/cuda/gemm_async/test_gemm_async.py	return the smallest power of 2 greater than or equal to x.	STRING
HIGH	…x/operator/tile_primitive/cuda/copy_async/test_tmem.py	return the smallest power of 2 greater than or equal to x.	STRING
HIGH	python/tvm/contrib/mkl.py	create an extern op that compute batched matrix mult of a and rhs with cblas this function serves as an example on how t	STRING
HIGH	python/tvm/contrib/cblas.py	create an extern op that compute batched matrix mult of a and rhs with cblas this function serves as an example on how t	STRING
HIGH	python/tvm/contrib/dnnl.py	create an extern op that compute batched matrix mult of a and rhs with cblas this function serves as an example on how t	STRING
HIGH	python/tvm/contrib/hipblas.py	create an extern op that compute batch matrix mult of a and rhs with cublas parameters ---------- lhs : tensor the left	STRING
HIGH	python/tvm/contrib/cublaslt.py	create an extern op that compute batch matrix mult of a and rhs with cublas parameters ---------- lhs : tensor the left	STRING
HIGH	python/tvm/contrib/cublas.py	create an extern op that compute batch matrix mult of a and rhs with cublas parameters ---------- lhs : tensor the left	STRING
HIGH	python/tvm/relax/frontend/__init__.py	stablehlo frontends for constructing relax programs, with the model importers	STRING
HIGH	python/tvm/relax/frontend/torch/__init__.py	stablehlo frontends for constructing relax programs, with the model importers	STRING
HIGH	python/tvm/relax/frontend/stablehlo/__init__.py	stablehlo frontends for constructing relax programs, with the model importers	STRING
HIGH	python/tvm/relax/frontend/nn/llm/_kernel_common.py	the rope mode of the paged kv cache. if it is none, the kv cache will not apply rope to q and k. if it is normal, rope w	STRING
HIGH	…x/test_runtime_builtin_paged_attention_kv_cache_cpu.py	the rope mode of the paged kv cache. if it is none, the kv cache will not apply rope to q and k. if it is normal, rope w	STRING
HIGH	docs/how_to/tutorials/optimize_llm.py	the rope mode of the paged kv cache. if it is none, the kv cache will not apply rope to q and k. if it is normal, rope w	STRING
HIGH	python/tvm/relax/backend/metal/coreml.py	partition the input module into hipblas-supported subgraphs. parameters ---------- mod: tvm.irmodule the irmodule to be	STRING
HIGH	python/tvm/relax/backend/cuda/cudnn.py	partition the input module into hipblas-supported subgraphs. parameters ---------- mod: tvm.irmodule the irmodule to be	STRING
HIGH	python/tvm/relax/backend/rocm/hipblas.py	partition the input module into hipblas-supported subgraphs. parameters ---------- mod: tvm.irmodule the irmodule to be	STRING
HIGH	tests/python/codegen/test_target_codegen_opencl.py	force the codegen wrapper into the fallback branch, then export+load+run.	STRING
HIGH	tests/python/codegen/test_target_codegen_cuda.py	force the codegen wrapper into the fallback branch, then export+load+run.	STRING
HIGH	tests/python/codegen/test_target_codegen_rocm.py	force the codegen wrapper into the fallback branch, then export+load+run.	STRING
HIGH	tests/python/relax/test_transform_convert_layout.py	layout_transform (nchw->nchw4c) \| <- texture conv2d (1) <- textures as output / \ conv2d (2) conv2d (3) \ / <- concat do	STRING
HIGH	…backend/adreno/test_transform_annotate_custom_scope.py	layout_transform (nchw->nchw4c) \| <- texture conv2d (1) <- textures as output / \ conv2d (2) conv2d (3) \ / <- concat do	STRING
HIGH	tests/python/relax/backend/adreno/test_texture_ops.py	layout_transform (nchw->nchw4c) \| <- texture conv2d (1) <- textures as output / \ conv2d (2) conv2d (3) \ / <- concat do	STRING
HIGH	tests/python/relax/test_transform_convert_layout.py	layout_transform (nchw->nchw4c) \| <- texture conv2d (1) <- textures as output / \ conv2d (2) conv2d (3) <- conv2d (2) em	STRING
HIGH	…backend/adreno/test_transform_annotate_custom_scope.py	layout_transform (nchw->nchw4c) \| <- texture conv2d (1) <- textures as output / \ conv2d (2) conv2d (3) <- conv2d (2) em	STRING
HIGH	tests/python/relax/backend/adreno/test_texture_ops.py	layout_transform (nchw->nchw4c) \| <- texture conv2d (1) <- textures as output / \ conv2d (2) conv2d (3) <- conv2d (2) em	STRING
HIGH	tests/python/relax/test_transform_convert_layout.py	verification of the pooling and many branches having textures layout_transform (nchw->nchw4c) \| <- texture conv2d (0) <-	STRING
HIGH	…backend/adreno/test_transform_annotate_custom_scope.py	verification of the pooling and many branches having textures layout_transform (nchw->nchw4c) \| <- texture conv2d (0) <-	STRING
HIGH	tests/python/relax/backend/adreno/test_texture_ops.py	verification of the pooling and many branches having textures layout_transform (nchw->nchw4c) \| <- texture conv2d (0) <-	STRING

AI Structural Patterns268 hits · 217 pts

Severity	File	Line	Context
LOW	python/tvm/topi/sort.py	67	CODE
LOW	python/tvm/topi/sort.py	219	CODE
LOW	python/tvm/topi/nn/conv2d.py	1278	CODE
LOW	python/tvm/topi/nn/conv2d.py	1409	CODE
LOW	python/tvm/topi/nn/conv1d_transpose.py	143	CODE
LOW	python/tvm/topi/nn/lstm.py	25	CODE
LOW	python/tvm/topi/vision/nms.py	476	CODE
LOW	python/tvm/topi/image/resize.py	224	CODE
LOW	python/tvm/topi/image/resize.py	373	CODE
LOW	python/tvm/topi/image/resize.py	495	CODE
LOW	python/tvm/topi/image/resize.py	746	CODE
LOW	python/tvm/topi/image/resize.py	959	CODE
LOW	python/tvm/topi/image/resize.py	1215	CODE
LOW	python/tvm/topi/testing/nms_python.py	37	CODE
LOW	python/tvm/topi/testing/lstm_python.py	24	CODE
LOW	python/tvm/topi/testing/sequence_mask_python.py	55	CODE
LOW	python/tvm/runtime/script_printer.py	53	CODE
LOW	python/tvm/runtime/script_printer.py	121	CODE
LOW	python/tvm/runtime/script_printer.py	253	CODE
LOW	python/tvm/runtime/script_printer.py	297	CODE
LOW	python/tvm/runtime/module.py	325	CODE
LOW	python/tvm/te/operation.py	435	CODE
LOW	python/tvm/s_tir/meta_schedule/relax_integration.py	163	CODE
LOW	python/tvm/s_tir/meta_schedule/relax_integration.py	277	CODE
LOW	python/tvm/s_tir/meta_schedule/tune.py	29	CODE
LOW	python/tvm/s_tir/meta_schedule/tune_context.py	98	CODE
LOW	python/tvm/s_tir/meta_schedule/tir_integration.py	45	CODE
LOW	python/tvm/s_tir/meta_schedule/tir_integration.py	167	CODE
LOW	python/tvm/s_tir/meta_schedule/database/database.py	435	CODE
LOW	python/tvm/s_tir/meta_schedule/runner/local_runner.py	236	CODE
LOW	python/tvm/s_tir/meta_schedule/runner/rpc_runner.py	236	CODE
LOW	python/tvm/s_tir/meta_schedule/cost_model/xgb_model.py	334	CODE
LOW	python/tvm/s_tir/meta_schedule/cost_model/xgb_model.py	634	CODE
LOW	python/tvm/s_tir/meta_schedule/cost_model/mlp_model.py	415	CODE
LOW	…n/tvm/s_tir/meta_schedule/testing/validate_database.py	328	CODE
LOW	…r/meta_schedule/search_strategy/evolutionary_search.py	59	CODE
LOW	…r/meta_schedule/feature_extractor/feature_extractor.py	64	CODE
LOW	python/tvm/s_tir/dlight/benchmark/bench.py	155	CODE
LOW	python/tvm/backend/cuda/op.py	3509	CODE
LOW	python/tvm/backend/cuda/op.py	3596	CODE
LOW	python/tvm/backend/cuda/op.py	3659	CODE
LOW	python/tvm/backend/cuda/op.py	3881	CODE
LOW	python/tvm/backend/cuda/op.py	3919	CODE
LOW	python/tvm/backend/cuda/op.py	3955	CODE
LOW	python/tvm/backend/cuda/operator/intrinsics/nvshmem.py	97	CODE
LOW	python/tvm/backend/cuda/operator/intrinsics/nvshmem.py	124	CODE
LOW	python/tvm/backend/cuda/operator/intrinsics/nvshmem.py	154	CODE
LOW	python/tvm/script/parser/core/evaluator.py	356	CODE
LOW	python/tvm/relax/base_py_module.py	496	CODE
LOW	python/tvm/relax/expr_functor.py	275	CODE
LOW	python/tvm/relax/expr_functor.py	793	CODE
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	1423	CODE
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	1539	CODE
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	1595	CODE
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	3660	CODE
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	3696	CODE
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	3809	CODE
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	3832	CODE
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	3845	CODE
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	3855	CODE
208 more matches not shown…

Excessive Try-Catch Wrapping135 hits · 152 pts

Severity	File	Line	Snippet	Context
LOW	python/tvm/ir/instrument.py	338	except Exception: # pylint: disable=broad-exception-caught	CODE
LOW	python/tvm/ir/base.py	37	except Exception:	CODE
LOW	python/tvm/runtime/_tensor.py	114	except Exception:	CODE
LOW	python/tvm/s_tir/schedule/_type_checker.py	350	except Exception:	CODE
LOW	…hon/tvm/s_tir/meta_schedule/post_optimization/space.py	237	except Exception: # pylint: disable=broad-except, invalid-name	CODE
LOW	…hon/tvm/s_tir/meta_schedule/post_optimization/space.py	251	except Exception: # pylint: disable=broad-except, invalid-name	CODE
LOW	python/tvm/s_tir/meta_schedule/runner/local_runner.py	310	except Exception as exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/s_tir/meta_schedule/runner/rpc_runner.py	127	except Exception as exception: # pylint: disable=broad-except	CODE
MEDIUM	python/tvm/s_tir/meta_schedule/runner/rpc_runner.py	119	def result(self) -> RunnerResult:	CODE
LOW	…n/tvm/s_tir/meta_schedule/testing/validate_database.py	769	except Exception as e: # pylint: disable=broad-except, invalid-name	CODE
LOW	python/tvm/s_tir/meta_schedule/testing/tune_utils.py	107	except Exception as exc: # pylint: disable=broad-except	CODE
LOW	python/tvm/s_tir/dlight/adreno/pool.py	91	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/s_tir/dlight/adreno/fallback.py	82	except Exception: # pylint: disable=broad-exception-caught	CODE
LOW	python/tvm/s_tir/dlight/adreno/fallback.py	94	except Exception: # pylint: disable=broad-exception-caught	CODE
LOW	python/tvm/s_tir/dlight/analysis/common_analysis.py	295	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/s_tir/dlight/analysis/common_analysis.py	398	except Exception:	CODE
LOW	python/tvm/backend/_autoload_backends.py	87	except Exception as e: # pylint: disable=broad-except	CODE
LOW	python/tvm/backend/loader.py	171	except Exception:	CODE
LOW	python/tvm/backend/metal/target_tags.py	30	except Exception: # pylint: disable=broad-except	CODE
MEDIUM	python/tvm/backend/metal/target_tags.py	24	def _register_metal_tag(name, max_threads, shared_mem, warp_size, mcpu):	CODE
LOW	…m/backend/cuda/operator/tile_primitive/layout_utils.py	310	except Exception:	CODE
LOW	…ator/tile_primitive/permute_layout/warp_xor_swizzle.py	242	except Exception as e:	CODE
LOW	…ckend/cuda/operator/tile_primitive/gemm/mma_m16n8k_.py	106	except Exception as e: # follower dim layout incompatible with anchor decomposition	CODE
LOW	…ckend/cuda/operator/tile_primitive/gemm/mma_m16n8k_.py	355	except Exception as e: # ICHECK failure -> layout not tileable as shape2d	CODE
LOW	…ckend/cuda/operator/tile_primitive/gemm/mma_m16n8k_.py	368	except Exception as e: # ICHECK failure -> layout not tileable as shape2d	CODE
LOW	…ckend/cuda/operator/tile_primitive/gemm/mma_m16n8k_.py	460	except Exception:	CODE
MEDIUM	…ckend/cuda/operator/tile_primitive/gemm/mma_m16n8k_.py	335	def _const(expr, name):	CODE
LOW	…nd/cuda/operator/tile_primitive/elementwise/_common.py	169	except Exception as e: # pylint: disable=broad-except	CODE
LOW⚡	…backend/cuda/operator/tile_primitive/copy_async/tma.py	326	except Exception as err:	CODE
LOW	…backend/cuda/operator/tile_primitive/copy_async/tma.py	403	except Exception:	CODE
LOW	…m/backend/cuda/operator/tile_primitive/copy/_common.py	436	except Exception as e:	CODE
LOW	…m/backend/cuda/operator/tile_primitive/copy/_common.py	483	except Exception as e:	CODE
LOW	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	196	except Exception:	CODE
LOW	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	268	except Exception:	CODE
LOW	python/tvm/contrib/download.py	123	except Exception as err:	CODE
LOW	python/tvm/script/parser/core/parser.py	337	except Exception as err: # pylint: disable=broad-except	CODE
LOW	python/tvm/script/parser/core/parser.py	628	except Exception as diag_err:	CODE
LOW	python/tvm/script/parser/core/parser.py	666	except Exception as err: # pylint: disable=broad-except	CODE
MEDIUM	python/tvm/script/parser/core/parser.py	58	def context():	CODE
LOW	python/tvm/script/parser/core/entry.py	143	except Exception as err: # pylint: disable=broad-exception-caught	CODE
LOW⚡	python/tvm/script/parser/core/entry.py	173	except Exception as error:	CODE
MEDIUM⚡	python/tvm/script/parser/core/entry.py	174	print(f"Error calling Python function {pyfunc.__name__}: {error}")	CODE
LOW	python/tvm/script/parser/core/evaluator.py	268	except Exception as err: # pylint: disable=broad-except	CODE
LOW	python/tvm/script/parser/core/evaluator.py	287	except Exception as err: # pylint: disable=broad-except	CODE
LOW	python/tvm/script/parser/core/evaluator.py	496	except Exception as err: # pylint: disable=broad-except	CODE
LOW	python/tvm/script/parser/ir/entry.py	105	except Exception: # pylint: disable=broad-exception-caught	CODE
LOW	python/tvm/testing/env.py	94	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	110	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	130	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	144	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	257	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	272	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	299	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	318	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	334	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	351	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	367	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	380	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/testing/env.py	402	except Exception: # pylint: disable=broad-except	CODE
LOW	python/tvm/support/popen_pool.py	440	except Exception as exception:	CODE
75 more matches not shown…

Self-Referential Comments51 hits · 134 pts

Severity	File	Line	Snippet	Context
MEDIUM	python/tvm/topi/testing/poolnd_python.py	163	# Create a padded array, and a boolean mask showing which values are padded values	COMMENT
MEDIUM	python/tvm/te/operation.py	179	# The following code is equivalent to numpy.cumsum	STRING
MEDIUM	…/backend/trn/operator/tile_primitive/binary/default.py	74	# Define the implementation function	COMMENT
MEDIUM	python/tvm/support/nvcc.py	514	# Define the vector-deprecation silencing macros as no-ops for every NVRTC	STRING
MEDIUM	python/tvm/relax/base_py_module.py	479	# Create a wrapper that handles both instance methods and static functions	COMMENT
MEDIUM	python/tvm/relax/relax_to_pyfunc_converter.py	87	# Create a copy of the current IRModule	COMMENT
MEDIUM	python/tvm/relax/relax_to_pyfunc_converter.py	128	# Create the Python function	COMMENT
MEDIUM	python/tvm/relax/frontend/tflite/tflite_frontend.py	8904	# Define a tf.Module containing a tf.function with network definition	STRING
MEDIUM	python/tvm/relax/frontend/nn/core.py	359	"""This method provides an iterator over module parameters,	STRING
MEDIUM	python/tvm/relax/frontend/nn/core.py	376	"""This method provides an iterator over module parameters,	STRING
MEDIUM	python/tvm/relax/frontend/onnx/onnx_frontend.py	4013	# Create a tensor of zeros then scatter our data through it.	COMMENT
MEDIUM	…n/tvm/relax/frontend/torch/base_fx_graph_translator.py	1486	# Create a placeholder for indices (empty tensor with same shape as output)	COMMENT
MEDIUM	…n/tvm/relax/frontend/torch/base_fx_graph_translator.py	1502	# Create a placeholder for indices (empty tensor with same shape as output)	COMMENT
MEDIUM	…n/tvm/relax/frontend/torch/base_fx_graph_translator.py	1518	# Create a placeholder for indices (empty tensor with same shape as output)	COMMENT
MEDIUM	python/tvm/relax/frontend/torch/fx_translator.py	1231	# Define the module	STRING
MEDIUM	…vm/relax/frontend/torch/exported_program_translator.py	190	# This method is called for batch_norm in training mode	COMMENT
MEDIUM	…vm/relax/frontend/torch/exported_program_translator.py	2346	# Define the module	STRING
MEDIUM	python/tvm/relax/testing/nn.py	228	# Define a linear layer	STRING
MEDIUM	python/tvm/relax/transform/transform.py	1115	# Define the pass input IRModule	STRING
MEDIUM	python/tvm/relax/transform/transform.py	1126	# Define the customized legalization function for "relax.add"	STRING
MEDIUM	python/tvm/tirx/stmt_functor.py	506	# Create the data_or_idx parameter based on what's available	COMMENT
MEDIUM	python/tvm/tirx/expr.py	325	# This class is not manipulated by C++. So use python's identity check function is sufficient	COMMENT
MEDIUM	python/tvm/tirx/expr.py	369	# This class is not manipulated by C++. So use python's identity check function is sufficient	COMMENT
MEDIUM	tests/python/driver/test_compile.py	63	# Define a simple Relax program	COMMENT
MEDIUM	tests/python/tirx-base/test_tir_expr_functor.py	815	# Create a simple expression	COMMENT
MEDIUM	tests/python/tirx-base/test_tir_expr_functor.py	818	# Create a mutator that replaces 'n' with a constant	COMMENT
MEDIUM	tests/python/tirx-base/test_tir_stmt_functor.py	994	# Create a new IntImm with negated value	COMMENT
MEDIUM	tests/python/runtime/test_executable.py	107	# Create a temporary directory for the library	COMMENT
MEDIUM	tests/python/runtime/test_executable.py	181	# Create an executable	COMMENT
MEDIUM	tests/python/runtime/test_executable.py	237	# Create an executable	COMMENT
MEDIUM	tests/python/s_tir/dlight/test_gpu_fallback.py	203	# This function is identical, except that it is explicitly	COMMENT
MEDIUM	tests/python/codegen/test_target_codegen_llvm.py	90	# Create an intrinsic that returns void.	COMMENT
MEDIUM	tests/python/tirx-transform/test_tir_functor.py	266	# Create a simple for loop	COMMENT
MEDIUM	tests/python/tirx-transform/test_tir_functor.py	298	# Create an evaluate statement with an expression	COMMENT
MEDIUM	tests/python/support/test_util.py	77	# Creating a new TempDirectory should fail now	COMMENT
MEDIUM	…/tirx-analysis/test_tir_analysis_verify_well_formed.py	446	# Create an undeclared buffer using an explicit data pointer that is NOT	COMMENT
MEDIUM	…ime_builtin_paged_attention_kv_cache_mla_flashinfer.py	449	# Create a mask similar to np.tril and np.triu.	COMMENT
MEDIUM	tests/python/relax/test_pytorch_integration.py	164	# Define a custom function	COMMENT
MEDIUM	tests/python/relax/test_base_py_module.py	121	# Create a simple IRModule without TIR functions for GPU testing	COMMENT
MEDIUM⚡	tests/python/relax/test_dlpack_integration.py	221	# Create a simple IRModule	COMMENT
MEDIUM	…st_runtime_builtin_paged_attention_kv_cache_mla_tir.py	379	# Create a mask similar to np.tril and np.triu.	COMMENT
MEDIUM	tests/python/tirx/transform/test_tirx_expr_functor.py	815	# Create a simple expression	COMMENT
MEDIUM	tests/python/tirx/transform/test_tirx_expr_functor.py	818	# Create a mutator that replaces 'n' with a constant	COMMENT
MEDIUM	tests/python/tirx/transform/test_stmt_functor.py	1008	# Create a new IntImm with negated value	COMMENT
MEDIUM	tests/scripts/task_config_build_gpu_other.sh	19	# This file is a compiler test to ensure that runtimes can compile	COMMENT
MEDIUM	docs/conf.py	23	# This file is execfile()d with the current directory set to its	COMMENT
MEDIUM	docs/how_to/tutorials/cross_compilation_and_rpc.py	411	# The following function shows how we run an OpenCL kernel remotely	COMMENT
MEDIUM	docs/how_to/tutorials/customize_opt.py	121	# Define a new pass for CUBLAS dispatch	COMMENT
MEDIUM	docs/how_to/tutorials/import_model.py	307	# Define a simple TF module and convert to TFLite.	COMMENT
MEDIUM	docs/how_to/tutorials/optimize_llm.py	530	# Create the KVCache	COMMENT
MEDIUM	docs/get_started/tutorials/ir_module.py	57	# Create a dummy model	COMMENT

Verbosity Indicators67 hits · 123 pts

Severity	File	Line	Snippet	Context
LOW	python/tvm/topi/vision/nms.py	220	# Step 1: Reorder data by sorted score	COMMENT
LOW	python/tvm/topi/vision/nms.py	241	# Step 2: Apply NMS - greedy suppression	COMMENT
LOW	python/tvm/ir/transform.py	295	The following code block decorates a module pass class.	STRING
LOW⚡	python/tvm/s_tir/meta_schedule/runner/local_runner.py	146	# Step 1: create the local runtime module	COMMENT
LOW⚡	python/tvm/s_tir/meta_schedule/runner/local_runner.py	149	# Step 2: Allocate input arguments	COMMENT
LOW⚡	python/tvm/s_tir/meta_schedule/runner/local_runner.py	157	# Step 3: Run time_evaluator	COMMENT
LOW⚡	python/tvm/s_tir/meta_schedule/runner/rpc_runner.py	394	# Step 3: Allocate input arguments	COMMENT
LOW⚡	python/tvm/s_tir/meta_schedule/runner/rpc_runner.py	402	# Step 4: Run time_evaluator	COMMENT
LOW⚡	…ackend/cuda/operator/tile_primitive/elementwise/reg.py	292	# Step 1: slice + permute (region offset absorbed into op_p.offset; per-iter	COMMENT
LOW⚡	…ackend/cuda/operator/tile_primitive/elementwise/reg.py	297	# Step 2: post-align thread-equality check. ``align`` is supposed to	COMMENT
LOW⚡	…ackend/cuda/operator/tile_primitive/elementwise/reg.py	309	# Step 3: drop thread iters; from here on operands have mem-only layouts.	COMMENT
LOW⚡	…ackend/cuda/operator/tile_primitive/elementwise/reg.py	314	# Step 4: enumerate spec.vec_impls widest-first; try carve tail for each	COMMENT
LOW⚡	…ackend/cuda/operator/tile_primitive/elementwise/reg.py	319	# Step 5: totals + emit. per_thread_total = ∏ extents of (carved) mem	COMMENT
LOW	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	82	# Step 1: identify reg / smem sides and pull their tensor shape + layout.	COMMENT
LOW⚡	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	126	# Step 3: ldstmatrix doesn't broadcast — require zero replica on both sides.	COMMENT
LOW⚡	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	132	# Step 4: R must have exactly one kind of lane axis from the valid set.	COMMENT
LOW⚡	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	140	# Step 5: group S by R's iter extents (one S group per R iter, outer→inner).	COMMENT
LOW⚡	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	144	# Step 6: permute R so thread iters come first (stride-desc), then memory	COMMENT
LOW⚡	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	149	# Step 7: apply R's perm to S in group units (1-to-1 with R's iters), and	COMMENT
LOW	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	290	# Step 10: emit one ldmatrix/stmatrix per mm, per warp.	COMMENT
LOW	python/tvm/script/parser/ir/parser.py	54	# Step 1: Check if this class inherits from BasePyModule	COMMENT
LOW⚡	python/tvm/script/parser/ir/parser.py	147	# We need to check if we're in a class that inherits from BasePyModule	COMMENT
LOW	python/tvm/relax/transform/transform.py	1686	The following code block decorates a function pass class.	STRING
LOW	python/tvm/relax/transform/transform.py	1830	The following code block decorates a dataflowblock pass class.	STRING
LOW⚡	python/tvm/tirx/build.py	183	# Step 0: Determine the target in environment	COMMENT
LOW⚡	python/tvm/tirx/build.py	191	# Step 1: Determine the target to search for tirx pipeline	COMMENT
LOW	python/tvm/tirx/build.py	202	# Step 2: Determine the host target	COMMENT
LOW⚡	python/tvm/tirx/build.py	214	# Step 3: Bind the target to the input module	COMMENT
LOW⚡	python/tvm/tirx/build.py	217	# Step 4: Apply the tirx pipeline	COMMENT
LOW⚡	python/tvm/tirx/build.py	229	# Step 5: Get host and device modules	COMMENT
LOW⚡	python/tvm/tirx/build.py	232	# Step 6: Apply finalization passes	COMMENT
LOW	python/tvm/tirx/transform/function_pass.py	110	The following code block decorates a function pass class.	STRING
LOW	include/tvm/topi/nn/pooling.h	577	// In the case of ceil_mode=True, we need to check if the last pooling window is valid.	COMMENT
LOW	include/tvm/ir/config_schema.h	125	// Step 1: validate/coerce and materialize options in declaration order	COMMENT
LOW	include/tvm/ir/config_schema.h	140	// Step 2: unknown-key policy	COMMENT
LOW	include/tvm/ir/config_schema.h	157	// Step 3: whole-object canonicalization (last)	COMMENT
LOW⚡	tests/python/relax/test_pytorch_integration.py	287	# Step 1: TIR matmul	COMMENT
LOW⚡	tests/python/relax/test_pytorch_integration.py	292	# Step 2: ReLU	COMMENT
LOW⚡	tests/python/relax/test_pytorch_integration.py	297	# Step 3: Softmax via call_dps_packed	COMMENT
LOW⚡	tests/python/relax/test_pytorch_integration.py	302	# Step 4: Identity function	COMMENT
LOW	tests/python/relax/test_base_py_module_printer.py	435	# We need to check if they exist in the IRModule's pyfuncs	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	496	# Step 1: Define and Export PyTorch Model	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	522	# Step 2: Convert to Relax and Prepare for Compilation	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	534	# Step 3: Cross-Compile for Target Device	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	588	# Step 4: Deploy to Remote Device via RPC	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	651	# Step 5: Run Inference on Remote Device	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	690	# Step 6: Performance Evaluation (Optional)	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	72	# Step 1: Your First Hybrid Module	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	123	# Step 2: Debugging — The Main Selling Point	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	187	# Step 3: A Realistic Pipeline — Python, TIR, and Packed Functions	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	260	# Step 4: Relax-to-Python Converter — Verify at Any Compilation Stage	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	332	# Step 5: R.call_py_func — Python Callbacks in Compiled IR	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	383	# Step 6: Cross-Level Calls and Symbolic Shapes	COMMENT
LOW⚡	docs/how_to/tutorials/bring_your_own_codegen.py	70	# Step 1: Import the backends to register their patterns	COMMENT
LOW⚡	docs/how_to/tutorials/bring_your_own_codegen.py	103	# Step 2: Define the model	COMMENT
LOW⚡	docs/how_to/tutorials/bring_your_own_codegen.py	125	# Step 3: Partition for the example NPU	COMMENT
LOW⚡	docs/how_to/tutorials/bring_your_own_codegen.py	153	# Step 4: Codegen, build and run on the example NPU	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	247	# # Step 1: Load the compiled library	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	250	# # Step 2: Create Virtual Machine	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	254	# # Step 3: Load parameters from the .npz file	COMMENT
7 more matches not shown…

Structural Annotation Overuse61 hits · 115 pts

Severity	File	Line	Snippet	Context
LOW	python/tvm/topi/vision/nms.py	220	# Step 1: Reorder data by sorted score	COMMENT
LOW	python/tvm/topi/vision/nms.py	241	# Step 2: Apply NMS - greedy suppression	COMMENT
LOW⚡	python/tvm/s_tir/meta_schedule/runner/local_runner.py	146	# Step 1: create the local runtime module	COMMENT
LOW⚡	python/tvm/s_tir/meta_schedule/runner/local_runner.py	149	# Step 2: Allocate input arguments	COMMENT
LOW⚡	python/tvm/s_tir/meta_schedule/runner/local_runner.py	157	# Step 3: Run time_evaluator	COMMENT
LOW⚡	python/tvm/s_tir/meta_schedule/runner/rpc_runner.py	394	# Step 3: Allocate input arguments	COMMENT
LOW⚡	python/tvm/s_tir/meta_schedule/runner/rpc_runner.py	402	# Step 4: Run time_evaluator	COMMENT
LOW⚡	…ackend/cuda/operator/tile_primitive/elementwise/reg.py	292	# Step 1: slice + permute (region offset absorbed into op_p.offset; per-iter	COMMENT
LOW⚡	…ackend/cuda/operator/tile_primitive/elementwise/reg.py	297	# Step 2: post-align thread-equality check. ``align`` is supposed to	COMMENT
LOW⚡	…ackend/cuda/operator/tile_primitive/elementwise/reg.py	309	# Step 3: drop thread iters; from here on operands have mem-only layouts.	COMMENT
LOW⚡	…ackend/cuda/operator/tile_primitive/elementwise/reg.py	314	# Step 4: enumerate spec.vec_impls widest-first; try carve tail for each	COMMENT
LOW⚡	…ackend/cuda/operator/tile_primitive/elementwise/reg.py	319	# Step 5: totals + emit. per_thread_total = ∏ extents of (carved) mem	COMMENT
LOW	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	82	# Step 1: identify reg / smem sides and pull their tensor shape + layout.	COMMENT
LOW⚡	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	126	# Step 3: ldstmatrix doesn't broadcast — require zero replica on both sides.	COMMENT
LOW⚡	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	132	# Step 4: R must have exactly one kind of lane axis from the valid set.	COMMENT
LOW⚡	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	140	# Step 5: group S by R's iter extents (one S group per R iter, outer→inner).	COMMENT
LOW⚡	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	144	# Step 6: permute R so thread iters come first (stride-desc), then memory	COMMENT
LOW⚡	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	149	# Step 7: apply R's perm to S in group units (1-to-1 with R's iters), and	COMMENT
LOW	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	290	# Step 10: emit one ldmatrix/stmatrix per mm, per warp.	COMMENT
LOW	python/tvm/script/parser/ir/parser.py	54	# Step 1: Check if this class inherits from BasePyModule	COMMENT
LOW	python/tvm/relax/backend/cuda/flashinfer.py	325	# NOTE: This function is still under development,	COMMENT
LOW⚡	python/tvm/tirx/build.py	183	# Step 0: Determine the target in environment	COMMENT
LOW⚡	python/tvm/tirx/build.py	191	# Step 1: Determine the target to search for tirx pipeline	COMMENT
LOW	python/tvm/tirx/build.py	202	# Step 2: Determine the host target	COMMENT
LOW⚡	python/tvm/tirx/build.py	214	# Step 3: Bind the target to the input module	COMMENT
LOW⚡	python/tvm/tirx/build.py	217	# Step 4: Apply the tirx pipeline	COMMENT
LOW⚡	python/tvm/tirx/build.py	229	# Step 5: Get host and device modules	COMMENT
LOW⚡	python/tvm/tirx/build.py	232	# Step 6: Apply finalization passes	COMMENT
LOW	include/tvm/ir/config_schema.h	125	// Step 1: validate/coerce and materialize options in declaration order	COMMENT
LOW	include/tvm/ir/config_schema.h	140	// Step 2: unknown-key policy	COMMENT
LOW	include/tvm/ir/config_schema.h	157	// Step 3: whole-object canonicalization (last)	COMMENT
LOW⚡	tests/python/relax/test_pytorch_integration.py	287	# Step 1: TIR matmul	COMMENT
LOW⚡	tests/python/relax/test_pytorch_integration.py	292	# Step 2: ReLU	COMMENT
LOW⚡	tests/python/relax/test_pytorch_integration.py	297	# Step 3: Softmax via call_dps_packed	COMMENT
LOW⚡	tests/python/relax/test_pytorch_integration.py	302	# Step 4: Identity function	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	496	# Step 1: Define and Export PyTorch Model	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	522	# Step 2: Convert to Relax and Prepare for Compilation	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	534	# Step 3: Cross-Compile for Target Device	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	588	# Step 4: Deploy to Remote Device via RPC	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	651	# Step 5: Run Inference on Remote Device	COMMENT
LOW	docs/how_to/tutorials/cross_compilation_and_rpc.py	690	# Step 6: Performance Evaluation (Optional)	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	72	# Step 1: Your First Hybrid Module	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	123	# Step 2: Debugging — The Main Selling Point	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	187	# Step 3: A Realistic Pipeline — Python, TIR, and Packed Functions	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	260	# Step 4: Relax-to-Python Converter — Verify at Any Compilation Stage	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	332	# Step 5: R.call_py_func — Python Callbacks in Compiled IR	COMMENT
LOW⚡	…s/how_to/tutorials/mix_python_and_tvm_with_pymodule.py	383	# Step 6: Cross-Level Calls and Symbolic Shapes	COMMENT
LOW⚡	docs/how_to/tutorials/bring_your_own_codegen.py	70	# Step 1: Import the backends to register their patterns	COMMENT
LOW⚡	docs/how_to/tutorials/bring_your_own_codegen.py	103	# Step 2: Define the model	COMMENT
LOW⚡	docs/how_to/tutorials/bring_your_own_codegen.py	125	# Step 3: Partition for the example NPU	COMMENT
LOW⚡	docs/how_to/tutorials/bring_your_own_codegen.py	153	# Step 4: Codegen, build and run on the example NPU	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	247	# # Step 1: Load the compiled library	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	250	# # Step 2: Create Virtual Machine	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	254	# # Step 3: Load parameters from the .npz file	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	259	# # Step 4: Prepare input data	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	263	# # Step 5: Run inference (pass input followed by all parameters)	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	266	# # Step 6: Extract result (output may be tuple or single Tensor)	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	342	# # Step 1: Cross-compile for ARM target (on local machine)	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	347	# # Step 2: Connect to remote device RPC server	COMMENT
LOW⚡	docs/how_to/tutorials/export_and_load_executable.py	350	# # Step 3: Upload the compiled library and parameters	COMMENT
1 more matches not shown…

Redundant / Tautological Comments48 hits · 77 pts

Severity	File	Line	Snippet	Context
LOW	python/tvm/topi/sparse_reshape.py	134	# Check if prev_shape and new_shape are equal	COMMENT
LOW	…n/tvm/backend/trn/operator/tile_primitive/dim_utils.py	225	# Check if there is a direct mapping	COMMENT
LOW	python/tvm/backend/cuda/lang/tile_scheduler.py	377	# Check if we're in the full block region	COMMENT
LOW⚡	python/tvm/script/parser/ir/parser.py	146	# Check if Python functions are allowed in this context	COMMENT
LOW	python/tvm/relax/relax_to_pyfunc_converter.py	419	# Check if it's a function parameter	COMMENT
LOW	python/tvm/relax/relax_to_pyfunc_converter.py	424	# Check if it's a bound variable	COMMENT
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	1081	# Check if the tensors are quantized. Parse if yes.	COMMENT
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	3813	# Check if the input tensor is quantized, call QNN op	COMMENT
LOW	…n/tvm/relax/frontend/torch/base_fx_graph_translator.py	174	# Check if this is a null_value expression	COMMENT
LOW	…n/tvm/relax/frontend/torch/base_fx_graph_translator.py	2056	# Check if all indices can be squeezed to 1D for sequential take	COMMENT
LOW	python/tvm/relax/frontend/torch/fx_translator.py	105	# Check if input is integer type and convert to float32 if needed	COMMENT
LOW	python/tvm/relax/frontend/torch/fx_translator.py	115	# Check if input is integer type and convert to float32 if needed	COMMENT
LOW	…vm/relax/frontend/torch/exported_program_translator.py	109	# Check if input is integer type and convert to float32 if needed	COMMENT
LOW	…vm/relax/frontend/torch/exported_program_translator.py	119	# Check if input is integer type and convert to float32 if needed	COMMENT
LOW	python/tvm/relax/backend/cuda/cutlass.py	137	# Check if any dimensions are symbolic.	COMMENT
LOW	python/tvm/relax/backend/cuda/cudnn.py	55	# Check if the data types of input and weights are supported by cuDNN BYOC	COMMENT
LOW⚡	python/tvm/relax/script/parser/parser.py	139	# Check if it calls into a func	COMMENT
LOW	python/tvm/relax/script/parser/parser.py	163	# Check if it is a recursive function	COMMENT
LOW	python/tvm/tirx/stmt_functor.py	750	# Check if anything changed	COMMENT
LOW	python/tvm/tirx/buffer.py	359	# Check if all positional args are integers/PrimExprs with dtype int32 or int64 (the shape) # noqa: E501	COMMENT
LOW	python/tvm/tirx/script/builder/external_kernel.py	162	# Check if NVSHMEM is used - requires cubin output for device library linking	COMMENT
LOW	tests/lint/check_asf_header.py	308	# Check if file should be skipped	COMMENT
LOW	tests/lint/check_asf_header.py	312	# Check if this file type is supported	COMMENT
LOW	tests/python/contrib/test_example_npu.py	98	# Check if the example NPU runtime is available	COMMENT
LOW	tests/python/codegen/test_target_codegen_aarch64.py	634	# Check if the vscale_range() attribute exists	COMMENT
LOW⚡	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	471	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	50	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	97	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	126	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	176	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	220	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	276	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	296	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	319	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	333	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	354	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	388	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	410	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	426	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	444	# Check if the generated ir is expected	COMMENT
LOW	tests/python/tvmscript/test_tvmscript_ir_builder_tir.py	505	# Check if the generated ir is expected	COMMENT
LOW⚡	tests/python/relax/test_tvmscript_pyfunc.py	145	# Check if functions attribute exists	COMMENT
LOW⚡	tests/python/relax/test_tvmscript_pyfunc.py	157	# Check if script method exists	COMMENT
LOW⚡	tests/python/relax/test_tvmscript_pyfunc.py	169	# Check if inheritance flag exists (this might not be set in all implementations)	COMMENT
LOW⚡	tests/python/relax/test_tvmscript_pyfunc.py	176	# Check if original class is preserved (this might not be set in all implementations)	COMMENT
LOW	tests/python/relax/test_base_py_module.py	75	# Check if target contains "cuda" instead of exact match	COMMENT
LOW	tests/scripts/ci.py	552	# Check if it should be a flag	COMMENT
LOW	docs/how_to/tutorials/customize_opt.py	125	# Check if CUBLAS is enabled	COMMENT

Modern Structural Boilerplate63 hits · 64 pts

Severity	File	Line	Snippet	Context
LOW	python/tvm/target/tag.py	21	__all__ = ["list_tags", "register_tag"]	CODE
LOW	python/tvm/runtime/vm.py	90	def _setup_device(self, dev: Device, memory_cfg: str \| dict[Device, str]) -> None:	CODE
LOW	python/tvm/runtime/vm.py	250	def set_input(self, func_name: str, args: Any, *kwargs: Any) -> None:	CODE
LOW	python/tvm/runtime/vm.py	325	def set_instrument(self, instrument: Function) -> None:	CODE
LOW	python/tvm/s_tir/dlight/analysis/common_analysis.py	33	logger = logging.getLogger(__name__) # pylint: disable=invalid-name	CODE
LOW	python/tvm/s_tir/tensor_intrin/riscv_cpu.py	31	logger = logging.getLogger(__name__)	CODE
LOW	python/tvm/backend/__init__.py	27	__all__ = ["is_loaded", "load"]	CODE
LOW	python/tvm/backend/loader.py	117	def _set_module_alias(alias: str, module, *, direct: bool = False) -> None:	CODE
LOW	python/tvm/backend/loader.py	187	__all__ = ["is_loaded", "load"]	CODE
LOW	python/tvm/backend/adreno/__init__.py	39	__all__ = ["register_backend", "target_tags"]	CODE
LOW	python/tvm/backend/trn/layout.py	123	__all__ = ["is_trainium_layout", "to_psum_layout", "trainium_layout"]	CODE
LOW	python/tvm/backend/trn/__init__.py	57	__all__ = [	CODE
LOW	python/tvm/backend/trn/op.py	136	__all__ = [	CODE
LOW	python/tvm/backend/trn/pipeline.py	58	__all__ = ["finalize_device_passes_trn", "trn_pipeline"]	CODE
LOW	python/tvm/backend/trn/script.py	58	__all__ = ["NKINamespace", "OpWrapper"]	CODE
LOW	python/tvm/backend/trn/transform/__init__.py	53	__all__ = ["LowerTIRx", "LowerTrainiumLayout", "TrnNaiveAllocator", "TrnPrivateBufferAlloc"]	CODE
LOW	python/tvm/backend/trn/operator/__init__.py	22	__all__ = ["tile_primitive"]	CODE
LOW	python/tvm/backend/metal/__init__.py	82	__all__ = [	CODE
LOW	python/tvm/backend/metal/op.py	79	__all__ = [	CODE
LOW	python/tvm/backend/metal/script.py	55	__all__ = ["MetalNamespace"]	CODE
LOW	python/tvm/backend/cuda/__init__.py	94	__all__ = [	CODE
LOW	python/tvm/backend/cuda/script.py	573	__all__ = ["CUDANamespace", "NVSHMEMNamespace", "PTXNamespace"]	CODE
LOW	python/tvm/backend/cuda/lang/__init__.py	21	__all__ = [	CODE
LOW	python/tvm/backend/cuda/operator/__init__.py	19	__all__ = ["intrinsics", "tile_primitive"]	CODE
LOW	python/tvm/backend/cuda/operator/intrinsics/__init__.py	42	__all__ = [	CODE
LOW	…ator/tile_primitive/permute_layout/warp_xor_swizzle.py	423	__all__ = [	CODE
LOW	…nd/cuda/operator/tile_primitive/elementwise/_common.py	392	__all__ = [	CODE
LOW	…da/operator/tile_primitive/elementwise/ops/__init__.py	121	__all__ = ["ALL_OPS", "OpSpec", "Plan", "SrcSpec", "VecImpl"]	CODE
LOW	…ckend/cuda/operator/tile_primitive/copy/ld_stmatrix.py	450	__all__ = ["copy_schedule_ldstmatrix"]	CODE
LOW	python/tvm/backend/webgpu/__init__.py	25	__all__ = ["register_backend"]	CODE
LOW	python/tvm/backend/hexagon/__init__.py	51	__all__ = ["register_backend", "target_tags"]	CODE
LOW	python/tvm/backend/opencl/__init__.py	58	__all__ = ["register_backend"]	CODE
LOW	python/tvm/backend/vulkan/__init__.py	69	__all__ = ["register_backend"]	CODE
LOW	python/tvm/backend/rocm/__init__.py	59	__all__ = ["register_backend"]	CODE
LOW	python/tvm/contrib/hexagon/hexagon_unary_ops.py	25	logger = logging.getLogger(__name__)	CODE
LOW	python/tvm/script/parser/core/doc_core.py	954	__all__ = [	CODE
LOW	python/tvm/script/parser/ir/__init__.py	26	__all__ = [	CODE
LOW	python/tvm/arith/analyzer.py	165	def set_z3_timeout_ms(self, timeout_ms: int) -> None:	CODE
LOW	python/tvm/arith/analyzer.py	181	def set_z3_rlimit(self, rlimit: int) -> None:	CODE
LOW	python/tvm/arith/analyzer.py	375	def set_maximum_rewrite_steps(self, maximum: int) -> None:	CODE
LOW	python/tvm/testing/env.py	54	__all__ = [	CODE
LOW	python/tvm/relax/expr_functor.py	1455	def set_var_remap(self, old_var: Var, new_var: Var) -> None:	CODE
LOW	python/tvm/relax/block_builder.py	738	def update_func(self, gv: GlobalVar, updated_func: BaseFunc) -> None:	CODE
LOW	python/tvm/relax/expr.py	883	def _update_type(expr: Expr, ty: Type \| None) -> None:	CODE
LOW	python/tvm/relax/frontend/tflite/tflite_frontend.py	36	__all__ = ["from_tflite"]	CODE
LOW	python/tvm/relax/frontend/nn/core.py	78	def set_default_dtype(dtype: str) -> None:	CODE
LOW	python/tvm/relax/frontend/nn/core.py	342	def set_state(self, state_vars: list[rx.Var]) -> None:	CODE
LOW	python/tvm/relax/frontend/nn/modules.py	49	def set_state(self, state_vars: list[rx.Var]) -> None:	CODE
LOW	python/tvm/relax/frontend/nn/modules.py	838	def set_state(self, state_vars: list[rx.Var]) -> None:	CODE
LOW	python/tvm/relax/frontend/nn/llm/kv_cache.py	85	__all__ = [	CODE
LOW	python/tvm/relax/training/optimizer.py	140	def _set_params_and_dtype(self, params: list[Var]) -> None:	CODE
LOW	…thon/tvm/relax/backend/contrib/example_npu/__init__.py	31	__all__ = ["patterns"]	CODE
LOW	python/tvm/relax/script/parser/dist.py	106	__all__ = ["DTensor", "device_mesh"]	CODE
LOW	python/tvm/relax/script/builder/ir.py	800	__all__ = [	CODE
LOW	python/tvm/tirx/script/tile.py	110	__all__ = [	CODE
LOW	python/tvm/tirx/script/builder/ir.py	3254	__all__ = [	CODE
LOW	python/tvm/tirx/script/builder/tirx.py	1631	__all__ = [	CODE
LOW	python/tvm/tirx/operator/__init__.py	41	__all__ = ["get_tirx_op"]	CODE
LOW	python/tvm/tirx/operator/tile_primitive/__init__.py	30	__all__ = ["DispatchContext", "fail", "list_registered_schedules", "predicate", "register_dispatch"]	CODE
LOW⚡	tests/python/relax/test_vm_build.py	1009	def set_input_attempt_stateless(vm: relax.VirtualMachine, device: tvm.runtime.Device) -> None:	CODE
3 more matches not shown…

Cross-Language Confusion12 hits · 60 pts

Severity	File	Line	Snippet	Context
HIGH	python/tvm/backend/cuda/operator/intrinsics/header.py	643	GmmaDescriptor(GmmaDescriptor && t) noexcept : desc_(t.desc_) {}	CODE
HIGH	python/tvm/backend/cuda/operator/intrinsics/header.py	652	GmmaDescriptor& operator=(GmmaDescriptor && t) noexcept {	CODE
HIGH	python/tvm/contrib/cutlass/gemm_profiler.py	170	for (int i = 1; i < argc && i < 4; ++i) {	STRING
HIGH	python/tvm/relax/frontend/tflite/tflite_flexbuffer.py	87	# Keys are strings here. The format is all strings separated by null, followed by back	COMMENT
HIGH	…ython/s_tir/meta_schedule/test_meta_schedule_runner.py	626	global repeated_args_before # pylint: disable=global-variable-undefined, invalid-name	CODE
HIGH	…ython/s_tir/meta_schedule/test_meta_schedule_runner.py	640	global repeated_args_before # pylint: disable=global-variable-undefined, invalid-name	CODE
HIGH	…ython/s_tir/meta_schedule/test_meta_schedule_runner.py	734	global repeated_args_before # pylint: disable=global-variable-undefined, invalid-name	CODE
HIGH	…ython/s_tir/meta_schedule/test_meta_schedule_runner.py	753	global repeated_args_before # pylint: disable=global-variable-undefined, invalid-name	CODE
HIGH	…ython/s_tir/meta_schedule/test_meta_schedule_runner.py	846	global repeated_args_before # pylint: disable=global-variable-undefined, invalid-name	CODE
HIGH	…ython/s_tir/meta_schedule/test_meta_schedule_runner.py	859	global repeated_args_before # pylint: disable=global-variable-undefined, invalid-name	CODE
HIGH	tests/python/arith/test_arith_z3.py	631	# i.e. EnterConstraint's solver.push()/pop() must be balanced.	COMMENT
HIGH	…s/python/tirx-transform/test_tir_transform_simplify.py	377	simplification that ``!(x \|\| y) == !x && !y``.	STRING

AI Slop Vocabulary24 hits · 50 pts

Severity	File	Line	Snippet	Context
LOW	python/tvm/topi/nn/qnn.py	67	# When disabled, just pass through the input values.	COMMENT
LOW	python/tvm/topi/nn/qnn.py	157	# When disabled simply return the input tensor.	COMMENT
MEDIUM⚡	…ackend/cuda/operator/tile_primitive/elementwise/reg.py	288	# trivial / identity layouts are fine — the algorithm is robust to	COMMENT
LOW	…d/cuda/operator/tile_primitive/elementwise/ops/cast.py	39	# Schedule wraps with Tx.cast(..., dst.dtype) — just pass through.	COMMENT
LOW	python/tvm/script/parser/core/diagnostics.py	87	# Since `findsource` is an internal API of inspect, we just use it	COMMENT
LOW	python/tvm/relax/relax_to_pyfunc_converter.py	642	# Simple fallback: just add the tensors	COMMENT
LOW	python/tvm/relax/frontend/onnx/onnx_frontend.py	2111	# from other frameworks and not actually used in TVM. We'll just use	COMMENT
LOW	…n/tvm/relax/frontend/torch/base_fx_graph_translator.py	2395	# by the translator, and therefore we just return a copy of the input.	COMMENT
LOW	…vm/relax/frontend/torch/exported_program_translator.py	1986	"bernoulli.p": lambda node: self.env[node.args[0]], # Dropout: just return input	CODE
LOW	python/tvm/relax/training/loss.py	364	# In other cases where ignore_index is not needed, just use the simpe product.	STRING
LOW	python/tvm/tirx/op.py	2841	# mode 2, simply use it with multiple Exprs:	COMMENT
LOW	tests/python/tirx-transform/test_tir_functor.py	164	# For demonstration, just return the original for now	COMMENT
MEDIUM	tests/python/relax/test_relax_to_pyfunc_converter.py	354	"""Test that operator mapping is comprehensive."""	STRING
MEDIUM	tests/python/relax/test_base_py_module_printer.py	357	"""Test comprehensive error handling and validation."""	STRING
MEDIUM	tests/python/relax/test_base_py_module_printer.py	361	"""Robust data processing with comprehensive error handling."""	STRING
MEDIUM	docs/how_to/tutorials/e2e_opt_model.py	97	# leverage MetaSchedule to tune the model and store the tuning logs to the database. We also	COMMENT
MEDIUM⚡	docs/how_to/tutorials/export_and_load_executable.py	327	# for a comprehensive guide on:	COMMENT
MEDIUM	docs/how_to/tutorials/optimize_llm.py	169	# - QKV projection: We leverage the horizontal fusion on QKV projection and fuse them into one	COMMENT
MEDIUM	docs/how_to/tutorials/optimize_llm.py	171	# - Attention: We leverage the horizontal fusion on attention and fuse the QKV projection and	COMMENT
MEDIUM	docs/deep_dive/tensor_ir/tutorials/tir_creation.py	263	# The aforementioned lambda expression encapsulates the computation:	COMMENT
MEDIUM	…cs/deep_dive/tensor_ir/tutorials/tir_transformation.py	175	# Alternatively, we can output the IRModule in conjunction with the historical trace.	COMMENT
MEDIUM	docs/get_started/tutorials/quick_start.py	113	# Apache TVM leverage ``pipeline`` to transform and optimize program.	COMMENT
MEDIUM	docs/get_started/tutorials/quick_start.py	125	# In this tutorial we only demonstrate the overall flow, by leverage ``zero`` optimization	COMMENT
LOW	apps/ios_rpc/tests/ios_rpc_test.py	59	# If we don't want to do metal and only use cpu, just set target to be target	COMMENT

Docstring Block Structure4 hits · 20 pts

Severity	File	Line	Snippet	Context
HIGH	python/tvm/topi/nn/pad.py	28	Calculates the output shape of a tensor after applying padding. Args: data (tvm.te.Tensor): The input	STRING
HIGH	python/tvm/backend/cuda/operator/intrinsics/utils.py	40	Validate that cta_group is 1 or 2 and return it as int. Args: cta_group: The cta_group value (can be int or	STRING
HIGH	python/tvm/backend/cuda/operator/intrinsics/utils.py	62	Validate that value is within range and is a power of two. Args: value: The value to validate min_v	STRING
HIGH	…end/cuda/operator/tile_primitive/gemm_async/tcgen05.py	339	Schedule an asynchronous GEMM operation using tcgen05.mma (Blackwell Tensor Core). Computes C = A @ B (with optiona	STRING

Slop Phrases3 hits · 8 pts

Severity	File	Line	Snippet	Context
MEDIUM	python/tvm/relax/frontend/tflite/tflite_frontend.py	8922	# Alternatively you can use tensorflow.keras.application wrapped arounf a tf.function	STRING
MEDIUM	include/tvm/s_tir/transform.h	201	* which is the three blocks as discussed above.	COMMENT
MEDIUM⚡	docs/how_to/tutorials/export_and_load_executable.py	319	# you can use TVM's RPC mechanism to cross-compile, upload, and run the model	COMMENT

Fake / Example Data7 hits · 7 pts

Severity	File	Line	Snippet	Context
LOW	python/tvm/te/operation.py	37	def placeholder(shape, dtype=None, name="placeholder"):	CODE
LOW	python/tvm/relax/frontend/torch/fx_translator.py	1114	if node.op == "placeholder":	CODE
LOW	python/tvm/relax/frontend/torch/dynamo.py	86	if node.op != "placeholder":	CODE
LOW	…vm/relax/frontend/torch/exported_program_translator.py	1517	if node.op == "placeholder":	CODE
LOW	…vm/relax/frontend/torch/exported_program_translator.py	1613	placeholders = [n for n in nodes if n.op == "placeholder"]	CODE
LOW	…vm/relax/frontend/torch/exported_program_translator.py	2126	for node in exported_program.graph.find_nodes(op="placeholder", target=spec.target):	CODE
LOW	include/tvm/te/operation.h	334	std::string name = "placeholder");	CODE

Overly Generic Function Names2 hits · 2 pts

Severity	File	Line	Snippet	Context
LOW	tests/python/relax/test_tvmscript_printer_relax.py	40	def test_function():	CODE
LOW	tests/python/relax/test_expr_functor.py	549	def test_function():	CODE

Dead Code1 hit · 2 pts

Severity	File	Line	Snippet	Context
MEDIUM	tests/python/relax/test_tvmscript_parser.py	748		CODE

Example Usage Blocks1 hit · 2 pts

Severity	File	Line	Snippet	Context
LOW	.agents/scripts/monitor_gpu.sh	7	# Usage:	COMMENT

Analysis Overview

What These Metrics Mean

Score History

Severity Breakdown

Directory Score Breakdown

Pattern Findings