Repository Analysis

pytorch/pytorch

Tensors and Dynamic neural networks in Python with strong GPU acceleration

12.7 Low AI signal View on GitHub

Analysis Overview

This report presents the forensic synthetic code analysis of pytorch/pytorch, a Python project with 101,799 GitHub stars. SynthScan v2.0 examined 4,138,504 lines of code across 10713 source files, recording 52877 pattern matches distributed across 26 syntactic categories. The overall adjusted score of 12.7 places this repository in the Low AI signal band.

The scanner applied 160+ deterministic lexical heuristics, multi-line block detectors, abstract syntax tree depth profilers, and a cross-file Jaccard similarity matrix to construct a statistically normalised synthetic code estimate. All matches are individually weighted by severity coefficient and contextual multiplier before summation, and the resulting headline score is temporally discounted to account for the repository's development history relative to the commercial emergence of large language model coding tooling (November 2022 onward).

12.7
Adjusted Score
12.7
Raw Score
100%
Time Factor
2026-07-14
Last Push
101.8K
Stars
Python
Language
4.1M
Lines of Code
10.7K
Files
52.9K
Pattern Hits
2026-07-14
Scan Date
0.09
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 132HIGH 874MEDIUM 2009LOW 49862

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 52877 distinct pattern matches across 26 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 Identifiers30020 hits · 23141 pts
SeverityFileLineSnippetContext
LOWsetup.py500def extract_variant_from_version(version: str) -> str:CODE
LOWsetup.py620def get_latest_nightly_version(variant: str = "cpu") -> str:CODE
LOWsetup.py653def download_and_extract_nightly_wheel(version: str) -> None:CODE
LOWsetup.py1056def configure_extension_build() -> tuple[STRING
LOW.spin/cmds.py131def regenerate_clangtidy_files():CODE
LOW.spin/cmds.py465def regenerate_github_workflows():CODE
LOWtools/gen_vulkan_spv.py157def extract_leading_whitespace(line: str) -> str:CODE
LOWtools/render_junit.py52def convert_junit_to_testcases(xml: JUnitXml | TestSuite) -> list[TestCase]: # type: ignore[no-any-unimported]CODE
LOWtools/create_worktree.py108def clone_submodule_recursive(CODE
LOWtools/build_with_debinfo.py138def torch_python_link_command() -> str:CODE
LOWtools/packaging/build_wheel.py49def get_supported_python_versions() -> list[str]:CODE
LOWtools/packaging/build_wheel.py70def _find_manylinux_interpreters() -> list[str]:CODE
LOWtools/experimental/torchfuzz/test_determinism.py71def test_deterministic_output():CODE
LOWtools/experimental/torchfuzz/ops_fuzzer.py32def _get_template_filtered_operators(CODE
LOWtools/experimental/torchfuzz/visualize_graph.py136def visualize_operation_graph(CODE
LOWtools/experimental/torchfuzz/visualize_graph.py159def operation_graph_to_networkx(graph: OperationGraph):CODE
LOWtools/experimental/torchfuzz/multi_process_fuzzer.py485def _print_operation_distribution(results: list[FuzzerResult]) -> None:CODE
LOWtools/experimental/torchfuzz/tensor_fuzzer.py190def _compute_contiguous_strides(size: tuple[int, ...]) -> list[int]:CODE
LOWtools/experimental/torchfuzz/tensor_fuzzer.py213def _compute_non_contiguous_dense_strides(size: tuple[int, ...]) -> list[int]:CODE
LOWtools/experimental/torchfuzz/tensor_fuzzer.py304def _compute_storage_size_needed(CODE
LOWtools/experimental/torchfuzz/tensor_fuzzer.py436def fuzz_non_contiguous_dense_tensor(CODE
LOWtools/experimental/torchfuzz/fuzzer.py28def _parse_supported_ops_with_weights(spec: str) -> tuple[list[str], dict[str, float]]:CODE
LOWtools/experimental/torchfuzz/fuzzer.py57def _resolve_generate_only_path(CODE
LOWtools/experimental/torchfuzz/test_streams_template.py31 def test_template_inherits_default_ops(self):CODE
LOWtools/experimental/torchfuzz/test_streams_template.py37 def test_template_uses_backward_check(self):CODE
LOWtools/experimental/torchfuzz/test_streams_template.py47 def test_codegen_creates_streams(self):CODE
LOWtools/experimental/torchfuzz/test_streams_template.py51 def test_codegen_has_stream_context(self):CODE
LOWtools/experimental/torchfuzz/test_streams_template.py55 def test_codegen_has_final_sync(self):CODE
LOWtools/experimental/torchfuzz/test_streams_template.py63 def test_codegen_has_backward(self):CODE
LOWtools/experimental/torchfuzz/test_streams_template.py67 def test_codegen_has_requires_grad(self):CODE
LOWtools/experimental/torchfuzz/test_streams_template.py72 def test_codegen_cross_stream_sync(self):CODE
LOWtools/experimental/torchfuzz/test_streams_template.py84 def test_codegen_event_based_sync(self):CODE
LOWtools/experimental/torchfuzz/test_streams_template.py96 def test_codegen_deterministic(self):CODE
LOWtools/experimental/torchfuzz/test_streams_template.py102 def test_codegen_is_valid_python(self):CODE
LOWtools/experimental/torchfuzz/codegen.py345def convert_graph_to_python_code(CODE
LOWtools/experimental/torchfuzz/codegen.py517def generate_simple_operation_code(CODE
LOWtools/experimental/torchfuzz/type_promotion.py106def get_promotion_table_for_strings() -> dict:CODE
LOWtools/experimental/torchfuzz/type_promotion.py162def get_scalar_promotion_pairs(CODE
LOWtools/experimental/torchfuzz/operators/registry.py38def _concrete_operator_classes(module: ModuleType) -> list[type[Operator]]:CODE
LOWtools/experimental/torchfuzz/operators/registry.py72 def _register_default_operators(self):CODE
LOWtools/experimental/torchfuzz/operators/registry.py148def set_operator_weight_by_torch_op(torch_op_name: str, weight: float) -> None:CODE
LOWtools/experimental/torchfuzz/operators/registry.py159def set_operator_weights_by_torch_op(weights: dict[str, float]) -> None:CODE
LOWtools/experimental/torchfuzz/cuda/_codegen.py567 def _generate_random_placement(self, tensor_size):CODE
LOWtools/test/test_test_selections.py114 def test_some_times_with_not_sort_by_time(self) -> None:CODE
LOWtools/test/test_test_selections.py149 def test_serial_parallel_interleaving(self) -> None:CODE
LOWtools/test/test_test_selections.py565 def test_strips_test_osdc_target(self) -> None:CODE
LOWtools/test/test_test_selections.py573 def test_build_env_with_test_substring_is_preserved(self) -> None:CODE
LOWtools/test/test_test_selections.py582 def test_multi_segment_name_keeps_full_prefix(self) -> None:CODE
LOWtools/test/test_test_selections.py591 def test_unrelated_targets_are_not_matched(self) -> None:CODE
LOWtools/test/test_scoped_library_linter.py28 def test_reports_torch_dot_library_dot_library(self) -> None:CODE
LOWtools/test/test_scoped_library_linter.py53 def test_reports_imported_library_name(self) -> None:CODE
LOWtools/test/test_scoped_library_linter.py76 def test_allows_noqa_alone_on_previous_line(self) -> None:CODE
LOWtools/test/test_scoped_library_linter.py88 def test_allows_multiline_call_if_any_line_has_noqa(self) -> None:CODE
LOWtools/test/test_scoped_library_linter.py102 def test_reports_multiline_call_without_noqa(self) -> None:CODE
LOWtools/test/test_scoped_library_linter.py118 def test_empty_on_syntax_error(self) -> None:CODE
LOWtools/test/test_scoped_library_linter.py124 def test_no_match_for_non_library_call(self) -> None:CODE
LOWtools/test/test_utils.py7 def test_create_from_namespaced_tuple(self) -> None:CODE
LOWtools/test/test_utils.py18 def test_namespace_levels_more_than_max(self) -> None:CODE
LOWtools/test/test_test_run.py24 def test_union_with_inclusions(self) -> None:CODE
LOWtools/test/test_test_run.py33 def test_union_with_non_overlapping_exclusions(self) -> None:CODE
29960 more matches not shown…
Over-Commented Block5429 hits · 5093 pts
SeverityFileLineSnippetContext
LOWCMakeLists.txt741 # Please note that the use of the following flags is required when linkingCOMMENT
LOWCMakeLists.txt941# Note for ROCM platform: 1. USE_ROCM is always ON untilCOMMENT
LOWCMakeLists.txt1221 # skip unwanted includes from windows.hCOMMENT
LOWsetup.py1# Welcome to the PyTorch setup.py.COMMENT
LOWsetup.py21# which we have faithfully adhered to in our build system is that CFLAGSCOMMENT
LOWsetup.py41# DEBUG_CUDA=1COMMENT
LOWsetup.py61# USE_MSLK=0COMMENT
LOWsetup.py81# MKL-DNN threading mode: TBB or OMP (default)COMMENT
LOWsetup.py101# disables distributed MPI backend buildCOMMENT
LOWsetup.py121# This option can be used if AVX512 doesn't perform well on a machine.COMMENT
LOWsetup.py141# specify which AMD GPU targets to build for.COMMENT
LOWsetup.py161# layer normalization. Default: enabled.COMMENT
LOWsetup.py181# Specify a NVCC to use. This is used in our CI to point to a cached nvccCOMMENT
LOWsetup.py201# LIBRARY_PATHCOMMENT
LOWsetup.py221# USE_SYSTEM_PSIMD=ONCOMMENT
LOWtools/git_add_generated_dirs.sh1#!/bin/shCOMMENT
LOWtools/build_with_debinfo.py1#!/usr/bin/env python3COMMENT
LOWtools/git_reset_generated_dirs.sh1#!/bin/shCOMMENT
LOWtools/iwyu/run.sh1# Run include-what-you-use on a file or folderCOMMENT
LOWtools/test/test_stable_shim_usage_linter.py141 shim_functions = get_shim_functions([sample_shim])COMMENT
LOWtools/test/stable_shim_usage_linter_data/sample_usage.h141// Case 18: Correct, fallback function version is lower than current block, target still equal.COMMENT
LOWtools/test/stable_shim_usage_linter_data/sample_shim.h1#ifndef SAMPLE_SHIM_HCOMMENT
LOWtools/test/stable_shim_usage_linter_data/sample_shim.h21AOTI_TORCH_EXPORT void old_function_2(void* ptr);COMMENT
LOWtools/test/stable_shim_usage_linter_data/sample_shim.h41AOTI_TORCH_EXPORT int legacy_fallback(int arg);COMMENT
LOWtools/test/stable_shim_usage_linter_data/sample_shim.h101COMMENT
LOWtools/linter/adapters/flake8_linter.py1# /// scriptCOMMENT
LOWtools/linter/adapters/flake8_linter.py201 # "B901": `return x` inside a generatorCOMMENT
LOWtools/linter/adapters/pyproject_linter.py161 return [COMMENT
LOWtools/linter/adapters/pyproject_linter.py181 # }COMMENT
LOWtools/linter/adapters/ruff_linter.py161COMMENT
LOWtools/linter/adapters/pyrefly_linter.py1# /// scriptCOMMENT
LOWtools/linter/adapters/pyrefly_linter.py21# "optree==0.17.0",COMMENT
LOWtools/autograd/load_derivatives.py341 "forward definition of gradient as element_wise but it does not "COMMENT
LOWtools/autograd/load_derivatives.py561 r.name for r in f.func.returnsCOMMENT
LOWtools/autograd/gen_python_functions.py1# Generates Python bindings for ATen functionsCOMMENT
LOWtools/autograd/gen_python_functions.py21#COMMENT
LOWtools/autograd/gen_python_functions.py1181 return sort_overloads(grouped, symint=symint)COMMENT
LOWtools/autograd/gen_python_functions.py1201# If they have different number of parameters then they are not ambiguous - butCOMMENT
LOWtools/autograd/gen_autograd_functions.py81void will_release_variables() override {COMMENT
LOWtools/autograd/gen_variable_type.py1# Generates VariableType.h/cppCOMMENT
LOWtools/autograd/gen_variable_type.py421# We check the following properties:COMMENT
LOWtools/autograd/gen_variable_type.py1241 # NOTE(crcrpar): in-place foreach functions' arguments include tensorlistCOMMENT
LOWtools/autograd/gen_trace_type.py241# This replacement implemented in-place with minimum modifications ofCOMMENT
LOWtools/autograd/derivatives.yaml1# Defines derivative formulas and Python signatures of methods on VariableCOMMENT
LOWtools/autograd/derivatives.yaml21# of how to register separate derivates for different dispatch keys).COMMENT
LOWtools/autograd/derivatives.yaml41# - If that argument is differentiable, in the sense that a gradient with respectCOMMENT
LOWtools/autograd/derivatives.yaml61# are in scope:COMMENT
LOWtools/autograd/derivatives.yaml81#COMMENT
LOWtools/autograd/derivatives.yaml101# (NB: if your function computes gradient for a list of tensors,COMMENT
LOWtools/autograd/derivatives.yaml121# Another interpretation of this expression would read as `var_0` is neededCOMMENT
LOWtools/autograd/derivatives.yaml141#COMMENT
LOWtools/autograd/derivatives.yaml161# (implemented using zeros_like). These gradients are (hopefully) notCOMMENT
LOWtools/autograd/derivatives.yaml181# according to these three cases:COMMENT
LOWtools/autograd/derivatives.yaml201# It is an efficient way to represent a Tensor filled with zeros becauseCOMMENT
LOWtools/autograd/derivatives.yaml1161COMMENT
LOWtools/autograd/derivatives.yaml1821 self: to_dense_backward(grad, self, masked_grad)COMMENT
LOWtools/autograd/derivatives.yaml1841- name: _to_sparse_csr(Tensor self, int? dense_dim=None) -> TensorCOMMENT
LOWtools/autograd/templates/python_nn_functions.cpp1#define TORCH_ASSERT_ONLY_METHOD_OPERATORSCOMMENT
LOWtools/autograd/templates/python_functions.cpp1#include <torch/csrc/autograd/generated/python_functions.h>COMMENT
LOWtools/autograd/templates/python_fft_functions.cpp1#define TORCH_ASSERT_ONLY_METHOD_OPERATORSCOMMENT
5369 more matches not shown…
Unused Imports4959 hits · 4322 pts
SeverityFileLineSnippetContext
LOWsetup.py245CODE
LOWtools/extract_scripts.py3CODE
LOWtools/nvcc_fix_deps.py16CODE
LOWtools/gen_vulkan_spv.py3CODE
LOWtools/gen_cupti_stubs.py24CODE
LOWtools/embed_libomp_macos.py16CODE
LOWtools/render_junit.py3CODE
LOWtools/strtobool.py3CODE
LOWtools/generate_torch_version.py1CODE
LOWtools/build_with_debinfo.py19CODE
LOWtools/build_pytorch_libs.py1CODE
LOWtools/nightly.py49CODE
LOWtools/experimental/torchfuzz/__init__.py4CODE
LOWtools/experimental/torchfuzz/__init__.py4CODE
LOWtools/experimental/torchfuzz/__init__.py4CODE
LOWtools/experimental/torchfuzz/__init__.py4CODE
LOWtools/experimental/torchfuzz/__init__.py4CODE
LOWtools/experimental/torchfuzz/__init__.py4CODE
LOWtools/experimental/torchfuzz/__init__.py12CODE
LOWtools/experimental/torchfuzz/__init__.py12CODE
LOWtools/experimental/torchfuzz/__init__.py12CODE
LOWtools/experimental/torchfuzz/__init__.py13CODE
LOWtools/experimental/torchfuzz/__init__.py13CODE
LOWtools/experimental/torchfuzz/__init__.py13CODE
LOWtools/experimental/torchfuzz/__init__.py14CODE
LOWtools/experimental/torchfuzz/__init__.py14CODE
LOWtools/experimental/torchfuzz/__init__.py14CODE
LOW…ols/experimental/torchfuzz/operators/loss_functions.py18CODE
LOWtools/experimental/torchfuzz/operators/bitwise.py13CODE
LOWtools/experimental/torchfuzz/operators/logical.py8CODE
LOWtools/experimental/torchfuzz/operators/reduction.py33CODE
LOWtools/experimental/torchfuzz/operators/activations.py33CODE
LOWtools/experimental/torchfuzz/operators/__init__.py3CODE
LOWtools/experimental/torchfuzz/operators/__init__.py4CODE
LOWtools/experimental/torchfuzz/operators/__init__.py5CODE
LOWtools/experimental/torchfuzz/operators/__init__.py6CODE
LOWtools/experimental/torchfuzz/operators/__init__.py7CODE
LOWtools/experimental/torchfuzz/operators/__init__.py8CODE
LOWtools/experimental/torchfuzz/operators/__init__.py9CODE
LOWtools/experimental/torchfuzz/operators/__init__.py10CODE
LOWtools/experimental/torchfuzz/operators/__init__.py10CODE
LOWtools/experimental/torchfuzz/operators/__init__.py10CODE
LOWtools/experimental/torchfuzz/operators/__init__.py10CODE
LOWtools/experimental/torchfuzz/operators/__init__.py10CODE
LOWtools/experimental/torchfuzz/operators/__init__.py10CODE
LOWtools/experimental/torchfuzz/operators/__init__.py10CODE
LOWtools/experimental/torchfuzz/operators/__init__.py10CODE
LOWtools/experimental/torchfuzz/operators/__init__.py20CODE
LOWtools/experimental/torchfuzz/operators/__init__.py20CODE
LOWtools/experimental/torchfuzz/operators/__init__.py20CODE
LOWtools/experimental/torchfuzz/operators/__init__.py20CODE
LOWtools/experimental/torchfuzz/operators/__init__.py26CODE
LOWtools/experimental/torchfuzz/operators/__init__.py26CODE
LOWtools/experimental/torchfuzz/operators/__init__.py26CODE
LOWtools/experimental/torchfuzz/operators/__init__.py26CODE
LOWtools/experimental/torchfuzz/operators/__init__.py26CODE
LOWtools/experimental/torchfuzz/operators/__init__.py26CODE
LOWtools/experimental/torchfuzz/operators/__init__.py26CODE
LOWtools/experimental/torchfuzz/operators/__init__.py26CODE
LOWtools/experimental/torchfuzz/operators/__init__.py36CODE
4899 more matches not shown…
Deep Nesting3900 hits · 2918 pts
SeverityFileLineSnippetContext
LOWsetup.py509CODE
LOWsetup.py653CODE
LOW.spin/cmds.py108CODE
LOW.spin/cmds.py310CODE
LOWtools/extract_scripts.py52CODE
LOWtools/clean.py7CODE
LOWtools/gen_vulkan_spv.py315CODE
LOWtools/gen_vulkan_spv.py376CODE
LOWtools/stale_issues.py12CODE
LOWtools/stale_issues.py81CODE
LOWtools/gen_cupti_stubs.py113CODE
LOWtools/embed_libomp_macos.py28CODE
LOWtools/render_junit.py62CODE
LOWtools/generate_torch_version.py51CODE
LOWtools/download_mnist.py40CODE
LOWtools/create_worktree.py187CODE
LOWtools/nightly.py891CODE
LOWtools/nightly.py1160CODE
LOWtools/nightly.py320CODE
LOWtools/experimental/torchfuzz/test_determinism.py71CODE
LOWtools/experimental/torchfuzz/ops_fuzzer.py254CODE
LOWtools/experimental/torchfuzz/visualize_graph.py41CODE
LOWtools/experimental/torchfuzz/visualize_graph.py159CODE
LOWtools/experimental/torchfuzz/multi_process_fuzzer.py83CODE
LOWtools/experimental/torchfuzz/multi_process_fuzzer.py217CODE
LOWtools/experimental/torchfuzz/multi_process_fuzzer.py517CODE
LOWtools/experimental/torchfuzz/tensor_fuzzer.py99CODE
LOWtools/experimental/torchfuzz/tensor_fuzzer.py213CODE
LOWtools/experimental/torchfuzz/tensor_fuzzer.py320CODE
LOWtools/experimental/torchfuzz/tensor_fuzzer.py484CODE
LOWtools/experimental/torchfuzz/fuzzer.py80CODE
LOWtools/experimental/torchfuzz/fuzzer.py278CODE
LOWtools/experimental/torchfuzz/codegen.py517CODE
LOWtools/experimental/torchfuzz/codegen.py198CODE
LOW…s/experimental/torchfuzz/operators/tensor_pointwise.py61CODE
LOWtools/experimental/torchfuzz/operators/layout.py192CODE
LOWtools/experimental/torchfuzz/operators/constant.py46CODE
LOWtools/experimental/torchfuzz/cuda/_codegen.py159CODE
LOWtools/experimental/torchfuzz/cuda/_codegen.py398CODE
LOWtools/experimental/torchfuzz/cuda/_codegen.py581CODE
LOWtools/test/test_gen_backend_stubs.py27CODE
LOWtools/test/test_cmake.py22CODE
LOWtools/test/heuristics/test_interface.py413CODE
LOWtools/test/heuristics/test_interface.py437CODE
LOWtools/linter/adapters/stable_shim_usage_linter.py35CODE
LOWtools/linter/adapters/stable_shim_usage_linter.py138CODE
LOWtools/linter/adapters/pyfmt_linter.py150CODE
LOWtools/linter/adapters/generated_shims_version_linter.py147CODE
LOWtools/linter/adapters/no_workflows_on_fork.py90CODE
LOWtools/linter/adapters/no_merge_conflict_csv_linter.py62CODE
LOWtools/linter/adapters/stable_shim_version_linter.py35CODE
LOWtools/linter/adapters/stable_shim_version_linter.py102CODE
LOWtools/linter/adapters/stable_shim_version_linter.py50CODE
LOWtools/linter/adapters/test_device_bias_linter.py45CODE
LOWtools/linter/adapters/test_device_bias_linter.py46CODE
LOWtools/linter/adapters/test_device_bias_linter.py170CODE
LOWtools/linter/adapters/_stable_shim_utils.py332CODE
LOWtools/linter/adapters/header_only_linter.py45CODE
LOWtools/linter/adapters/clangformat_linter.py173CODE
LOWtools/linter/adapters/docstring_linter.py195CODE
3840 more matches not shown…
Decorative Section Separators862 hits · 2661 pts
SeverityFileLineSnippetContext
MEDIUMtools/experimental/torchfuzz/codegen.py294# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/codegen.py296# ---------------------------------------------------------------------------COMMENT
MEDIUM…ols/experimental/torchfuzz/operators/loss_functions.py49# ---------------------------------------------------------------------------COMMENT
MEDIUM…ols/experimental/torchfuzz/operators/loss_functions.py51# ---------------------------------------------------------------------------COMMENT
MEDIUM…ols/experimental/torchfuzz/operators/loss_functions.py98# ---------------------------------------------------------------------------COMMENT
MEDIUM…ols/experimental/torchfuzz/operators/loss_functions.py100# ---------------------------------------------------------------------------COMMENT
MEDIUM…ols/experimental/torchfuzz/operators/loss_functions.py214# ---------------------------------------------------------------------------COMMENT
MEDIUM…ols/experimental/torchfuzz/operators/loss_functions.py216# ---------------------------------------------------------------------------COMMENT
MEDIUM…ols/experimental/torchfuzz/operators/loss_functions.py249# ---------------------------------------------------------------------------COMMENT
MEDIUM…ols/experimental/torchfuzz/operators/loss_functions.py251# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/bitwise.py66# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/bitwise.py68# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/bitwise.py182# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/bitwise.py184# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/bitwise.py202# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/bitwise.py204# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/bitwise.py217# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/bitwise.py219# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/logical.py36# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/logical.py38# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/logical.py82# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/logical.py84# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/logical.py102# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/logical.py104# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/reduction.py98# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/reduction.py100# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/reduction.py170# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/reduction.py172# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/reduction.py209# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/reduction.py211# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/reduction.py499# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/reduction.py501# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/reduction.py625# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/reduction.py627# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/activations.py44# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/activations.py51# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/activations.py109# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/activations.py111# ---------------------------------------------------------------------------COMMENT
MEDIUM…ls/experimental/torchfuzz/operators/tensor_creation.py107# ---------------------------------------------------------------------------COMMENT
MEDIUM…ls/experimental/torchfuzz/operators/tensor_creation.py109# ---------------------------------------------------------------------------COMMENT
MEDIUM…ls/experimental/torchfuzz/operators/tensor_creation.py269# ---------------------------------------------------------------------------COMMENT
MEDIUM…ls/experimental/torchfuzz/operators/tensor_creation.py271# ---------------------------------------------------------------------------COMMENT
MEDIUM…ls/experimental/torchfuzz/operators/tensor_creation.py317# ---------------------------------------------------------------------------COMMENT
MEDIUM…ls/experimental/torchfuzz/operators/tensor_creation.py319# ---------------------------------------------------------------------------COMMENT
MEDIUM…ls/experimental/torchfuzz/operators/tensor_creation.py356# ---------------------------------------------------------------------------COMMENT
MEDIUM…ls/experimental/torchfuzz/operators/tensor_creation.py358# ---------------------------------------------------------------------------COMMENT
MEDIUM…ls/experimental/torchfuzz/operators/tensor_creation.py410# ---------------------------------------------------------------------------COMMENT
MEDIUM…ls/experimental/torchfuzz/operators/tensor_creation.py412# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/comparison.py84# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/comparison.py86# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/comparison.py146# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/comparison.py148# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/comparison.py206# ---------------------------------------------------------------------------COMMENT
MEDIUMtools/experimental/torchfuzz/operators/comparison.py208# ---------------------------------------------------------------------------COMMENT
MEDIUM…s/experimental/torchfuzz/operators/elementwise_math.py33# ---------------------------------------------------------------------------COMMENT
MEDIUM…s/experimental/torchfuzz/operators/elementwise_math.py35# ---------------------------------------------------------------------------COMMENT
MEDIUM…s/experimental/torchfuzz/operators/elementwise_math.py181# ---------------------------------------------------------------------------COMMENT
MEDIUM…s/experimental/torchfuzz/operators/elementwise_math.py183# ---------------------------------------------------------------------------COMMENT
MEDIUM…s/experimental/torchfuzz/operators/elementwise_math.py213# ---------------------------------------------------------------------------COMMENT
MEDIUM…s/experimental/torchfuzz/operators/elementwise_math.py215# ---------------------------------------------------------------------------COMMENT
802 more matches not shown…
Cross-Language Confusion487 hits · 2425 pts
SeverityFileLineSnippetContext
HIGHtools/linter/adapters/lintrunner_version_linter.py57 description="Lintrunner is not installed, did you forget to run `make setup-lint && make lint`?",CODE
HIGH.ci/lumen_cli/tests/test_run_plan.py99 "export A=x && pytest -q",CODE
HIGH.ci/lumen_cli/tests/test_run_plan.py100 "export B=y && pytest -q tests/unit",CODE
HIGHtest/test_datapipe.py565 self.assertTrue(df.equals(cdf))CODE
HIGHtest/test_schema_check.py459 # Empty containers return falseCOMMENT
HIGHtest/test_jit.py6521 # 1. nan == nan should return trueSTRING
HIGHtest/test_pytree.py1271 '"children_spec": [{"type": null, "context": null, 'STRING
HIGHtest/test_pytree.py1272 '"children_spec": []}, {"type": null, "context": null, 'STRING
HIGHtest/test_pytree.py1273 '"children_spec": []}]}, {"type": null, "context": null, 'STRING
HIGHtest/test_pytree.py1275 '"[4, 5, 6]", "children_spec": [{"type": null, "context": null, 'STRING
HIGHtest/test_pytree.py1276 '"children_spec": []}, {"type": null, "context": null, "children_spec": 'STRING
HIGHtest/test_pytree.py1277 '[]}, {"type": null, "context": null, "children_spec": []}]}]}]'STRING
HIGHtest/test_cuda.py10153 if (row < M && col < N) {CODE
HIGHtest/test_opaque_obj_v2.py470 queue.push(b)CODE
HIGHtest/test_meta.py1262 with MetaCrossRefDispatchMode.push(STRING
HIGHtest/inductor/test_torchbind.py122 q.push(torch.ones(2, 3))CODE
HIGHtest/inductor/test_torchbind.py350 self.tq.push(x.cos())CODE
HIGHtest/inductor/test_torchbind.py351 self.tq.push(x.sin())CODE
HIGHtest/cpp_extensions/test_libtorch_agnostic.py2147 # Runtime predates the shim: dlsym returns null, so we fall back toCOMMENT
HIGHtest/nn/test_pooling.py843 "MaxUnpool2d": r"Assertion `maxind >= 0 && maxind < outputImageSize` failed",CODE
HIGHtest/nn/test_pooling.py844 "MaxUnpool3d": r"Assertion `index >= 0 && index < outputImageSize` failed",CODE
HIGHtest/nn/test_embedding.py1165 msg = "Expected idx >= 0 && idx < num_embeddings"CODE
HIGHtest/distributed/test_distributed_spawn.py42 f"e.g.\ntouch /tmp/barrier && TEMP_DIR=/tmp BACKEND='nccl' WORLD_SIZE=2 python {__file__}",CODE
HIGHtest/distributed/test_fake_pg.py76 # default init path leaves the backend options null).COMMENT
HIGHtest/fx/test_fx_split.py211 # pyre-fixme[61]: `name` is undefined, or not always defined.COMMENT
HIGHtest/quantization/eager/test_equalize_eager.py56 identical modules should return trueSTRING
HIGHtest/quantization/eager/test_equalize_eager.py57 modules with high difference in weights should return falseSTRING
HIGHtest/cpython/v3_13/test_contextlib.py812 stack.push(_expect_ok)CODE
HIGHtest/cpython/v3_13/test_contextlib.py815 stack.push(cm)CODE
HIGHtest/cpython/v3_13/test_contextlib.py817 stack.push(_suppress_exc)CODE
HIGHtest/cpython/v3_13/test_contextlib.py820 stack.push(cm)CODE
HIGHtest/cpython/v3_13/test_contextlib.py822 stack.push(_expect_exc)CODE
HIGHtest/cpython/v3_13/test_contextlib.py824 stack.push(_expect_exc)CODE
HIGHtest/cpython/v3_13/test_contextlib.py893 stack.push(lambda *exc: False)CODE
HIGHtest/cpython/v3_13/test_contextlib.py898 stack.push(lambda *exc: True)CODE
HIGHtest/cpython/v3_13/test_contextlib.py1066 stack.push(suppress_exc)CODE
HIGHtest/cpython/v3_13/test_contextlib.py1137 stack.push(suppress_exc)CODE
HIGHtest/cpython/v3_13/test_contextlib.py1144 stack.push(lambda *exc: True)CODE
HIGHtest/cpython/v3_13/test_contextlib.py1145 stack.push(lambda *exc: 1/0)CODE
HIGHtest/cpython/v3_13/test_contextlib.py1146 stack.push(lambda *exc: {}[1])CODE
HIGHtest/cpython/v3_13/test_contextlib.py997 stack.push(suppress_exc)CODE
HIGHtest/cpython/v3_13/test_contextlib.py1163 stack.push(cm)CODE
HIGHtest/jit/test_torchbind.py98 stackstring.push("lel")CODE
HIGHtest/jit/test_torchbind.py61 ss1.push(ss2.pop())CODE
HIGHtest/lazy/test_functionalization.py98 %1 = [BFloat16[3]] aten::_to_copy(%0), dtype=BFloat16, layout=null, device=null, pin_memory=null, non_blocking=0, memoSTRING
HIGHtest/export/test_torchbind.py1106 tq.push(torch.zeros(2, 2))CODE
HIGHtest/export/test_torchbind.py1107 tq.push(torch.ones(2, 2))CODE
HIGHtest/export/test_torchbind.py1109 tq.push(torch.ones(2, 2) * 2)CODE
HIGHtest/export/test_torchbind.py1175 tq.push(x.cos())CODE
HIGHtest/export/test_torchbind.py1176 tq.push(x.sin())CODE
HIGHtest/export/test_torchbind.py1236 tq.push(x.cos())CODE
HIGHtest/export/test_torchbind.py1237 tq.push(x.sin())CODE
HIGHtest/export/test_torchbind.py1251 tq2.push(torch.randn(4, 5, requires_grad=False))CODE
HIGHtest/export/test_torchbind.py1257 tq3.push(torch.randn(2, 3, requires_grad=False))CODE
HIGHtest/export/test_torchbind.py1263 tq4.push(torch.randn(2, 3, requires_grad=False))CODE
HIGHtest/export/test_torchbind.py1269 tq5.push(torch.randn(2, 3, requires_grad=True))CODE
HIGHtest/export/test_torchbind.py1275 tq6.push(torch.randn(2, 3, requires_grad=True, dtype=torch.float64))CODE
HIGHtest/export/test_torchbind.py1287 tq.push(x.cos())CODE
HIGHtest/export/test_torchbind.py1288 tq.push(x.sin())CODE
HIGHtest/export/test_torchbind.py1297 tq1.push(torch.randn(2, 3, requires_grad=False))CODE
427 more matches not shown…
Self-Referential Comments664 hits · 1900 pts
SeverityFileLineSnippetContext
MEDIUMsetup.py517 # Create a temporary directory for downloadingCOMMENT
MEDIUMsetup.py663 # Create a temporary directory for downloadingCOMMENT
MEDIUMtools/update_masked_docs.py30# This file is generated, do not modify it!STRING
MEDIUMtools/experimental/torchfuzz/ops_fuzzer.py472 # Create the operation nodeCOMMENT
MEDIUMtools/experimental/torchfuzz/multi_process_fuzzer.py22 # Create a mock tqdm class for type safetyCOMMENT
MEDIUMtools/experimental/torchfuzz/tensor_fuzzer.py251 # Create a valid permutation that's still denseCOMMENT
MEDIUMtools/experimental/torchfuzz/tensor_fuzzer.py344 # Create a local Random instance to avoid interfering with global stateCOMMENT
MEDIUMtools/experimental/torchfuzz/tensor_fuzzer.py499 # Create a local random instance to avoid interfering with global stateCOMMENT
MEDIUMtools/experimental/torchfuzz/tensor_fuzzer.py507 # Create a scalar value based on dtypeCOMMENT
MEDIUMtools/experimental/torchfuzz/tensor_fuzzer.py527 # Create a scalar value based on dtypeCOMMENT
MEDIUMtools/experimental/torchfuzz/fuzzer.py129 # Create a unique folder for this iterationCOMMENT
MEDIUMtools/experimental/torchfuzz/operators/item.py27 # Create a tensor spec that can produce a scalar via .item()COMMENT
MEDIUMtools/test/test_gb_registry_linter.py438 # Create a callsite with one of the duplicate typesCOMMENT
MEDIUMtools/test/heuristics/test_interface.py29 # Create a dummy heuristic classCOMMENT
MEDIUMtools/linter/adapters/stable_shim_usage_linter.py125 "# This file is automatically updated by the stable_shim_usage_linter.\n"CODE
MEDIUMtools/linter/adapters/s3_init.py122 # Create the output folderCOMMENT
MEDIUMtools/autograd/gen_variable_type.py431# The following code templates implement the checks for this invariant:COMMENT
MEDIUMtools/autograd/derivatives.yaml1483# Defining a backward will make codegen spit out the forward call asCOMMENT
MEDIUMtools/setup_helpers/gen_version_header.py59 # Create the output dir if it doesn't exist.COMMENT
MEDIUM.ci/lumen_cli/tests/test_utils.py26 # Create a temporary directory for the test to run inCOMMENT
MEDIUM.ci/lumen_cli/cli/lib/core/vllm/vllm_test_library.yaml2# This file defines the test plans for vllm CI developmentCOMMENT
MEDIUM.ci/libtorch/extract_libtorch_from_wheel.py339 # Create the zipCOMMENT
MEDIUMtest/test_python_dispatch.py743 # Create a fragmentSTRING
MEDIUMtest/test_linalg.py10011 # Create the results file in the parent so we can always clean it up,STRING
MEDIUMtest/test_functionalization.py1825 # Create a view of xCOMMENT
MEDIUMtest/test_hub.py316 # Create an entry with malicious pathCOMMENT
MEDIUMtest/test_hub.py322 # Create a single entry with malicious path for legacy formatCOMMENT
MEDIUMtest/test_hub.py390 # Create a legacy-style malicious zip (single entry with traversal)COMMENT
MEDIUMtest/test_hub.py412 # Create a zip that looks like a GitHub repository but has malicious contentCOMMENT
MEDIUMtest/test_matmul_cuda.py1072 # Create a doubly-non-contiguous 2D matrix: stride on both dims > 1COMMENT
MEDIUMtest/test_torch.py3469 # Create the 4 possible combinations of scalar sizes for source / indexCOMMENT
MEDIUMtest/test_torch.py3518 # Create the 8 possible combinations of scalar sizes for target / index / sourceCOMMENT
MEDIUMtest/test_torch.py8718 # This method is primarily exposed for torchvision's resizeCOMMENT
MEDIUMtest/test_torch.py11020# The following block extends TestTorch with negative dim wrapping testsCOMMENT
MEDIUMtest/test_dynamic_shapes.py6672 # Create a backed symbolCOMMENT
MEDIUMtest/test_dynamic_shapes.py6677 # Create an unbacked symbolCOMMENT
MEDIUMtest/test_dynamic_shapes.py4099 # Create a symint during fake tensor tracingCOMMENT
MEDIUMtest/test_dynamic_shapes.py5204 # Create a scalar tensor to scatter into the selected indexCOMMENT
MEDIUMtest/test_dynamic_shapes.py6204 # Create a tensor with shape [..., 2] where batch dim is unbackedCOMMENT
MEDIUMtest/test_dynamic_shapes.py6618 # Create a derived expression: u0 // 2COMMENT
MEDIUMtest/test_dynamic_shapes.py6762 # Create a backed foreign symbol that this env cannot translate.COMMENT
MEDIUMtest/test_optim.py2069 # Create a random 2D tensor for compatibility with Muon.COMMENT
MEDIUMtest/test_nn.py8654 # Create an appropriately-sized input with a single spatial element.COMMENT
MEDIUMtest/test_subclass.py231 # Define a subclass that does not rewrap for any function in its __torch_dispatch__ impl.COMMENT
MEDIUMtest/test_indexing.py1952 # Create the 8 possible combinations of scalar sizes for target / index / sourceCOMMENT
MEDIUMtest/test_indexing.py2298 # Create the 4 possible combinations of scalar sizes for index / sourceCOMMENT
MEDIUMtest/test_custom_ops.py1542 # Define a named tuple for a Point with x and y coordinatesCOMMENT
MEDIUMtest/test_cuda.py5078 # Create a pipe to communicate between parent and child processesCOMMENT
MEDIUMtest/test_cuda.py2828 # Define a function to create generator states, with optional graph registrationCOMMENT
MEDIUMtest/test_cuda.py2842 # Define a function to perform specific RNG actions using the generator's statesCOMMENT
MEDIUMtest/test_cuda.py2859 # Define a function to retrieve the final offsets of the original and new generator statesCOMMENT
MEDIUMtest/test_cuda.py5504 # Create a simple model with a backward passCOMMENT
MEDIUMtest/test_cuda.py5690 # Create an instance of the modelCOMMENT
MEDIUMtest/test_cuda.py5695 # Create a sample input tensorCOMMENT
MEDIUMtest/test_cuda.py2529 # This function is intended to test the lazy creation and reuse of per-threadSTRING
MEDIUMtest/test_cuda.py10219 # Define a simple vector addition kernelSTRING
MEDIUMtest/test_opaque_obj_v2.py99# Create a class with an unsupported descriptorCOMMENT
MEDIUMtest/test_opaque_obj_v2.py218 # Create a new counter to match device mesh's __getitem__COMMENT
MEDIUMtest/test_dlpack.py421 # Create a new tensor, and wrap it using TensorDLPackWrapper.COMMENT
MEDIUMtest/test_dlpack.py424 # Create a new tensor from the wrapper.COMMENT
604 more matches not shown…
Excessive Try-Catch Wrapping1357 hits · 1385 pts
SeverityFileLineSnippetContext
LOWsetup.py698 except Exception:CODE
LOWsetup.py706 except Exception:CODE
LOWtools/render_junit.py33 except Exception as err:CODE
MEDIUMtools/nightly_hotpatch.py94 print("Error: PyTorch is not installed in the current Python environment.")CODE
LOWtools/nightly_hotpatch.py130 except Exception as e:CODE
MEDIUMtools/nightly_hotpatch.py131 print(f"An error occurred while downloading the patch: {e}")CODE
MEDIUMtools/nightly_hotpatch.py178 print("Error: The 'patch' utility is not installed or not found in PATH.")CODE
LOWtools/nightly_hotpatch.py180 except Exception as e:CODE
MEDIUMtools/nightly_hotpatch.py181 print(f"An error occurred while applying the patch: {e}")CODE
LOWtools/generate_torch_version.py30 except Exception:CODE
LOWtools/generate_torch_version.py47 except Exception:CODE
LOWtools/nightly.py721 except Exception as e:CODE
LOWtools/nightly.py945 except Exception:CODE
LOWtools/experimental/torchfuzz/ops_fuzzer.py238 except Exception:CODE
LOWtools/experimental/torchfuzz/multi_process_fuzzer.py183 except Exception as e:CODE
LOWtools/experimental/torchfuzz/multi_process_fuzzer.py363 except Exception as e:CODE
LOWtools/experimental/torchfuzz/fuzzer.py266 except Exception as e:CODE
MEDIUMtools/experimental/torchfuzz/fuzzer.py274 print(f"Error: {error_message}")CODE
LOWtools/experimental/torchfuzz/fuzzer.py437 except Exception as e:CODE
LOWtools/experimental/torchfuzz/fuzzer.py476 except Exception as e:CODE
LOWtools/test/gen_operators_yaml_test.py141 except Exception:CODE
LOWtools/linter/adapters/gha_linter.py61 except Exception as err:CODE
LOWtools/linter/adapters/actionlint_linter.py168 except Exception:STRING
LOWtools/linter/adapters/grep_linter.py136 except Exception as err:CODE
LOWtools/linter/adapters/grep_linter.py216 except Exception as err:CODE
LOWtools/linter/adapters/grep_linter.py397 except Exception as err:CODE
LOWtools/linter/adapters/cmake_linter.py144 except Exception:STRING
LOWtools/linter/adapters/pyfmt_linter.py138 except Exception as err:CODE
LOWtools/linter/adapters/pyfmt_linter.py193 except Exception:CODE
LOWtools/linter/adapters/no_workflows_on_fork.py242 except Exception:CODE
LOWtools/linter/adapters/no_merge_conflict_csv_linter.py97 except Exception:CODE
LOWtools/linter/adapters/stable_shim_version_linter.py94 except Exception as e:CODE
LOWtools/linter/adapters/nativefunctions_linter.py75 except Exception as err:CODE
LOWtools/linter/adapters/clangformat_linter.py260 except Exception:CODE
LOWtools/linter/adapters/docstring_linter.py92 except Exception as e:STRING
MEDIUMtools/linter/adapters/docstring_linter.py86def _grandfather(self) -> dict[str, dict[str, Any]]:CODE
LOWtools/linter/adapters/pyproject_linter.py250 except Exception:CODE
LOWtools/linter/adapters/clangtidy_linter.py335 except Exception:STRING
LOWtools/linter/adapters/cmake_minimum_required_linter.py186 except Exception:CODE
LOWtools/linter/adapters/cmake_minimum_required_linter.py246 except Exception:CODE
LOWtools/linter/adapters/ruff_linter.py479 except Exception: # Catch all exceptions for lintrunnerCODE
LOWtools/linter/adapters/newlines_linter.py96 except Exception as err:CODE
LOWtools/linter/adapters/newlines_linter.py135 except Exception as err:CODE
LOWtools/linter/adapters/codespell_linter.py132 except Exception as err:CODE
LOWtools/linter/adapters/codespell_linter.py159 except Exception as err:CODE
LOWtools/linter/adapters/codespell_linter.py216 except Exception:CODE
LOWtools/dynamo/verify_dynamo.py187 except Exception:CODE
LOWtools/testing/upload_artifacts.py121 except Exception as e:CODE
LOWtools/testing/upload_artifacts.py207 except Exception as e:CODE
LOWtools/testing/upload_artifacts.py219 except Exception as e:CODE
LOWtools/testing/target_determination/determinator.py30 except Exception as e:CODE
MEDIUMtools/testing/target_determination/determinator.py31 print(f"Error in {heuristic.__class__.__name__}: {e}", file=file)CODE
LOW…ing/target_determination/heuristics/mentioned_in_pr.py33 except Exception as e:CODE
LOW…ing/target_determination/heuristics/mentioned_in_pr.py42 except Exception as e:CODE
LOW…ing/target_determination/heuristics/mentioned_in_pr.py53 except Exception:CODE
LOWtools/stats/monitor.py244 except Exception as e:CODE
LOWtools/stats/monitor.py323 except Exception as e:CODE
LOWtools/stats/monitor.py472 except Exception as e:CODE
LOWtools/stats/monitor.py499 except Exception:CODE
LOWtools/stats/monitor.py524 except Exception:CODE
1297 more matches not shown…
Hallucination Indicators132 hits · 1255 pts
SeverityFileLineSnippetContext
CRITICALtools/autograd/load_derivatives.py292 is_foreach = f.func.name.name.base.startswith("_foreach_")CODE
CRITICALtools/autograd/gen_autograd_functions.py694 info.func.func.name.name.base.startswith("_foreach") and is_outputSTRING
CRITICALtest/test_per_overload_api.py73 torch.ops.aten.linear.default.decompose(x, y),CODE
CRITICALtest/inductor/test_gpu_select_algorithm.py104 w_int8 = torch.ao.quantization.fx._decomposed.quantize_per_channel(CODE
CRITICALtest/inductor/test_gpu_select_algorithm.py159 w_int8 = torch.ao.quantization.fx._decomposed.quantize_per_channel(CODE
CRITICALtest/inductor/test_cpu_select_algorithm.py1734 w_int8 = torch.ao.quantization.fx._decomposed.quantize_per_channel(CODE
CRITICALtest/inductor/test_cpu_select_algorithm.py1802 w_int8 = torch.ao.quantization.fx._decomposed.quantize_per_channel(CODE
CRITICALtest/onnx/test_pytorch_onnx_onnxruntime.py12879 model = torch.ao.nn.intrinsic.quantized.ConvReLU1d(16, 33, 3, stride=2)CODE
CRITICALtest/onnx/test_pytorch_onnx_onnxruntime.py12893 model = torch.ao.nn.intrinsic.quantized.ConvReLU2d(16, 33, 3, stride=2)CODE
CRITICALtest/onnx/test_pytorch_onnx_onnxruntime.py12908 model = torch.ao.nn.intrinsic.quantized.ConvReLU3d(CODE
CRITICALtest/fx/test_dce_pass.py232 torch._ops.ops.aten.add_.Tensor(a, 1)CODE
CRITICALtest/fx/test_dce_pass.py314 torch._ops.ops.aten.add.out(b, b, out=a, alpha=2)CODE
CRITICALtest/dynamo/test_minifier.py337 conv = torch.ao.nn.intrinsic.qat.ConvBnReLU2d(CODE
CRITICALtest/dynamo/test_minifier.py359 "torch.ao.nn.intrinsic.qat.modules.conv_fused.ConvBnReLU2d(",CODE
CRITICALtest/dynamo/test_minifier.py387 conv = torch.ao.nn.intrinsic.qat.ConvBnReLU2d(CODE
CRITICALtest/dynamo/test_reconstruct.py375 return self.a.b.c.d.e(x)CODE
CRITICALtest/quantization/core/test_quantized_module.py1742 cell_dq = torch.ao.nn.quantized.dynamic.LSTM(input_size=input_size,STRING
CRITICALtest/quantization/core/test_quantized_module.py1750 ref_dq = torch.ao.nn.quantized.dynamic.LSTM(input_size=input_size,STRING
CRITICALtest/quantization/core/test_quantized_module.py1804 cell_dq = torch.ao.nn.quantized.dynamic.GRU(input_size=input_size,STRING
CRITICALtest/quantization/core/test_quantized_op.py678 qY = torch.ao.nn.quantized.functional.elu(qX, output_scale, output_zero_point, alpha=alpha)CODE
CRITICALtest/quantization/core/test_quantized_op.py1037 qY = torch.ao.nn.quantized.functional.hardswish(CODE
CRITICALtest/quantization/core/test_quantized_op.py3055 qY = torch.ao.nn.quantized.functional.hardtanh(qX, -1, 6)CODE
CRITICALtest/quantization/core/test_quantized_op.py9148 qY_hat = torch.ao.nn.quantized.functional.hardtanh(qX, min_val, max_val)CODE
CRITICALtest/quantization/fx/test_quantize_fx.py4944 return torch.ao.quantization.fx.lstm_utils._get_lstm_with_individually_observed_parts(CODE
CRITICALtest/quantization/fx/test_quantize_fx.py4964 return torch.ao.quantization.fx.lstm_utils._get_reference_quantized_lstm_module(CODE
CRITICALtest/quantization/fx/test_quantize_fx.py6770 activation=torch.ao.quantization.observer.HistogramObserver.with_args(CODE
CRITICALtest/quantization/jit/test_quantize_jit.py3844 eager_model.sub.module.fc1.weight.detach()CODE
CRITICALtest/quantization/jit/test_quantize_jit.py3847 eager_model.sub.module.fc1.bias.detach()CODE
CRITICALtest/quantization/jit/test_quantize_jit.py3850 eager_model.sub.module.fc2.weight.detach()CODE
CRITICALtest/quantization/jit/test_quantize_jit.py3853 eager_model.sub.module.fc2.bias.detach()CODE
CRITICALtest/quantization/jit/test_quantize_jit.py3798 eager_model.sub2.fc1.module.weight.detach()STRING
CRITICALtest/quantization/jit/test_quantize_jit.py3801 eager_model.sub2.fc1.module.bias.detach()STRING
CRITICALtest/quantization/eager/test_quantize_eager_ptq.py1078 original_m.sub.custom.conv.weight.detach()STRING
CRITICALtest/quantization/eager/test_quantize_eager_ptq.py1081 original_m.sub.custom.conv.bias.detach()STRING
CRITICALtest/cpp/jit/test_backend.cpp49 __torch__.torch.classes.__backends__.test_backend.__new__(__torch__.torch.classes.__backends__.test_backend)CODE
CRITICALtest/jit/test_class_type.py556 three = jit._imported_class_test.very.very.nested.FooUniqueName(a)CODE
CRITICALtest/export/test_torchbind.py344 torch.ops._TorchScriptTesting.takes_foo.default._dispatch_cache.clear()CODE
CRITICALtest/export/test_torchbind.py355 torch.ops._TorchScriptTesting.takes_foo.default.py_impl(CODE
CRITICALtest/export/test_export.py12870 x = self.n3.n4.n5.n6.n7.n8(x + 1)CODE
CRITICALtest/export/test_export.py12872 self.n3.n4.n5.buf.add_(1)CODE
CRITICALtest/export/test_export.py12873 self.n3.n4.n5.n6.buf.add_(1)CODE
CRITICALtest/export/test_export.py12874 self.n3.n4.n5.n6.n7.n8.buf.add_(1)CODE
CRITICALtest/export/test_export.py12909 x = self.n1.n2.n3.n4.n5.n6.n7(x + 1)CODE
CRITICALtest/export/test_export.py12910 self.n1.n2.n3.buf.add_(1)CODE
CRITICALtest/export/test_export.py12911 self.n1.n2.n3.n4.n5.n6.buf.add_(1)CODE
CRITICALtest/export/test_export.py12912 self.n1.n2.n3.n4.n5.n6.n7.n8.buf.add_(1)CODE
CRITICALtest/export/test_export.py13289 x = self.n1.n2.n3.n4.n5(x + 1)CODE
CRITICALtest/export/test_export.py13290 x = self.n1.n2.n3.n4.n5.n6(x + 1)CODE
CRITICALtest/export/test_export.py13292 self.n1.n2.n3.n4.buf.add_(1)CODE
CRITICALtest/export/test_export.py13293 self.n1.n2.n3.n4.n5.buf.add_(1)CODE
CRITICALtest/export/test_export.py13294 self.n1.n2.n3.n4.n5.n6.buf.add_(1)CODE
CRITICALtest/export/test_export.py13388 x = self.n5.n6.n7.n8.n9(x + 1)CODE
CRITICALtest/export/test_export.py13391 self.n5.n6.n7.buf.add_(1)CODE
CRITICALtest/export/test_export.py13392 self.n5.n6.n7.n8.buf.add_(1)CODE
CRITICALtest/export/test_export.py13409 x = self.n4.n5.n6.n7.n8(x + 1)CODE
CRITICALtest/export/test_export.py13410 x = self.n4.n5.n6.n7.n8.n9(x + 1)CODE
CRITICALtest/export/test_export.py13411 self.n4.n5.n6.buf.add_(1)CODE
CRITICALtest/export/test_export.py13412 self.n4.n5.n6.n7.n8.buf.add_(1)CODE
CRITICALtest/export/test_export.py13429 x = self.n3.n4.n5.n6.n7.n8(x + 1)CODE
CRITICALtest/export/test_export.py13430 x = self.n3.n4.n5.n6.n7.n8.n9(x + 1)CODE
72 more matches not shown…
Modern Structural Boilerplate1172 hits · 1116 pts
SeverityFileLineSnippetContext
LOWtools/packaging/build_wheel.py19logger = logging.getLogger(__name__)CODE
LOWtools/experimental/torchfuzz/__init__.py17__all__ = [CODE
LOWtools/experimental/torchfuzz/operators/registry.py116def set_operator_weight(op_name: str, weight: float) -> None:CODE
LOWtools/experimental/torchfuzz/operators/registry.py142def set_operator_weights(weights: dict[str, float]) -> None:CODE
LOWtools/experimental/torchfuzz/operators/registry.py148def set_operator_weight_by_torch_op(torch_op_name: str, weight: float) -> None:CODE
LOWtools/experimental/torchfuzz/operators/registry.py159def set_operator_weights_by_torch_op(weights: dict[str, float]) -> None:CODE
LOWtools/experimental/torchfuzz/operators/__init__.py62__all__ = [CODE
LOWtools/test/test_cmake.py101def set_env_var(key: str, value: str | None) -> None:CODE
LOWtools/linter/adapters/_linter/__init__.py13__all__ = [CODE
LOWtools/linter/clang_tidy/generate_build_files.py25def update_submodules() -> None:CODE
LOW…s/testing/target_determination/heuristics/interface.py78 def set_test_score(self, test_run: TestRun, new_score: float) -> None:CODE
LOWtools/code_coverage/package/tool/summarize_jsons.py189def update_set() -> None:CODE
LOW.ci/lumen_cli/cli/run.py11logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/test_cli/register_test.py11logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/lib/core/vllm/lib.py24logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/lib/core/vllm/vllm_test.py25logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/lib/core/vllm/vllm_build.py33logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/lib/core/torchtitan/lib.py11logger = logging.getLogger(__name__)CODE
LOW…i/lumen_cli/cli/lib/core/torchtitan/torchtitan_test.py16logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/lib/common/docker_helper.py13logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/lib/common/gh_summary.py17logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/lib/common/utils.py16logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/lib/common/path_helper.py10logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/lib/common/pip_helper.py14logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/lib/common/git_helper.py12logger = logging.getLogger(__name__)CODE
LOW.ci/lumen_cli/cli/build_cli/register_build.py10logger = logging.getLogger(__name__)CODE
LOW.ci/manywheel/repair_wheel.py258def set_rpath(sofile: Path, rpath: str, force_rpath: bool) -> None:CODE
LOWtest/test_fx_passes.py21logger = logging.getLogger(__name__)CODE
LOWtest/inductor/test_aoti_cross_compile_windows.py51 def set_base_path(cls, path: Path | str | None = None) -> None:CODE
LOWtest/inductor/test_aoti_cross_compile_windows.py56 def set_win_torch_libs_path(cls, path: str | None = None) -> None:CODE
LOW…inductor/extension_backends/triton/device_interface.py57 def set_device(device: int) -> None:CODE
LOW…inductor/extension_backends/triton/device_interface.py77 def set_device(device) -> None:CODE
LOW…ension/torch_openreg/torch_openreg/openreg/__init__.py45def set_device(device) -> None:STRING
LOW…ension/torch_openreg/torch_openreg/openreg/__init__.py77__all__ = [STRING
LOW…xtension/torch_openreg/torch_openreg/openreg/random.py7__all__ = [CODE
LOW…torch_agn_2_13_extension/libtorch_agn_2_13/__init__.py30__all__ = [CODE
LOW…ibtorch_agn_2_9_extension/libtorch_agn_2_9/__init__.py22__all__ = [CODE
LOW…torch_agn_2_12_extension/libtorch_agn_2_12/__init__.py22__all__ = [CODE
LOW…torch_agn_2_11_extension/libtorch_agn_2_11/__init__.py22__all__ = [CODE
LOW…torch_agn_2_10_extension/libtorch_agn_2_10/__init__.py22__all__ = [CODE
LOWtest/distributed/test_c10d_functional_native.py722 def _set_group_name(self, name: str) -> None:CODE
LOWtest/distributed/checkpoint/test_pg_transport.py42logger = logging.getLogger(__name__)CODE
LOWtest/distributed/checkpoint/test_dtensor_checkpoint.py66 def set_extra_state(self, state: dict[str, int | torch._tensor.Tensor]) -> None:CODE
LOWtest/distributed/checkpoint/test_dtensor_resharding.py39logger = logging.getLogger(__name__)CODE
LOWtest/distributed/checkpoint/test_checkpoint.py215 def set_up_storage_reader(self, metadata: Metadata, is_coordinator: bool) -> None:CODE
LOWtest/distributed/elastic/test_control_plane.py201 def set_content(self, content: str, content_type: str) -> None:CODE
LOWtest/distributed/elastic/test_control_plane.py205 def set_status(self, status: int) -> None:CODE
LOW…tributed/elastic/rendezvous/dynamic_rendezvous_test.py215 def set_state_internal(self, state: _RendezvousState) -> None:CODE
LOWtest/distributed/elastic/rendezvous/api_test.py216 def set_closed(self) -> None:CODE
LOWtest/distributed/elastic/utils/util_test.py26 def set_timeout(self, timeout: float) -> None:CODE
LOWtest/distributed/pipelining/test_schedule.py72logger = logging.getLogger(__name__)CODE
LOWtest/distributed/pipelining/test_schedule_multiproc.py62logger = logging.getLogger(__name__)CODE
LOW…/distributed/_composable/fsdp/test_fully_shard_comm.py984 def set_forward_prefetch(model: Transformer, num_to_prefetch: int) -> None:CODE
LOW…/distributed/_composable/fsdp/test_fully_shard_comm.py1079 def set_backward_prefetch(model: Transformer, num_to_prefetch: int) -> None:CODE
LOW…/distributed/_composable/fsdp/test_fully_shard_comm.py1205 def set_backward_prefetch(model: Transformer) -> None:CODE
LOW…t/distributed/tensor/test_dtensor_dispatch_overhead.py22logger = logging.getLogger(__name__)CODE
LOWtest/dynamo/test_reorder_logs.py19logger = logging.getLogger(__name__)CODE
LOWtest/jit/fixtures_srcs/generate_models.py16logger = logging.getLogger(__name__)CODE
LOWtest/ao/sparsity/test_kernels.py31logger = logging.getLogger(__name__)CODE
LOWtest/profiler/test_cpp_thread.py115 def set_trace(self, trace_obj) -> None:CODE
1112 more matches not shown…
Docstring Block Structure203 hits · 1000 pts
SeverityFileLineSnippetContext
HIGHtest/test_transformers.py127 Compare test tensor against golden and reference tensors. Golden is the highest precision possible serving as tSTRING
HIGHtest/onnx/onnx_test_common.py163Run ORT on the given ONNX model and inputs Used in test_fx_to_onnx_with_onnxruntime.py Args: onnx_modeSTRING
HIGHtest/onnx/torchlib/ops_test_common.py327Run a model with ONNX Runtime in a separate process. Args: serialized_model: Serialized ONNX model proto. STRING
HIGH.claude/skills/docstring/SKILL.md295 Sample from the Gumbel-Softmax distribution and optionally discretize. Args: logits (Tensor): `[..., nSTRING
HIGHtorch/library.py1707Returns the computed kernel for a given operator and dispatch key. This function retrieves the kernel that would beSTRING
HIGHtorch/_torch_docs.py4512 unflatten(input, dim, sizes) -> Tensor Expands a dimension of the input tensor over multiple dimensions. .. seealso::STRING
HIGHtorch/_torch_docs.py6410 lu_unpack(LU_data, LU_pivots, unpack_data=True, unpack_pivots=True, *, out=None) -> (Tensor, Tensor, Tensor) Unpacks tSTRING
HIGHtorch/_torch_docs.py7381 aminmax(input, *, dim=None, keepdim=False, out=None) -> (Tensor min, Tensor max) Computes the minimum and maximum valuSTRING
HIGHtorch/_torch_docs.py11499 triangular_solve(b, A, upper=True, transpose=False, unitriangular=False, *, out=None) -> (Tensor, Tensor) Solves a sysSTRING
HIGHtorch/functional.py80broadcast_shapes(*shapes) -> Size Similar to :func:`broadcast_tensors` but for shapes. This is equivalent to STRING
HIGHtorch/functional.py176einsum(equation, *operands) -> Tensor Sums the product of the elements of the input :attr:`operands` along dimensioSTRING
HIGHtorch/_lobpcg.py359Find the k largest (or smallest) eigenvalues and the corresponding eigenvectors of a symmetric positive definite genSTRING
HIGHtorch/serialization.py714 Restores `storage` using a deserializer function registered for the `location`. This function looks in the regSTRING
HIGHtorch/_functorch/einops.py16A native implementation of `einops.rearrange`, a reader-friendly smart element reordering for multidimensional tensoSTRING
HIGHtorch/_functorch/_aot_autograd/fx_utils.py47 Given a joint graph with descriptors (meta['desc'] on placeholders and output), returns the node for every inpuSTRING
HIGHtorch/_functorch/_aot_autograd/fx_utils.py103Get all output nodes and their corresponding tangent nodes from a joint graph. Similar to get_all_input_and_grad_noSTRING
HIGHtorch/_functorch/_aot_autograd/fx_utils.py231Get parameter nodes mapped by their fully qualified names. This function traverses the graph to find all parameter STRING
HIGHtorch/_functorch/_aot_autograd/fx_utils.py259Get buffer nodes mapped by their fully qualified names. This function traverses the graph to find all buffer input STRING
HIGHtorch/_functorch/_aot_autograd/fx_utils.py287Get all parameter nodes from a graph as a list. You can rely on this providing the correct order of parameters you STRING
HIGHtorch/_functorch/_aot_autograd/fx_utils.py308Get all buffer nodes from a graph as a list. You can rely on this providing the correct order of buffers you need STRING
HIGHtorch/linalg/__init__.py97 .. seealso:: :func:`torch.linalg.cholesky_ex` for a version of this operation that skips the (slow) eSTRING
HIGHtorch/linalg/__init__.py243 .. note:: Consider using :func:`torch.linalg.solve` if possible for multiplying a matrix on the left by the inSTRING
HIGHtorch/linalg/__init__.py308 Args: A (Tensor): tensor of shape `(*, n, n)` where `*` is zero or more batch dimensions. Keyword args: left STRING
HIGHtorch/linalg/__init__.py426 linalg.slogdet(A, *, out=None) -> (Tensor, Tensor) Computes the sign and natural logarithm of the absolute value of thSTRING
HIGHtorch/linalg/__init__.py504 .. warning:: This function assumes that :attr:`A` is `diagonalizable`_ (for example, when all the eigenvaSTRING
HIGHtorch/linalg/__init__.py615 .. seealso:: :func:`torch.linalg.eig` computes the full eigenvalue decomposition. Args: A (Tensor): tensSTRING
HIGHtorch/linalg/__init__.py675 .. note:: The eigenvalues of real symmetric or complex Hermitian matrices are always real. .. warning:: The eigenvectSTRING
HIGHtorch/linalg/__init__.py798 .. seealso:: :func:`torch.linalg.eigh` computes the full eigenvalue decomposition. Args: A (Tensor): tenSTRING
HIGHtorch/linalg/__init__.py838 householder_product(A, tau, *, out=None) -> Tensor Computes the first `n` columns of a product of Householder matricesSTRING
HIGHtorch/linalg/__init__.py939 Args: A (Tensor): tensor of shape `(*, n, n)` where `*` is zero or more batch dimensions consistinSTRING
HIGHtorch/linalg/__init__.py993 Args: A (Tensor): tensor of shape `(*, n, n)` where `*` is zero or more batch dimensions consistinSTRING
HIGHtorch/linalg/__init__.py1080 torch.linalg.lstsq(A, B, rcond=None, *, driver=None) -> (Tensor, Tensor, Tensor, Tensor) Computes a solution to the leSTRING
HIGHtorch/linalg/__init__.py1204 matrix_power(A, n, *, out=None) -> Tensor Computes the `n`-th power of a square matrix for an integer `n`. Supports iSTRING
HIGHtorch/linalg/__init__.py1355 linalg.norm(input, ord=None, dim=None, keepdim=False, *, out=None, dtype=None) -> Tensor Computes a vector or matrix nSTRING
HIGHtorch/linalg/__init__.py1494 linalg.vector_norm(x, ord=2, dim=None, keepdim=False, *, dtype=None, out=None) -> Tensor Computes a vector norm. If :STRING
HIGHtorch/linalg/__init__.py1576 linalg.matrix_norm(A, ord='fro', dim=(-2, -1), keepdim=False, *, dtype=None, out=None) -> Tensor Computes a matrix norSTRING
HIGHtorch/linalg/__init__.py1741 linalg.svd(A, full_matrices=True, *, driver=None, out=None) -> (Tensor, Tensor, Tensor) Computes the singular value deSTRING
HIGHtorch/linalg/__init__.py1912 .. seealso:: :func:`torch.linalg.svd` computes the full singular value decomposition. Args: A (Tensor): STRING
HIGHtorch/linalg/__init__.py1945 linalg.cond(A, p=None, *, out=None) -> Tensor Computes the condition number of a matrix with respect to a matrix norm.STRING
HIGHtorch/linalg/__init__.py2315 .. seealso:: :func:`torch.linalg.solve_triangular` computes the solution of a triangular system of linear STRING
HIGHtorch/linalg/__init__.py2479 .. warning:: The LU decomposition is almost never unique, as often there are different permutation matriceSTRING
HIGHtorch/linalg/__init__.py2648 lu(A, *, pivot=True, out=None) -> (Tensor, Tensor, Tensor) Computes the LU decomposition with partial pivoting of a maSTRING
HIGHtorch/linalg/__init__.py2738 linalg.tensorinv(A, ind=2, *, out=None) -> Tensor Computes the multiplicative inverse of :func:`torch.tensordot`. If STRING
HIGHtorch/linalg/__init__.py2806 linalg.tensorsolve(A, B, dims=None, *, out=None) -> Tensor Computes the solution `X` to the system `torch.tensordot(A,STRING
HIGHtorch/linalg/__init__.py2872 qr(A, mode='reduced', *, out=None) -> (Tensor, Tensor) Computes the QR decomposition of a matrix. Letting :math:`\matSTRING
HIGHtorch/linalg/__init__.py3020 Args: A (Tensor): tensor of shape `(*, m, n)` with `m >= n`, where `*` is zero or more batch dimensSTRING
HIGHtorch/nn/functional.py2226 Sample from the Gumbel-Softmax distribution (`Link 1`_ `Link 2`_) and optionally discretize. Args: logiSTRING
HIGHtorch/nn/functional.py6484 Args: query (Tensor): Query tensor; shape :math:`(N, ..., Hq, L, E)`. key (Tensor): Key tensor; shaSTRING
HIGHtorch/nn/attention/bias.py192 Handles the logic for computing attention with the specified causal bias. Args: query (TenSTRING
HIGHtorch/nn/attention/flex_attention.py1034 Creates a BlockMask instance from key-value block information. Args: kv_num_blocks (TensorSTRING
HIGHtorch/nn/attention/flex_attention.py1205 Returns a new BlockMask by selecting batch, head, and Q-block rows. BlockMask indexing accepts up to tSTRING
HIGHtorch/nn/parallel/distributed.py71Factory method to create a BucketCapacityConfig from user inputs. Args: bucket_cap_mb: Single buckeSTRING
HIGHtorch/nn/parallel/distributed.py133Compute bucket size limits for initial bucketing and rebuilding. Args: static_graph: Whether the coSTRING
HIGHtorch/nn/utils/stateless.py18 Unties all tied tensors in the module to parameters_and_buffers. This function returns a new untied_parametersSTRING
HIGHtorch/nn/utils/parametrize.py457Register a parametrization to a tensor in a module. Assume that ``tensor_name="weight"`` for simplicity. When accesSTRING
HIGHtorch/nn/utils/parametrize.py701Remove the parametrizations on a tensor in a module. - If ``leave_parametrized=True``, ``module[tensor_name]`` willSTRING
HIGHtorch/nn/utils/weight_norm.py85Apply weight normalization to a parameter in the given module. .. math:: \mathbf{w} = g \dfrac{\mathbf{v}}STRING
HIGHtorch/nn/utils/prune.py590Compute and returns a mask for the input tensor ``t``. Starting from a base ``default_mask`` (which should be aSTRING
HIGHtorch/nn/utils/prune.py699Compute and returns a mask for the input tensor ``t``. Starting from a base ``default_mask`` (which should be aSTRING
HIGHtorch/nn/utils/prune.py837Apply pruning reparameterization without pruning any units. Applies pruning reparameterization to the tensor corresSTRING
143 more matches not shown…
Cross-File Repetition170 hits · 850 pts
SeverityFileLineSnippetContext
HIGHtools/linter/adapters/clangtidy_linter.py0(?mx) ^ (?p<file>.*?): (?p<line>\d+): (?:(?p<column>-?\d+):)? \s(?p<severity>\s+?):? \s(?p<message>.*) \s(?p<code>\[.*\]STRING
HIGHtools/linter/adapters/mypy_linter.py0(?mx) ^ (?p<file>.*?): (?p<line>\d+): (?:(?p<column>-?\d+):)? \s(?p<severity>\s+?):? \s(?p<message>.*) \s(?p<code>\[.*\]STRING
HIGHtools/linter/adapters/pyrefly_linter.py0(?mx) ^ (?p<file>.*?): (?p<line>\d+): (?:(?p<column>-?\d+):)? \s(?p<severity>\s+?):? \s(?p<message>.*) \s(?p<code>\[.*\]STRING
HIGH.ci/pytorch/test_example_code/cnn_smoke_win_arm64.py0it's used to check basic rnn features with cuda. for example, it would throw exception if missing some components are miSTRING
HIGH.ci/pytorch/test_example_code/rnn_smoke_win_arm64.py0it's used to check basic rnn features with cuda. for example, it would throw exception if missing some components are miSTRING
HIGH.ci/pytorch/test_example_code/cnn_smoke.py0it's used to check basic rnn features with cuda. for example, it would throw exception if missing some components are miSTRING
HIGH.ci/pytorch/test_example_code/rnn_smoke.py0it's used to check basic rnn features with cuda. for example, it would throw exception if missing some components are miSTRING
HIGHtest/test_transformers.py0clones the query, key, and value tensors and moves them to the specified dtype.STRING
HIGHtest/inductor/test_flex_attention.py0clones the query, key, and value tensors and moves them to the specified dtype.STRING
HIGHtest/inductor/test_flex_decoding.py0clones the query, key, and value tensors and moves them to the specified dtype.STRING
HIGHbenchmarks/transformer/score_mod.py0clones the query, key, and value tensors and moves them to the specified dtype.STRING
HIGHtest/test_dynamic_spec_make_fx.py0reset the global intvar uid counter so uids start at 0 per test.STRING
HIGHtest/dynamo/test_dynamic_spec.py0reset the global intvar uid counter so uids start at 0 per test.STRING
HIGHtest/export/test_dynamic_spec_export.py0reset the global intvar uid counter so uids start at 0 per test.STRING
HIGHtest/inductor/test_codecache.py0verify that we can populate and hot load functions from the cache.STRING
HIGHtest/inductor/test_config.py0verify that we can populate and hot load functions from the cache.STRING
HIGHtest/dynamo/test_aot_autograd_cache.py0verify that we can populate and hot load functions from the cache.STRING
HIGH…ension/torch_openreg/torch_openreg/openreg/__init__.py0context-manager that changes the selected device. args: device (torch.device or int or str): device index to select. it'STRING
HIGHtorch/cuda/__init__.py0context-manager that changes the selected device. args: device (torch.device or int or str): device index to select. it'STRING
HIGHtorch/mtia/__init__.py0context-manager that changes the selected device. args: device (torch.device or int or str): device index to select. it'STRING
HIGHtorch/xpu/__init__.py0context-manager that changes the selected device. args: device (torch.device or int or str): device index to select. it'STRING
HIGHtest/nn/attention/test_fa3.py0test that torch.compile preserves tensor metadata (shape, stride, dtype).STRING
HIGHtest/nn/attention/_fa_test_common.py0test that torch.compile preserves tensor metadata (shape, stride, dtype).STRING
HIGHtest/nn/attention/test_fa4.py0test that torch.compile preserves tensor metadata (shape, stride, dtype).STRING
HIGHtest/distributed/test_c10d_ucc.py0this unit test verifies whether the future object is passed properly using nccl backend. the hook callback function creaSTRING
HIGHtest/distributed/test_c10d_gloo.py0this unit test verifies whether the future object is passed properly using nccl backend. the hook callback function creaSTRING
HIGHtest/distributed/test_c10d_nccl.py0this unit test verifies whether the future object is passed properly using nccl backend. the hook callback function creaSTRING
HIGHtest/dynamo/test_aot_compile.py0\ <internal traceback> torch._dynamo.exc.unsupported: call to `torch._dynamo.graph_break()` explanation: user-inserted gSTRING
HIGHtest/dynamo/test_error_messages.py0\ <internal traceback> torch._dynamo.exc.unsupported: call to `torch._dynamo.graph_break()` explanation: user-inserted gSTRING
HIGHtest/dynamo/test_exc.py0\ <internal traceback> torch._dynamo.exc.unsupported: call to `torch._dynamo.graph_break()` explanation: user-inserted gSTRING
HIGHtest/quantization/eager/test_quantize_eager_qat.py0create a qat module from a float module args: `mod` a float module, either produced by torch.ao.quantization utilities oSTRING
HIGHtorch/ao/nn/qat/modules/linear.py0create a qat module from a float module args: `mod` a float module, either produced by torch.ao.quantization utilities oSTRING
HIGHtorch/ao/nn/intrinsic/qat/modules/conv_fused.py0create a qat module from a float module args: `mod` a float module, either produced by torch.ao.quantization utilities oSTRING
HIGHtorch/ao/nn/qat/modules/conv.py0create a qat module from a float module args: `mod` a float module, either produced by torch.ao.quantization utilities oSTRING
HIGHtorch/ao/nn/qat/modules/embedding_ops.py0create a qat module from a float module args: `mod` a float module, either produced by torch.ao.quantization utilities oSTRING
HIGHtest/cpython/v3_13/test_cmath.py0assert that floats x and y are identical, in the sense that: (1) both x and y are nans, or (2) both x and y are infinitiSTRING
HIGHtest/cpython/v3_13/test_complex.py0assert that floats x and y are identical, in the sense that: (1) both x and y are nans, or (2) both x and y are infinitiSTRING
HIGHtest/cpython/v3_13/test_float.py0assert that floats x and y are identical, in the sense that: (1) both x and y are nans, or (2) both x and y are infinitiSTRING
HIGHtest/jit/mydecorator.py0helper function used in test_decorator.py. we define it in a separate file on purpose to test that the names in differenSTRING
HIGHtest/jit/myfunction_b.py0helper function used in test_decorator.py. we define it in a separate file on purpose to test that the names in differenSTRING
HIGHtest/jit/myfunction_a.py0helper function used in test_decorator.py. we define it in a separate file on purpose to test that the names in differenSTRING
HIGHtest/functorch/test_rearrange.py0adapted from https://github.com/arogozhnikov/einops/blob/36c7bb16e57d6e57f8f3050f9e07abdf3f00469f/einops/parsing.py. mitSTRING
HIGHtest/functorch/test_parsing.py0adapted from https://github.com/arogozhnikov/einops/blob/36c7bb16e57d6e57f8f3050f9e07abdf3f00469f/einops/parsing.py. mitSTRING
HIGHtorch/_appdirs.py0adapted from https://github.com/arogozhnikov/einops/blob/36c7bb16e57d6e57f8f3050f9e07abdf3f00469f/einops/parsing.py. mitSTRING
HIGHfunctorch/einops/_parsing.py0adapted from https://github.com/arogozhnikov/einops/blob/36c7bb16e57d6e57f8f3050f9e07abdf3f00469f/einops/parsing.py. mitSTRING
HIGHtest/functorch/test_codegen_mutation_epilogue.py0capture codegen artifacts from the structured trace log.STRING
HIGHtest/functorch/test_codegen_backward_epilogue.py0capture codegen artifacts from the structured trace log.STRING
HIGHtest/functorch/test_codegen_debug_assert.py0capture codegen artifacts from the structured trace log.STRING
HIGHtest/functorch/test_codegen_output_alias.py0capture codegen artifacts from the structured trace log.STRING
HIGHtest/functorch/test_codegen_dedup.py0capture codegen artifacts from the structured trace log.STRING
HIGHtorch/storage.py0return a copy of this object in custom device memory. if this object is already in device memory and on the correct deviSTRING
HIGHtorch/_utils.py0return a copy of this object in custom device memory. if this object is already in device memory and on the correct deviSTRING
HIGHtorch/utils/backend_registration.py0return a copy of this object in custom device memory. if this object is already in device memory and on the correct deviSTRING
HIGHtorch/_export/passes/replace_with_hop_pass_util.py0split gm into sub-graph-modules using `sequential_split_and_maybe_inline_subgraphs`, and then recursively call itself onSTRING
HIGHtorch/_export/passes/replace_autocast_with_hop_pass.py0split gm into sub-graph-modules using `sequential_split_and_maybe_inline_subgraphs`, and then recursively call itself onSTRING
HIGHtorch/_export/passes/replace_set_grad_with_hop_pass.py0split gm into sub-graph-modules using `sequential_split_and_maybe_inline_subgraphs`, and then recursively call itself onSTRING
HIGHtorch/nn/qat/dynamic/modules/linear.py0intrinsic qat modules. this file is in the process of migration to `torch/ao/nn/intrinsic/qat`, and is kept here for comSTRING
HIGHtorch/nn/qat/modules/linear.py0intrinsic qat modules. this file is in the process of migration to `torch/ao/nn/intrinsic/qat`, and is kept here for comSTRING
HIGHtorch/nn/qat/modules/conv.py0intrinsic qat modules. this file is in the process of migration to `torch/ao/nn/intrinsic/qat`, and is kept here for comSTRING
HIGHtorch/nn/qat/modules/embedding_ops.py0intrinsic qat modules. this file is in the process of migration to `torch/ao/nn/intrinsic/qat`, and is kept here for comSTRING
110 more matches not shown…
Redundant / Tautological Comments489 hits · 714 pts
SeverityFileLineSnippetContext
LOWtools/nightly_hotpatch.py168 # Check if the patch was applied successfullyCOMMENT
LOWtools/packaging/build_wheel.py85 # Check if it's PyPy (skip it)COMMENT
LOWtools/packaging/build_wheel.py103 # Check if this version is supportedCOMMENT
LOWtools/experimental/torchfuzz/ops_fuzzer.py71 # Check if the operator supports any of the allowed operationsCOMMENT
LOWtools/experimental/torchfuzz/multi_process_fuzzer.py167 # Check if output should be ignored and which pattern matchedCOMMENT
LOWtools/experimental/torchfuzz/operators/layout.py333 # Check if there's at least one singleton dimensionCOMMENT
LOWtools/linter/adapters/grep_linter.py201 # Check if file is too large to compute replacementCOMMENT
LOWtools/linter/adapters/stable_shim_version_linter.py114 # Check if this is the AOTI shim - only enforce versioning on new linesCOMMENT
LOWtools/linter/adapters/stable_shim_version_linter.py163 # Check if this is a newly added lineCOMMENT
LOWtools/linter/adapters/_stable_shim_utils.py346 # Check if we're entering a block commentCOMMENT
LOWtools/linter/adapters/_stable_shim_utils.py411 # Check if this #elif has a version conditionCOMMENT
LOWtools/linter/adapters/s3_init.py42 # Open the file in binary mode and hash it.COMMENT
LOWtools/linter/adapters/update_s3.py25 # Open the file in binary mode and hash it.COMMENT
LOWtools/linter/adapters/newlines_linter.py44 # Check if file is too largeCOMMENT
LOWtools/linter/adapters/codespell_linter.py98 # Check if file is too largeCOMMENT
LOWtools/amd_build/build_amd.py150# Check if the compiler is hip-clang.COMMENT
LOWtools/testing/target_determination/heuristics/utils.py174 # Check if the file extension is in the docs-only setCOMMENT
LOW…ols/lite_interpreter/gen_selected_mobile_ops_header.py87# Write the file selected_mobile_ops.h with optionally:STRING
LOW…ols/lite_interpreter/gen_selected_mobile_ops_header.py126# Write the file selected_mobile_ops.h with optionally:STRING
LOWtools/vendoring/quack/vendor.sh127# Set UPSTREAM_DIR to a quack checkout at $sha. With a local checkout, validateCOMMENT
LOWtools/stats/import_test_stats.py57 # Check if the file is new enough (see: FILE_CACHE_LIFESPAN_SECONDS). A real checkCOMMENT
LOWtools/stats/check_disabled_tests.py77 # Check if the test is a failureCOMMENT
LOW.ci/docker/common/install_torch_tpu.sh56 # Check if xtrace (set -x) is enabledCOMMENT
LOW.ci/pytorch/common_utils.sh63 # Check if the first argument contains multiple paths separated by spacesCOMMENT
LOW.ci/pytorch/common_utils.sh200 # Check if the wheel has been already been builtCOMMENT
LOW.ci/pytorch/build.sh138 # Set ROCM_ARCH to gfx906 for CI builds, if user doesn't override.COMMENT
LOW.ci/pytorch/test.sh168# Set OMP_NUM_THREADS to nproc/4 on k8s ARC runners if not already set.COMMENT
LOW.ci/pytorch/test.sh1314 # Check if regression was consistent (majority of runs)COMMENT
LOW.ci/pytorch/smoke_test/smoke_test.py529 # Check if NVSHMEM is compiled in current buildCOMMENT
LOW.ci/pytorch/smoke_test/smoke_test.py545 # Check if NVSHMEM is available on current system.COMMENT
LOW.ci/pytorch/smoke_test/check_binary_symbols.py418 # Check if the stable C shim existsSTRING
LOWtest/test_linalg.py9360 # Set filenames to be unique on a per test basisSTRING
LOWtest/test_reductions.py3697 # Check if reduction happens along the specified dim with and without keepdim. Check withCOMMENT
LOWtest/test_reductions.py3718 # Check if function raises error on specified zero'd dimension as reduction dim.COMMENT
LOWtest/test_reductions.py3756 # Check if function raises error on specified zero'd dimension as reduction dim.COMMENT
LOWtest/test_reductions.py3783 # Check if reduction happens along the specified dimension.COMMENT
LOWtest/test_bundled_images.py90 # Check if fb::image_decode_to_NCHW works as expectedCOMMENT
LOWtest/test_jit.py12881 # Check if the peak sizes at most differ by an empirically obtained factorSTRING
LOWtest/test_view_ops.py1097 # Check if result is a view of x by comparing data pointersCOMMENT
LOWtest/test_view_ops.py1099 # Check if ref is a view of x by comparing data pointersCOMMENT
LOWtest/test_custom_ops.py333 # Check if ret is a tuple and has exactly one and the same elementCOMMENT
LOWtest/test_meta.py417 # Check if it's a view, by testing if any of the returns haveCOMMENT
LOWtest/test_fx_experimental.py920 # Assign ops to different partitions so a single-output submodule exists.STRING
LOWtest/test_modules.py99 # Check if this module creates parameters or registers buffers.COMMENT
LOWtest/test_modules.py110 # Check if a parameter or buffer was created with a tensor not passed to the constructor.COMMENT
LOWtest/test_modules.py268 # Check if the inplace variant of the module gives the same result as the out of placeCOMMENT
LOWtest/test_transformers.py2759 # Check if masked rows are zero in outputCOMMENT
LOWtest/test_transformers.py2766 # Check if gradients for masked rows are zeroCOMMENT
LOWtest/inductor/test_flex_attention.py6916 # Check if kernel code contains warp specialization parametersCOMMENT
LOWtest/inductor/test_max_autotune.py5902 # Set env to make it work in CI.COMMENT
LOWtest/inductor/test_cutlass_fallback.py18# Check if CUTLASS is availableCOMMENT
LOWtest/inductor/test_cooperative_reductions.py216 # Iterate through elements for comparisonCOMMENT
LOWtest/inductor/test_max_autotune_blackwell.py723 # Set env to make it work in CI.COMMENT
LOWtest/inductor/test_pad_mm.py634 # Check if any bmm operations had dtype changesCOMMENT
LOWtest/inductor/test_ck_backend.py503 # Set env to make it work in CI.COMMENT
LOWtest/inductor/test_subgraph_choice.py174 # Set env to make it work in CI.COMMENT
LOWtest/inductor/test_foreach.py1073 # Set requires_grad to be False to avoid mutating a leaf variableCOMMENT
LOWtest/inductor/test_op_dtype_prop.py184 # Check if the op's output should be upcasted/downcasted.COMMENT
LOWtest/inductor/test_aot_inductor_package.py422 # Check if the .so file was build successfullyCOMMENT
LOWtest/inductor/test_aot_inductor_package.py465 # Check if the .so file was build successfullyCOMMENT
429 more matches not shown…
AI Structural Patterns845 hits · 620 pts
SeverityFileLineSnippetContext
LOWtools/testing/test_run.py135CODE
LOW.ci/pytorch/smoke_test/smoke_test.py75CODE
LOW.ci/pytorch/smoke_test/max_autotune.py35CODE
LOW.ci/pytorch/test_example_code/cnn_smoke_win_arm64.py21CODE
LOW.ci/pytorch/test_example_code/cnn_smoke.py21CODE
LOWtest/test_mps.py14904CODE
LOWtest/test_mps.py11157CODE
LOWtest/test_linalg.py4526CODE
LOWtest/test_utils.py133CODE
LOWtest/test_fx_passes.py155CODE
LOWtest/test_fx_passes.py167CODE
LOWtest/test_fx_passes.py179CODE
LOWtest/test_fx_passes.py207CODE
LOWtest/test_fx_passes.py806CODE
LOWtest/test_fx_passes.py814CODE
LOWtest/test_fx_passes.py933CODE
LOWtest/test_fx_passes.py1116CODE
LOWtest/test_static_runtime.py20CODE
LOWtest/test_functionalization.py486CODE
LOWtest/test_functionalization.py1176CODE
LOWtest/test_decomp.py1592CODE
LOWtest/test_mobile_optimizer.py321CODE
LOWtest/test_xnnpack_integration.py1200CODE
LOWtest/test_xnnpack_integration.py1215CODE
LOWtest/test_torch.py3457CODE
LOWtest/test_dynamic_shapes.py698CODE
LOWtest/test_dynamic_shapes.py2291CODE
LOWtest/test_dynamic_shapes.py4579CODE
LOWtest/test_jit_fuser_te.py2693CODE
LOWtest/test_nn.py3136CODE
LOWtest/test_nn.py7383CODE
LOWtest/test_jit.py697CODE
LOWtest/test_jit.py1240CODE
LOWtest/test_jit.py6582CODE
LOWtest/test_jit.py6917CODE
LOWtest/test_jit.py11757CODE
LOWtest/test_jit.py11732CODE
LOWtest/test_jit.py11746CODE
LOWtest/test_jit.py15988CODE
LOWtest/test_jit.py1857CODE
LOWtest/test_custom_ops.py1226CODE
LOWtest/test_custom_ops.py4395CODE
LOWtest/test_torchfuzz_repros.py52CODE
LOWtest/test_torchfuzz_repros.py93CODE
LOWtest/test_torchfuzz_repros.py138CODE
LOWtest/test_torchfuzz_repros.py181CODE
LOWtest/test_torchfuzz_repros.py230CODE
LOWtest/test_torchfuzz_repros.py264CODE
LOWtest/run_test.py1233CODE
LOWtest/test_opaque_obj_v2.py385CODE
LOWtest/test_opaque_obj_v2.py2925CODE
LOWtest/test_opaque_obj_v2.py3287CODE
LOWtest/test_autograd.py7834CODE
LOWtest/test_autograd.py8034CODE
LOWtest/test_autograd.py8750CODE
LOWtest/test_autograd.py13607CODE
LOWtest/test_autograd.py16209CODE
LOWtest/test_autograd.py16223CODE
LOWtest/test_autograd.py16234CODE
LOWtest/test_autograd.py15832CODE
785 more matches not shown…
Structural Annotation Overuse369 hits · 619 pts
SeverityFileLineSnippetContext
LOWtools/experimental/torchfuzz/README.md528### Step 1: Create Operator ClassCOMMENT
LOWtools/experimental/torchfuzz/README.md564### Step 2: Register Your ModuleCOMMENT
LOWtools/experimental/torchfuzz/README.md581### Step 3: Add to Template (Optional)COMMENT
LOWtools/experimental/torchfuzz/README.md598### Step 4: Test Your OperatorCOMMENT
LOWtest/test_autograd.py8190 # Step 1: Checkpointed forward with batch_size=4 → static graph G0COMMENT
LOWtest/test_autograd.py8194 # Step 2: Non-checkpointed call with batch_size=7 triggers automaticCOMMENT
LOWtest/test_autograd.py8202 # Step 3: Backward. Checkpoint recomputes fn(x) but now hits G1COMMENT
LOWtest/cpp_extensions/cudnn_extension.cpp43 // Step 1: Check inputs. This will throw an error if inputs are invalid, so noCOMMENT
LOWtest/cpp_extensions/cudnn_extension.cpp46 // Step 2: Create descriptorsCOMMENT
LOWtest/cpp_extensions/cudnn_extension.cpp59 // Step 3: Apply CuDNN functionCOMMENT
LOWtest/cpp_extensions/cudnn_extension.cpp71 // Step 4: Destroy descriptorsCOMMENT
LOWtest/cpp_extensions/cudnn_extension.cpp73 // Step 5: Return something (optional)COMMENT
LOWtest/distributed/tensor/test_redistribute.py3087 # Step 1: reduce_scatter on mesh dim 0 (size 2)COMMENT
LOWtest/distributed/tensor/test_redistribute.py3091 # Step 2: reduce_scatter on mesh dim 1 (size 3)COMMENT
LOWtest/distributed/tensor/experimental/test_local_map.py842 # Step 1: BuggyDPMatmul without spmd_types -> wrong W gradient.COMMENT
LOWtest/distributed/tensor/experimental/test_local_map.py859 # Step 2: BuggyDPMatmul with spmd_types=True -> caught.COMMENT
LOWtest/distributed/tensor/experimental/test_local_map.py875 # Step 3: CorrectDPMatmul with spmd_types=True -> correct gradientsCOMMENT
LOWtest/distributed/tensor/parallel/test_tp_examples.py222 # Step 3: Run test by comparing outputs from single-gpu and multi-gpu models.COMMENT
LOWtest/distributed/tensor/parallel/test_tp_examples.py305 ) # Step 1: Initialize single-gpu models.CODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py308 ) # Step 2: Setup tp model, place onto device mesh.CODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py311 ) # Step 3: Setup optimizers for both modelsCODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py433 ) # Step 1: Initialize single-gpu models.CODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py436 ) # Step 2: Setup tp model, place onto device mesh.CODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py439 ) # Step 3: Setup optimizers for both modelsCODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py442 ) # Step 4: set `requires_grad` patternsCODE
LOWtest/dynamo/test_activation_offloading.py596 # Step 1: cold — allocates new pinned bufferCOMMENT
LOWtest/dynamo/test_activation_offloading.py609 # Step 2: warm — reuses pooled bufferCOMMENT
LOWtest/dynamo/test_repros.py7545 # Step 1: Compile a static shapes graphCOMMENT
LOWtest/dynamo/test_repros.py7554 # Step 2: Compile a dynamic shapes graphCOMMENT
LOWtest/dynamo/test_repros.py7564 # Step 3: Run with Step 1's inputsCOMMENT
LOWtest/dynamo/test_repros.py7587 # Step 1: Compile a static shapes graphCOMMENT
LOWtest/dynamo/test_repros.py7596 # Step 2: Compile a dynamic shapes graphCOMMENT
LOWtest/dynamo/test_repros.py7606 # Step 3: Run with Step 1's inputsCOMMENT
LOWtest/profiler/test_profiler.py4295 # Step 3: under the patch, RECORD -> DEVICE_STOPPED fires stop_trace +COMMENT
LOW.claude/skills/aoti-debug/triton-index-out-of-bounds.md31## Step 1: Collect AOTI PackageCOMMENT
LOW.claude/skills/aoti-debug/triton-index-out-of-bounds.md43## Step 2: Locate the Failing Kernel in C++ WrapperCOMMENT
LOW.claude/skills/aoti-debug/triton-index-out-of-bounds.md72## Step 3: Understand the Kernel LogicCOMMENT
LOW.claude/skills/aoti-debug/triton-index-out-of-bounds.md99## Step 4: Identify the Dynamic Shape VariableCOMMENT
LOW.claude/skills/aoti-debug/triton-index-out-of-bounds.md135## Step 5: Trace Back to Model InputCOMMENT
LOW.claude/skills/aoti-debug/triton-index-out-of-bounds.md160## Step 6: Map to Model CodeCOMMENT
LOW.claude/skills/aoti-debug/SKILL.md93### Step 1: Sanity ChecksCOMMENT
LOW.claude/skills/aoti-debug/SKILL.md107### Step 2: Pinpoint the CUDA IMACOMMENT
LOW.claude/skills/aoti-debug/SKILL.md121### Step 3: Identify Problematic Kernels with Intermediate Value DebuggerCOMMENT
LOW.claude/skills/metal-kernel/SKILL.md24## Step 1: Update native_functions.yamlCOMMENT
LOW.claude/skills/metal-kernel/SKILL.md89## Step 2: Implement Metal KernelCOMMENT
LOW.claude/skills/metal-kernel/SKILL.md236## Step 3: Implement Host-Side StubCOMMENT
LOW.claude/skills/metal-kernel/SKILL.md283## Step 4: CompileCOMMENT
LOW.claude/skills/pyrefly-type-coverage/SKILL.md14### Step 1: Remove file-level type-check suppressionsCOMMENT
LOW.claude/skills/pyrefly-type-coverage/SKILL.md26### Step 2: Add a sub-config entry to `pyrefly.toml`COMMENT
LOW.claude/skills/pyrefly-type-coverage/SKILL.md45### Step 3: Run pyreflyCOMMENT
LOW.claude/skills/pyrefly-type-coverage/SKILL.md70### Step 4: Add annotationsCOMMENT
LOW.claude/skills/pyrefly-type-coverage/SKILL.md200### Step 5: IterateCOMMENT
LOW.claude/skills/pyrefly-type-coverage/SKILL.md209### Step 6: LintCOMMENT
LOW.claude/skills/pyrefly-type-coverage/SKILL.md220### Step 7: TestCOMMENT
LOW.claude/skills/pt2-bug-basher/SKILL.md260# Step 1: Generate the minifier launcherCOMMENT
LOW.claude/skills/pt2-bug-basher/SKILL.md263# Step 2: Run the minifierCOMMENT
LOW.claude/skills/pt2-bug-basher/SKILL.md266# Step 3: Run the minimized reproCOMMENT
LOW.claude/skills/at-dispatch-v2/SKILL.md44### Step 1: Add the Dispatch_v2.h includeCOMMENT
LOW.claude/skills/at-dispatch-v2/SKILL.md55### Step 2: Identify the old dispatch patternCOMMENT
LOW.claude/skills/at-dispatch-v2/SKILL.md64### Step 3: Map the old macro to type groupsCOMMENT
309 more matches not shown…
Fake / Example Data701 hits · 596 pts
SeverityFileLineSnippetContext
LOWtest/test_fx_passes.py583 if node.op not in ("placeholder", "output")CODE
LOWtest/test_fx_passes.py589 node.name for node in fused_submodule.graph.nodes if node.op == "placeholder"CODE
LOWtest/test_fx_passes.py552 if node.op not in ("placeholder", "output")CODE
LOWtest/test_fx_passes.py558 node.name for node in fused_submodule.graph.nodes if node.op == "placeholder"CODE
LOWtest/test_fx_passes.py619 node.name for node in fused_submodule.graph.nodes if node.op == "placeholder"CODE
LOWtest/test_fx_passes.py650 node.name for node in fused_submodule.graph.nodes if node.op == "placeholder"CODE
LOWtest/test_fx_passes.py1187 if not test_case.match_placeholder and node.op == "placeholder":CODE
LOWtest/test_proxy_tensor.py1194 if node.op == "placeholder"CODE
LOWtest/test_proxy_tensor.py1236 placeholders = [node for node in gm.graph.nodes if node.op == "placeholder"]CODE
LOWtest/test_proxy_tensor.py1912 meta_inp = _get_node(fx_g, lambda x: x.op == 'placeholder')CODE
LOWtest/test_dynamic_shapes.py7097 if node.op == "placeholder":CODE
LOWtest/test_opaque_obj_v2.py4006 if n.op == "placeholder" and is_opaque_node(n)CODE
LOWtest/test_opaque_obj_v2.py4029 placeholders = [n for n in captured_gm.graph.nodes if n.op == "placeholder"]CODE
LOWtest/test_opaque_obj_v2.py4043 restored_phs = [n for n in restored_gm.graph.nodes if n.op == "placeholder"]CODE
LOWtest/test_opaque_obj_v2.py4087 placeholders = [n for n in gm.graph.nodes if n.op == "placeholder"]CODE
LOWtest/test_fx_experimental.py1384 self.assertEqual(nodes[0].op, "placeholder")CODE
LOWtest/test_fx_experimental.py345 if node.op not in {"output", "placeholder", "get_attr"}:CODE
LOWtest/test_fx_experimental.py403 if node.op not in {"output", "placeholder", "get_attr"}:CODE
LOWtest/test_fx_experimental.py468 if node.op not in {"placeholder", "get_attr", "output"}:CODE
LOWtest/test_fx_experimental.py1166 if node.op in ("placeholder", "get_attr", "output"):CODE
LOWtest/test_fx_experimental.py1221 if node.op in ("placeholder", "get_attr", "output"):CODE
LOWtest/test_fx_experimental.py1337 if node.op in ("placeholder", "get_attr", "output"):CODE
LOWtest/test_fx_experimental.py1362 if node.op in ("placeholder", "get_attr", "output"):CODE
LOWtest/test_fx_experimental.py1019 self.assertEqual(nodes[0].op, "placeholder")STRING
LOWtest/test_fx_experimental.py1594 if node.op == "placeholder":STRING
LOWtest/test_fx_experimental.py1153 elif node.op == "placeholder":STRING
LOWtest/test_fx_experimental.py1609 ("placeholder", "x"),STRING
LOWtest/test_fx_experimental.py1638 ("placeholder", "x"),STRING
LOWtest/test_dynamic_spec_make_fx.py35 if node.op != "placeholder":CODE
LOWtest/test_fx.py2198 if node.op == "placeholder":CODE
LOWtest/test_fx.py2213 node.meta.get("val") for node in gm.graph.nodes if node.op == "placeholder"CODE
LOWtest/test_fx.py2614 x: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py2624 x : torch.fx.Node = graph.create_node('placeholder', 'x')CODE
LOWtest/test_fx.py2633 x: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py907 if node.op in {"placeholder", "output"}:CODE
LOWtest/test_fx.py956 a: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py1038 if n.op == "placeholder":CODE
LOWtest/test_fx.py1103 placeholder_nodes.append(graph.create_node("placeholder", name))CODE
LOWtest/test_fx.py1455 FileCheck().check("x").check("placeholder").check("y").check(CODE
LOWtest/test_fx.py1456 "placeholder"CODE
LOWtest/test_fx.py1468 a: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py1922 a: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py2031 a: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py2082 "placeholder",CODE
LOWtest/test_fx.py2125 if node.op == "placeholder":CODE
LOWtest/test_fx.py2671 x: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py2736 x: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py2737 y: torch.fx.Node = graph.create_node("placeholder", "y")CODE
LOWtest/test_fx.py2753 x: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py2771 x: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py2772 y: torch.fx.Node = graph.create_node("placeholder", "y")CODE
LOWtest/test_fx.py2788 x: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py2789 y: torch.fx.Node = graph.create_node("placeholder", "y")CODE
LOWtest/test_fx.py2839 x: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py2857 x: torch.fx.Node = graph.create_node("placeholder", "x")CODE
LOWtest/test_fx.py2928 if not (input_node and input_node.op == "placeholder"):CODE
LOWtest/test_fx.py2930 f"Expected input_node to exist and have op='placeholder', "CODE
LOWtest/test_fx.py3285 if node.op == "placeholder":CODE
LOWtest/test_fx.py3421 if node.op == "placeholder" or node.op == "output":CODE
LOWtest/test_fx.py4239 placeholder_count = sum(i.op == "placeholder" for i in nf.graph.nodes)CODE
641 more matches not shown…
AI Slop Vocabulary306 hits · 587 pts
SeverityFileLineSnippetContext
MEDIUMsetup.py42# if used in conjunction with DEBUG or REL_WITH_DEB_INFO, will alsoCOMMENT
LOWtools/optional_submodules.py71 # Otherwise just call top-level function of choiceCOMMENT
MEDIUMtools/experimental/torchfuzz/visualize_graph.py79 # Create comprehensive labelCOMMENT
LOWtools/experimental/torchfuzz/operators/layout.py69 # For single-dim input, just use the exact element countCOMMENT
LOWtools/experimental/torchfuzz/operators/layout.py150 # For single-dim input, just use the exact element countCOMMENT
LOWtools/linter/adapters/gb_registry_linter.py210 # Empty registry, just add all entriesCOMMENT
MEDIUMtools/autograd/gen_python_functions.py596 # find and leverage the original ATen signature (to which it delegatesCOMMENT
MEDIUMtools/autograd/gen_autograd_functions.py196# `grads` in order to utilize and apply the existing derivative definitionsCOMMENT
MEDIUMtools/autograd/gen_autograd_functions.py684 # note(crcrpar): [nuanced return type of out-of-place foreach functions]STRING
MEDIUMtools/autograd/gen_variable_type.py426# pertaining to its output:COMMENT
MEDIUMtools/autograd/gen_variable_type.py1544 # See Note [nuanced return type of out-of-place foreach functions]COMMENT
LOWtools/autograd/gen_variable_type.py2079 # View ops create fw_grad that already is a view of the base's fw_grad so just use thatCOMMENT
LOWtools/autograd/gen_trace_type.py180 # For inplace: just add the input to the end to confirm with the JIT schemaCOMMENT
MEDIUMtools/alerts/create_alerts.py182# Creates a Dict of Job Name -> [JobData]. Essentially a Column in HUDSTRING
LOWtools/code_coverage/package/oss/init.py128 # if this argument is specified, just return itselfCOMMENT
MEDIUMtools/stats/upload_test_stats.py67 # To do this, we leverage an implementation detail of how we write outCOMMENT
LOW.ci/docker/manywheel/s390_scripts/build.sh51# So let's just use the same one pip and everyone usesCOMMENT
LOW.ci/docker/common/install_nccl.sh6# so we can just use the default NCCL version.COMMENT
LOW.ci/manywheel/build_common.sh177 # Note - just use whichever python we happen to be onCOMMENT
LOW.ci/pytorch/common_utils.sh145 # There won't be nvidia-smi in nogpu tests, so just set TORCH_CUDA_ARCH_LIST to the defaultCOMMENT
MEDIUM.ci/pytorch/run_tests.sh7# Essentially runs pytorch/test/run_test.py, but keeps track of which tests toCOMMENT
MEDIUM.ci/pytorch/test_fa3_abi_stable.sh22# To be as robust against upstream changes as possible, minimally only stageCOMMENT
MEDIUMtest/test_mps.py6712 # Essentially remained over integral types should rely on integers opsCOMMENT
LOWtest/test_mps.py15089 # fail, so just set CWD to this script's directorySTRING
LOWtest/test_mps.py15134 # fail, so just set CWD to this script's directorySTRING
LOWtest/test_python_dispatch.py2292 e.contiguous() # this will just return the original TensorImpl since is_contiguous = TrueCODE
LOWtest/test_import_stats.py7# time is greater than a certain threshold, but for now we just use themCOMMENT
LOWtest/test_tensor_creation_ops.py2636 # instead of repeating this calculation, we just use empty_strided which does the sameCOMMENT
MEDIUMtest/test_nn.py14825 # Probability-target counterpart of the scalar harness legs:COMMENT
LOWtest/test_pytree.py897 # fail, so just set CWD to this script's directorySTRING
MEDIUMtest/test_custom_ops.py1499 # Not comprehensive (it doesn't need to be), just a check that our mechanism worksCOMMENT
LOWtest/test_stateless.py905 # fail, so just set CWD to this script's directorySTRING
LOWtest/test_stateless.py927 # fail, so just set CWD to this script's directorySTRING
MEDIUMtest/test_jit_llga_fuser.py61 # so we disable AMP for JIT & leverage eager-mode AMP.COMMENT
LOWtest/test_cuda.py7186 # fail, so just set CWD to this script's directoryCOMMENT
LOWtest/test_cuda.py3326 # fail, so just set CWD to this script's directorySTRING
MEDIUMtest/run_test.py439 # in that they utilize the usual python test infrastructure.STRING
LOWtest/run_test.py519 # to just return successfully hereSTRING
MEDIUMtest/test_dlpack.py819 # Run the comprehensive C++ testSTRING
MEDIUMtest/test_meta.py788# useful for more comprehensive testing e.g., as seen inSTRING
MEDIUMtest/test_binary_ufuncs.py2057 # in a binary operation in conjunction with a Tensor on allCOMMENT
MEDIUMtest/test_autograd.py2975 # Definitely pick one tensor to addCOMMENT
LOWtest/test_autograd.py4750 # This op will just return grad_output two times in backwardCOMMENT
LOWtest/test_autograd.py11857 # fail, so just set CWD to this script's directorySTRING
MEDIUMtest/test_nestedtensor.py8996# OpInfo-based NJT tests. These tests utilize an NJT-specific op_db generated from the standardCOMMENT
MEDIUMtest/test_scaled_matmul_cuda.py1374 # by cosine similarity (robust to kernel differences) rather than element-wise.COMMENT
LOWtest/test_testing.py2486 # fail, so just set CWD to this script's directoryCOMMENT
MEDIUMtest/inductor/test_deterministic.py136 # Halving sweep, matching what the benchmark harness does.COMMENT
LOWtest/inductor/test_compiled_autograd.py186 # fail, so just set CWD to this script's directoryCOMMENT
LOWtest/inductor/test_triton_kernels.py5632 # we can just pass the function 'f' for dynamoSTRING
LOWtest/inductor/test_triton_kernels.py5916 # we can just pass the function 'f' for dynamoSTRING
MEDIUMtest/inductor/test_max_autotune_blackwell.py292 # Fundamentally we should fix the triton_blackwell_ws_persistent_device_tma templateCOMMENT
MEDIUMtest/inductor/test_cpu_repro.py472 # The cpp_wrapper C-shim can't utilize the Python error API, so errorCOMMENT
MEDIUMtest/inductor/test_compile_to_python.py431 # profiling harness. ``def call(`` matches both the flat ``def call(args)`` and theCOMMENT
MEDIUMtest/inductor/test_compile_to_python.py178 # than an Inductor profiling harness: none of these debug entry points appear.COMMENT
MEDIUMtest/inductor/test_control_flow.py57 # Simple robust loss: abs values + small constant to avoid inf/nanCOMMENT
MEDIUMtest/inductor/test_torchinductor.py17035 # odd number on purpose to trigger comprehensive paddingCOMMENT
MEDIUMtest/inductor/test_torchinductor_opinfo_properties.py634 """Get sample inputs from OpInfo using reference_inputs for comprehensive coverage."""STRING
MEDIUMtest/inductor/test_torchinductor_opinfo_properties.py635 # Use reference_inputs for more comprehensive test coverageCOMMENT
LOWtest/onnx/test_pytorch_onnx_onnxruntime.py13814 # Otherwise the entire model will just return a constant and not haveCOMMENT
246 more matches not shown…
Verbosity Indicators268 hits · 453 pts
SeverityFileLineSnippetContext
LOWtest/test_autograd.py8190 # Step 1: Checkpointed forward with batch_size=4 → static graph G0COMMENT
LOWtest/test_autograd.py8194 # Step 2: Non-checkpointed call with batch_size=7 triggers automaticCOMMENT
LOWtest/test_autograd.py8202 # Step 3: Backward. Checkpoint recomputes fn(x) but now hits G1COMMENT
LOWtest/test_testing.py890 # We need to check if this test is still needed or if this behavior is now enabled by default.COMMENT
LOWtest/cpp_extensions/cudnn_extension.cpp43 // Step 1: Check inputs. This will throw an error if inputs are invalid, so noCOMMENT
LOWtest/cpp_extensions/cudnn_extension.cpp46 // Step 2: Create descriptorsCOMMENT
LOWtest/cpp_extensions/cudnn_extension.cpp59 // Step 3: Apply CuDNN functionCOMMENT
LOWtest/cpp_extensions/cudnn_extension.cpp71 // Step 4: Destroy descriptorsCOMMENT
LOWtest/cpp_extensions/cudnn_extension.cpp73 // Step 5: Return something (optional)COMMENT
LOWtest/distributed/tensor/test_redistribute.py3087 # Step 1: reduce_scatter on mesh dim 0 (size 2)COMMENT
LOWtest/distributed/tensor/test_redistribute.py3091 # Step 2: reduce_scatter on mesh dim 1 (size 3)COMMENT
LOWtest/distributed/tensor/experimental/test_local_map.py842 # Step 1: BuggyDPMatmul without spmd_types -> wrong W gradient.COMMENT
LOWtest/distributed/tensor/experimental/test_local_map.py859 # Step 2: BuggyDPMatmul with spmd_types=True -> caught.COMMENT
LOWtest/distributed/tensor/experimental/test_local_map.py875 # Step 3: CorrectDPMatmul with spmd_types=True -> correct gradientsCOMMENT
LOWtest/distributed/tensor/parallel/test_tp_examples.py222 # Step 3: Run test by comparing outputs from single-gpu and multi-gpu models.COMMENT
LOWtest/distributed/tensor/parallel/test_tp_examples.py305 ) # Step 1: Initialize single-gpu models.CODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py308 ) # Step 2: Setup tp model, place onto device mesh.CODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py311 ) # Step 3: Setup optimizers for both modelsCODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py433 ) # Step 1: Initialize single-gpu models.CODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py436 ) # Step 2: Setup tp model, place onto device mesh.CODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py439 ) # Step 3: Setup optimizers for both modelsCODE
LOWtest/distributed/tensor/parallel/test_tp_examples.py442 ) # Step 4: set `requires_grad` patternsCODE
LOWtest/dynamo/test_activation_offloading.py596 # Step 1: cold — allocates new pinned bufferCOMMENT
LOWtest/dynamo/test_activation_offloading.py609 # Step 2: warm — reuses pooled bufferCOMMENT
LOWtest/dynamo/test_repros.py7545 # Step 1: Compile a static shapes graphCOMMENT
LOWtest/dynamo/test_repros.py7554 # Step 2: Compile a dynamic shapes graphCOMMENT
LOWtest/dynamo/test_repros.py7564 # Step 3: Run with Step 1's inputsCOMMENT
LOWtest/dynamo/test_repros.py7587 # Step 1: Compile a static shapes graphCOMMENT
LOWtest/dynamo/test_repros.py7596 # Step 2: Compile a dynamic shapes graphCOMMENT
LOWtest/dynamo/test_repros.py7606 # Step 3: Run with Step 1's inputsCOMMENT
LOWtest/profiler/test_profiler.py4295 # Step 3: under the patch, RECORD -> DEVICE_STOPPED fires stop_trace +COMMENT
LOWtorch/_custom_ops.py61 >>> # Step 1: define the custom op.CODE
LOWtorch/_custom_ops.py72 >>> # Step 2: Register an implementation for various PyTorch subsystemsCODE
LOWtorch/_custom_ops.py144 >>> # Step 1: define the custom op.CODE
LOWtorch/_custom_ops.py155 >>> # Step 2: Register an implementation for various PyTorch subsystemsCODE
LOWtorch/_ops.py421 # Step 1: dispatch on any user TorchDispatchModesCOMMENT
LOWtorch/_ops.py449 # Step 2: dispatch on any subclassesCOMMENT
LOWtorch/_ops.py682 # When we are unsetting a mode, we need to check if there isCOMMENT
LOWtorch/_ops.py732 # When we are setting a mode, we need to check if there isCOMMENT
LOWtorch/csrc/dynamo/init.cpp361 // even if __next__ is not defined, so we need to check if the method is trulyCOMMENT
LOWtorch/csrc/utils/python_arg_parser.cpp574 // Step 1: Try to dispatch based on the mode stack, *ignoring* infraCOMMENT
LOWtorch/csrc/utils/python_arg_parser.cpp624 // Step 1: Try to dispatch on any user TorchDispatchModes (including infraCOMMENT
LOWtorch/csrc/utils/python_arg_parser.cpp635 // Step 2: Try to dispatch based on any user subclasses,COMMENT
LOWtorch/csrc/jit/tensorexpr/loopnest.cpp1975 // We need to check if fusing the loops results in a loop-carried dependence.COMMENT
LOWtorch/csrc/lazy/ts_backend/ts_eager_fallback.cpp222 // Step 1: Convert all non-eager tensor inputs into eager tensors and put themCOMMENT
LOWtorch/csrc/lazy/ts_backend/ts_eager_fallback.cpp254 // Step 2: Call the underlying eager implementation of the operatorCOMMENT
LOWtorch/csrc/lazy/ts_backend/ts_eager_fallback.cpp257 // Step 3: We need to take special care to handle mutable aliases properly:COMMENT
LOWtorch/csrc/lazy/ts_backend/ts_eager_fallback.cpp268 // Step 4: Convert any eager output tensors back to the original input device.COMMENT
LOWtorch/_functorch/eager_transforms.py661 # Step 1: Construct grad_outputs by splitting the standard basisCOMMENT
LOWtorch/_functorch/eager_transforms.py778 # Step 2: The returned jacobian is one big tensor per input. In this step,COMMENT
LOWtorch/_functorch/eager_transforms.py792 # Step 3: Right now, `jacobian` is a List[List[Tensor]].COMMENT
LOWtorch/_functorch/_aot_autograd/functional_utils.py397 # For outputs aliasing inputs, we need to check if the requires-gradness has changed.COMMENT
LOWtorch/_functorch/_aot_autograd/functional_utils.py399 # For outputs aliasing inputs, we need to check if the dtype has changed.COMMENT
LOW…rch/_functorch/_aot_autograd/graph_capture_wrappers.py1341 # Step 1: wrap tensor inputs into subclasses if necessaryCOMMENT
LOW…rch/_functorch/_aot_autograd/graph_capture_wrappers.py1346 # Step 2: call the inner function, with our (maybe subclass) inputsCOMMENT
LOW…rch/_functorch/_aot_autograd/graph_capture_wrappers.py1379 # Step 3: Unwrap any subclass outputs back into dense tensorsCOMMENT
LOW…ctorch/_activation_offloading/activation_offloading.py782 # Step 1: Identify reload patternsCOMMENT
LOW…ctorch/_activation_offloading/activation_offloading.py787 # Step 2: Reorder nodes by directly manipulating the graphCOMMENT
LOW…ctorch/_activation_offloading/activation_offloading.py806 # Step 1: Decide which nodes to offload and mark themCOMMENT
LOW…ctorch/_activation_offloading/activation_offloading.py815 # Step 2: Add offload and reload nodes to the graphsCOMMENT
208 more matches not shown…
Modern AI Meta-Vocabulary76 hits · 236 pts
SeverityFileLineSnippetContext
MEDIUM.ci/macwheel/build.sh5# repair_wheel.py) are non-orchestrating stages that stay version-agnostic.COMMENT
MEDIUM.ci/macwheel/build_all_macos_wheels.sh8# Per-Python orchestration (env setup, deps, build, delocate) lives inCOMMENT
MEDIUM…ofiler/profiler_global_callback_stale_session_test.cpp12#include <torch/csrc/profiler/orchestration/observer.h>COMMENT
MEDIUMtest/functorch/test_compile_to_python.py307 # AROUND the orchestration (graph_compile._aot_stage2c_make_inference_function),COMMENT
MEDIUMtest/functorch/test_compile_to_python.py309 # orchestration -- dedup(orchestration(inner)) -- so the orchestration sees theCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py338 # orchestration. It collapses the aliased inputs into a synthetic base BEFORE theCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py339 # orchestration runs, so it must be composed OUTSIDE it (same requirement as dedup):COMMENT
MEDIUMtest/functorch/test_compile_to_python.py341 # which inverted the chain to orchestration(synthetic_base(inner))).COMMENT
MEDIUMtest/functorch/test_compile_to_python.py418 # so the inner compile MUST run under autocast for the orchestration to emitCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py420 # baked in; the orchestration then disables autocast at runtime to keep the denseCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py885 # exactly one forward orchestration wrapper and rejects an empty capture rather thanCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py895 # The inlined orchestration hoists its globals to module scope; a hoisted name thatCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py1008 # A valid forward orchestration wrapper whose codegen'd signature matches what theCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py1010 # broken sibling so the SIBLING is what trips the guard, not a missing orchestration.COMMENT
MEDIUMtest/functorch/test_compile_to_python.py1022 # front (before chain wiring) even when paired with a valid forward orchestration:COMMENT
MEDIUMtest/functorch/test_compile_to_python.py1082 # target origin (the last orchestration's), so a foreign-origin wrapper must beCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py1083 # dropped from the emitted source while the target orchestration still composes.COMMENT
MEDIUMtest/functorch/test_compile_to_python.py145 # _inner_call, and the AOTAutograd orchestration is inlined as a real def that theCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py283 # (the _alias_fn / gen_alias_from_base path, which the orchestration closes overCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py397 # those wrap the orchestration rather than the inner call -- see the dedup /COMMENT
MEDIUMtest/functorch/test_compile_to_python.py471 # The orchestration is spliced as a real top-level ``def _runtime_wrapper`` that theCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py483 # the orchestration is invoked with contextlib.nullcontext + a no-op in those twoCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py525 # that token is the orchestration's first parameter and appears in every composedCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py843 # The generated call invokes the orchestration positionally, so a changed signatureCOMMENT
MEDIUMtest/functorch/test_compile_to_python.py1067 # Two orchestration wrappers sharing one origin_id (so the origin filter keeps both)COMMENT
MEDIUMtorch/csrc/autograd/profiler_python.cpp29#include <torch/csrc/profiler/orchestration/python_tracer.h>COMMENT
MEDIUMtorch/csrc/autograd/profiler_kineto.cpp20#include <torch/csrc/profiler/orchestration/observer.h>COMMENT
MEDIUMtorch/csrc/mtia/profiler/MTIAMemoryProfiler.h2#include <torch/csrc/profiler/orchestration/python_tracer.h>COMMENT
MEDIUMtorch/csrc/profiler/collection.cpp2#include <torch/csrc/profiler/orchestration/vulkan.h>COMMENT
MEDIUMtorch/csrc/profiler/api.h3#include <torch/csrc/profiler/orchestration/observer.h>COMMENT
MEDIUMtorch/csrc/profiler/collection.h21#include <torch/csrc/profiler/orchestration/python_tracer.h>COMMENT
MEDIUMtorch/csrc/profiler/orchestration/python_tracer.cpp1#include <torch/csrc/profiler/orchestration/python_tracer.h>COMMENT
MEDIUMtorch/csrc/profiler/orchestration/observer.cpp1#include <torch/csrc/profiler/orchestration/observer.h>COMMENT
MEDIUMtorch/csrc/profiler/orchestration/vulkan.cpp1#include <torch/csrc/profiler/orchestration/vulkan.h>COMMENT
MEDIUMtorch/_functorch/_aot_autograd/runtime_wrappers.py979 # The orchestration closes over these two codegen'd epilogue functions directlyCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py320 # caught by the orchestration count/wiring guards below (a second orchestration tripsCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py322 # target is the origin of the LAST captured orchestration wrapper: a foreign loweringCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py323 # appends its orchestration before the outer one finishes, so the final orchestrationCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py421 # This is what lets the composer tell the inner call apart from the orchestration'sCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py424 # wrapping the inner call, composed INSIDE the orchestration) are distinguished fromCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py426 # orchestration in graph_compile._aot_stage2c_make_inference_function).COMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py429 # Name of the single-arg adapter emitted over the orchestration; the innermost outerCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py430 # wrapper closes over the orchestration's outer closure and is wired to this name.COMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py433 # The orchestration's outer closure is the object outer wrappers wrap. The composerCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py434 # never captures it directly (the captured orchestration ``fn`` is the innerCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py620 # The single-arg adapter over the orchestration (emitted only when outer wrappersCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py622 # orchestration's positional (_compiled_fn_, _first_ctx_, _on_before_call_, args)COMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py623 # signature to that. When there are no outer wrappers the orchestration is invokedCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py704 # ``call`` invokes the orchestration directly when nothing wraps it; when outerCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py706 # adapter, then the orchestration) -- so the orchestration-substitution comment livesCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py707 # on whichever site actually invokes the orchestration.COMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py93# calls the link it wraps -- the orchestration's outer closure for the innermost outerCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py94# wrapper), plus the orchestration's output-alias and mutation epilogue helpers, whichCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py108# _replay_aliases_ = _alias_fn # (3) orchestration's refCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py223# codegen'd runtime wrappers, not reimplemented: each (the orchestration and any chainCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py246 # An OUTER wrapper (dedup / synthetic base) closes over the orchestration's outerCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py248 # path, so wire it to the single-arg orchestration entry adapter the composer emits.COMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py359 # The generated ``call`` invokes the orchestration POSITIONALLY by its own name (seeCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py407 # ``_compiled_fn_``). The orchestration takes its inner as a call-time arg, not aCOMMENT
MEDIUMtorch/_functorch/_aot_autograd/to_standalone_python.py452 # Walk the OUTER chain outward from the orchestration closure: the innermost outerCOMMENT
16 more matches not shown…
Overly Generic Function Names260 hits · 152 pts
SeverityFileLineSnippetContext
LOWtools/test/heuristics/test_heuristics.py270 def helper(test_file: str, changed_file: str) -> bool:CODE
LOWtest/test_mps.py3699 def helper(binary_op):CODE
LOWtest/test_mps.py4679 def helper(dtype):CODE
LOWtest/test_mps.py4690 def helper(dtype):CODE
LOWtest/test_mps.py257 def helper(model_cpu, model_mps, dtype, iterations, batch_size, atol=3e-4, rtol=1e-5):CODE
LOWtest/test_mps.py321 def helper(model, optimizer, scaler, device, input, target, apply_grad_transform=False):CODE
LOWtest/test_mps.py1004 def helper(val, shape, dtype):CODE
LOWtest/test_mps.py1220 def helper(dtype):CODE
LOWtest/test_mps.py1454 def helper(input_shape, batch1_shape, batch2_shape):CODE
LOWtest/test_mps.py1520 def helper(bias_shape):CODE
LOWtest/test_mps.py1842 def helper(shape, ks, padding=0, dilation=1, ceil_mode=False, return_indices=False, test_ties=False):CODE
LOWtest/test_mps.py1916 def helper(size, memory_format):CODE
LOWtest/test_mps.py1942 def helper(shape):CODE
LOWtest/test_mps.py2051 def helper(shape, channels_last=False):CODE
LOWtest/test_mps.py2068 def helper(shape, eps=1, momentum=0.1, wts=False, training=False, channels_last=False,CODE
LOWtest/test_mps.py2621 def helper(input_shape, normalized_shape, eps=1e-05, elementwise_affine=True, dtype=torch.float32, non_contiguouCODE
LOWtest/test_mps.py2723 def helper(shape, eps=1, momentum=0.1, wts=False, channels_last=False, track_running_stats=True, test_module=FalCODE
LOWtest/test_mps.py2900 def helper(dim, layer='linear', dtype=torch.float32):CODE
LOWtest/test_mps.py2970 def helper(input_shape, wt_shape,CODE
LOWtest/test_mps.py3071 def helper(input_shape, wt_shape,CODE
LOWtest/test_mps.py3139 def helper(shape):CODE
LOWtest/test_mps.py3164 def helper(shape):CODE
LOWtest/test_mps.py3188 def helper(threshold, value, num_elems, inplace=False, requires_grad=True):CODE
LOWtest/test_mps.py3215 def helper(shape):CODE
LOWtest/test_mps.py3248 def helper(shape, value, xtype=torch.float32, ytype=None, ztype=None):CODE
LOWtest/test_mps.py3285 def helper(shape, value):CODE
LOWtest/test_mps.py3311 def helper(shape, value):CODE
LOWtest/test_mps.py3501 def helper(shape):CODE
LOWtest/test_mps.py3621 def helper(x_tensor, y_func, z_func, r_func=None):CODE
LOWtest/test_mps.py3733 def helper(operator):CODE
LOWtest/test_mps.py3825 def helper(shape, repeats):CODE
LOWtest/test_mps.py3924 def helper(shape, dtype=torch.float32, num_repeats=torch.Tensor(), dim=None):CODE
LOWtest/test_mps.py3949 def helper(dtype):CODE
LOWtest/test_mps.py4049 def helper(data, to_dtype):CODE
LOWtest/test_mps.py4067 def helper(data, to_dtype):CODE
LOWtest/test_mps.py4180 def helper(dtype):CODE
LOWtest/test_mps.py4237 def helper(x, return_inverse, return_counts):CODE
LOWtest/test_mps.py4257 def helper(x, dim, return_inverse, return_counts):CODE
LOWtest/test_mps.py4461 def helper(src_shape, dst_shape, src_dtype, dst_dtype):CODE
LOWtest/test_mps.py4599 def helper(dtype, noncontiguous, dim):CODE
LOWtest/test_mps.py4644 def helper(dtype):CODE
LOWtest/test_mps.py4662 def helper(dtype):CODE
LOWtest/test_mps.py4772 def helper(shape, dtype):CODE
LOWtest/test_mps.py4972 def helper(n, c):CODE
LOWtest/test_mps.py4989 def helper(shape, reduction):CODE
LOWtest/test_mps.py5018 def helper(shape, reduction):CODE
LOWtest/test_mps.py5088 def helper(shape, reduction):CODE
LOWtest/test_mps.py5362 def helper(shape):CODE
LOWtest/test_mps.py5375 def helper(shape):CODE
LOWtest/test_mps.py5386 def helper(shape):CODE
LOWtest/test_mps.py5399 def helper(shape):CODE
LOWtest/test_mps.py5410 def helper(shape):CODE
LOWtest/test_mps.py5423 def helper(shape):CODE
LOWtest/test_mps.py5434 def helper(shape):CODE
LOWtest/test_mps.py5447 def helper(shape):CODE
LOWtest/test_mps.py5458 def helper(shape):CODE
LOWtest/test_mps.py5471 def helper(shape):CODE
LOWtest/test_mps.py5500 def helper(n, c, h, w, reduction_type, dtype=torch.float32):CODE
LOWtest/test_mps.py5635 def helper(n, c, h, w, dtype=torch.float32):CODE
LOWtest/test_mps.py5800 def helper(shape):CODE
200 more matches not shown…
Dead Code78 hits · 152 pts
SeverityFileLineSnippetContext
MEDIUMtest/test_datapipe.py3741CODE
MEDIUMtest/test_functionalization.py770CODE
MEDIUMtest/test_functionalization.py771CODE
MEDIUMtest/test_functionalization.py772CODE
MEDIUMtest/test_functionalization.py773CODE
MEDIUMtest/test_jit.py14077CODE
MEDIUMtest/test_jit.py9927CODE
MEDIUMtest/test_jit_fuser.py734CODE
MEDIUMtest/test_jit_fuser.py735CODE
MEDIUMtest/test_jit_fuser.py737CODE
MEDIUMtest/test_jit_fuser.py739CODE
MEDIUMtest/test_jit_fuser.py742CODE
MEDIUMtest/test_jit_fuser.py746CODE
MEDIUMtest/test_autograd.py15715CODE
MEDIUMtest/test_fx.py5334CODE
MEDIUMtest/test_fx.py3939CODE
MEDIUMtest/inductor/test_ordered_set.py25CODE
MEDIUMtest/inductor/test_ordered_set.py976CODE
MEDIUMtest/onnx/test_utility_funs.py109CODE
MEDIUMtest/dynamo/test_misc.py9788CODE
MEDIUMtest/dynamo/test_misc.py12566CODE
MEDIUMtest/dynamo/test_generator.py1236CODE
MEDIUMtest/dynamo/test_exceptions.py657CODE
MEDIUMtest/dynamo/test_guard_manager.py918CODE
MEDIUMtest/quantization/fx/test_quantize_fx.py2564CODE
MEDIUMtest/quantization/fx/test_quantize_fx.py2566CODE
MEDIUMtest/quantization/fx/test_quantize_fx.py2569CODE
MEDIUMtest/cpython/v3_13/test_itertools.py2723CODE
MEDIUMtest/cpython/v3_13/test_generators.py566CODE
MEDIUMtest/cpython/v3_13/test_generators.py577CODE
MEDIUMtest/cpython/v3_13/test_exceptions.py1836CODE
MEDIUMtest/cpython/v3_13/test_set.py77CODE
MEDIUMtest/cpython/v3_13/test_set.py1109CODE
MEDIUMtest/cpython/v3_13/test_deque.py40CODE
MEDIUMtest/cpython/v3_13/test_builtin.py1288CODE
MEDIUMtest/jit/test_backends.py595CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py501CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py987CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py988CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py989CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py992CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py1003CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py1004CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py1005CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py1007CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py1013CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py1014CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py1151CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py1152CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py1207CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py1209CODE
MEDIUMtest/torch_np/numpy_tests/core/test_indexing.py1210CODE
MEDIUMtest/functorch/test_vmap.py1004CODE
MEDIUMtest/functorch/test_vmap.py1007CODE
MEDIUMtest/functorch/test_vmap.py1010CODE
MEDIUMtest/functorch/test_vmap.py1013CODE
MEDIUMtest/functorch/test_vmap.py1017CODE
MEDIUMtest/functorch/test_aotdispatch.py334CODE
MEDIUMtest/functorch/test_aotdispatch.py337CODE
MEDIUMtest/functorch/test_aotdispatch.py338CODE
18 more matches not shown…
Slop Phrases47 hits · 83 pts
SeverityFileLineSnippetContext
MEDIUMtools/autograd/derivatives.yaml37# If None of the output is differentiable, you can also add the functionCOMMENT
LOWtest/test_jit.py455 # you are deleting `constants.pkl`, make sure to update `torch.serialization.load`COMMENT
LOWtest/test_jit.py2536 # NB: We make sure to pass in a batch with a different max sequenceSTRING
MEDIUMtest/test_fx.py4161 # Used to test that you can use your own placeholder classCOMMENT
LOWtest/inductor/test_triton_kernels.py6156 # so we wanna make sure to correctly handle edge cases with tricky line breaksSTRING
MEDIUMtest/inductor/test_profiler.py83 # Debugging tips: you can add prof.export_chrome_trace("test.json") inline inCOMMENT
LOWtest/dynamo/test_repros.py1612 # NB: When you remove the expectedFailure, don't forget toCOMMENT
LOWtest/dynamo/test_repros.py1612 # NB: When you remove the expectedFailure, don't forget toCOMMENT
LOWtest/quantization/fx/test_model_report_fx.py1958 # if they passed in fusion parameter, make sure to test thatCOMMENT
MEDIUMtest/ao/sparsity/test_structured_sparsifier.py999 # Manually set the filter weights for demonstration purposesCOMMENT
LOW.claude/skills/triaging-issues/labels.json416 "description": "Optimization passes written in FX (don't forget to select a more specific label)"CODE
MEDIUMtorch/torch_version.py38 # Ideally this should work for most cases by attempting to groupCOMMENT
LOWtorch/headeronly/core/ScalarType.h341 // Don't forget to extend this when adding new QInt typesCOMMENT
MEDIUMtorch/csrc/jit/backends/backend_debug_info.h51 * don't serialize this info as part of LoweredModule as mentioned earlier.COMMENT
MEDIUMtorch/_functorch/_aot_autograd/graph_compile.py995 # Step 1) - Get the `joint_hop_gm`. As mentioned earlier, theCOMMENT
LOWtorch/_subclasses/meta_utils.py962 # Need to make sure to resize the meta storage too.COMMENT
LOWtorch/nn/utils/parametrizations.py135 # We always make sure to copy Q in every pathCOMMENT
MEDIUMtorch/nn/modules/loss.py2215 >>> S_min = 10 # Minimum target length, for demonstration purposesCODE
LOWtorch/nn/modules/activation.py1363 # When lifting this restriction, don't forget to eitherSTRING
LOWtorch/nn/modules/activation.py1363 # When lifting this restriction, don't forget to eitherSTRING
LOW…nnx/_internal/torchscript_exporter/symbolic_opset11.py980 # make sure to cast other to self's typeCOMMENT
LOW…nnx/_internal/torchscript_exporter/symbolic_opset11.py1014 # make sure to cast other to self's typeCOMMENT
LOW…onnx/_internal/torchscript_exporter/symbolic_opset9.py2159 # make sure to cast other to self's typeCOMMENT
LOW…onnx/_internal/torchscript_exporter/symbolic_opset9.py2188 # make sure to cast other to self's typeCOMMENT
LOW…nnx/_internal/torchscript_exporter/symbolic_opset12.py77 # make sure to cast other to self's typeCOMMENT
LOWtorch/distributed/elastic/rendezvous/etcd_rendezvous.py886 # make sure you don't call any long-blocking methods that do notCOMMENT
LOWtorch/distributed/tensor/_dtensor_spec.py448 # use, where we make sure to update the hash when the `tensor_meta`COMMENT
LOWtorch/fx/graph.py2492 # When generating Python code, we need to make sure to name thingsCOMMENT
MEDIUMtorch/fx/experimental/symbolic_shapes.py1536# (1) If you can use a higher level combinator prefer using those instead, they are definitely safe (modulo short-circuiCOMMENT
LOWtorch/fx/experimental/proxy_tensor.py743 # from_real_tensor to create fake values (don't forget toCOMMENT
LOWtorch/fx/experimental/proxy_tensor.py743 # from_real_tensor to create fake values (don't forget toCOMMENT
LOWtorch/fx/passes/reinplace.py764 # When that happens, we need to make sure to keep ourSTRING
MEDIUMtorch/optim/lr_scheduler.py368 >>> # Alternatively, you can use a single lambda function for all groups.CODE
LOWtorch/_inductor/cpp_builder.py2520 # When we build remotely, we need to make sure to carefully copy any filesCOMMENT
LOWtorch/_inductor/compile_fx_ext.py646 # Since our metrics were gathered in a subprocess make sure to add themCOMMENT
LOWtorch/_inductor/codegen/common.py1297# NB: if you add a new special function, don't forget to updateCOMMENT
LOWtorch/_inductor/codegen/common.py1297# NB: if you add a new special function, don't forget to updateCOMMENT
LOWtorch/utils/_python_dispatch.py956 # Next: we need to make sure to return inputs directly, if the output is a mutable alias (e.g. add_()).COMMENT
LOWtorch/utils/hipify/hipify_python.py734 # make sure to check the end-of-word marker presentCOMMENT
LOWtorch/utils/data/sampler.py73 # implementing a `__len__` method. In such cases, we must make sure to notSTRING
MEDIUMtorch/_dynamo/backends/onnxrt.py34# " (4) If it returns `True`, then you can use `onnxrt` backend.\n"COMMENT
MEDIUMbenchmarks/dynamo/common.py1036 # Hypothetically you can use this from other places, but it's currentlyCOMMENT
LOW.github/scripts/test_trymerge.py8# GraphQL queries in trymerge.py, please make sure to delete `gql_mocks.json`COMMENT
LOWaten/src/ATen/OpaqueTensorImpl.h49 // unnecessary; don't forget to change that if needed!COMMENT
LOWaten/src/ATen/SparseTensorImpl.cpp50 // unnecessary; don't forget to change that if needed!COMMENT
MEDIUMtorchgen/api/types/types_base.py179# which you can use to figure out if a given Binding is appropriateCOMMENT
LOWc10/core/StorageImpl.h106 // unnecessary; don't forget to change that if needed!COMMENT
TODO Padding53 hits · 77 pts
SeverityFileLineSnippetContext
LOWtest/test_mps.py9662 # TODO: Add tests for data typesCOMMENT
LOWtest/test_dataloader.py2697 # TODO: Fix test for 'loader_kill' that would cause running out of shared memory.COMMENT
LOWtest/test_reductions.py1616 # TODO: consider refactoring with bincount testCOMMENT
LOWtest/test_sparse.py5148 # TODO: implement batch support in _convert_indices_from_csr_to_cooCOMMENT
LOWtest/test_ops.py710 # TODO: consider adding some tolerance to this comparisonCOMMENT
LOWtest/test_ops.py2272 # TODO: add tests for `R->C` functionsCOMMENT
LOW…composable/test_composability/test_2d_composability.py509 # TODO: Add tests when using gradient_as_bucket_view and static_graph for DDP.COMMENT
LOWtest/distributed/tensor/test_matrix_ops.py914 # TODO: Add test cases where is_causal=False and an attention mask is provided.COMMENT
LOWtest/quantization/core/test_quantized_module.py1896 # TODO: add tests for conv and linearSTRING
LOWtest/cpp/api/support.h116 // TODO: implement abs on CharTensor (int8)COMMENT
LOWtest/jit/test_tracer.py642 # TODO: implementCOMMENT
LOWtest/torch_np/test_reductions.py364 # TODO: add tests with np.int32(3) etc, when implementedCOMMENT
LOWtest/torch_np/test_reductions.py524 # TODO: add tests with np.int32(3) etc, when implementedCOMMENT
LOWtest/functorch/test_ops.py1032 # TODO: implement batching ruleCOMMENT
LOWtest/functorch/test_ops.py1146 # TODO: implement batching ruleCOMMENT
LOWtest/functorch/test_ops.py1305 # TODO: implement batching ruleCOMMENT
LOWtest/functorch/test_ops.py1506 # TODO: implement batching ruleCOMMENT
LOWtest/functorch/test_ops.py1620 # TODO: implement batching ruleCOMMENT
LOWtest/functorch/test_ops.py2006 # TODO: implement batching ruleCOMMENT
LOWtest/functorch/test_vmap.py4483 # TODO: implement batching ruleCOMMENT
LOWtest/functorch/test_vmap.py4552 # TODO: implement batching ruleCOMMENT
LOWtorch/_prims/__init__.py400# TODO: implement dtype validation here, too, or on the corresponding refsCOMMENT
LOWtorch/csrc/Storage.cpp368// TODO: implement equalityCOMMENT
LOWtorch/csrc/inductor/aoti_torch/shim_common.cpp468// TODO: implement a more efficient version instead of calling into atenCOMMENT
LOWtorch/csrc/inductor/aoti_torch/shim_common.cpp867// TODO: implement a more efficient version instead of calling into atenCOMMENT
LOWtorch/csrc/inductor/aoti_torch/shim_common.cpp938// TODO: implement a more efficient version instead of calling into atenCOMMENT
LOWtorch/csrc/inductor/aoti_torch/shim_common.cpp956// TODO: implement a more efficient version instead of calling into atenCOMMENT
LOWtorch/csrc/inductor/aoti_torch/shim_common.cpp979// TODO: implement a more efficient version instead of calling into atenCOMMENT
LOWtorch/csrc/jit/mobile/nnc/aot_compiler.cpp152 // TODO: implement prealloc optimization and fill in temp_sizesCOMMENT
LOWtorch/distributed/elastic/rendezvous/etcd_rendezvous.py949 # TODO: implement timeoutCOMMENT
LOWtorch/distributed/_shard/sharded_optim/api.py82 # TODO: implement state_dictSTRING
LOWtorch/distributed/_shard/sharded_optim/api.py92 # TODO: implement load_state_dictSTRING
LOWtorch/distributed/_shard/sharded_optim/api.py99 # TODO: implement add_param_groupCOMMENT
LOWtorch/fx/traceback.py77 # TODO: add logging to tlparseCOMMENT
LOWtorch/_prims_common/wrappers.py55# TODO: implement ref.cast with an option to enforce safe castingCOMMENT
LOWtorch/masked/_ops.py846 # TODO: implement sparse CSR specific where operator for efficiencyCOMMENT
LOWtorch/_inductor/utils.py4296 # TODO: implement V3_BACKENDS_TUPLECOMMENT
LOWtorch/utils/model_dump/__init__.py356 # TODO: handle errors here and just ignore the file?COMMENT
LOWtorch/ao/quantization/fx/_lower_to_native_backend.py399# TODO: add tests for lowering these opsCOMMENT
LOWtorch/ao/quantization/fx/fuse.py119 # TODO: add validation that root_node is a module and has the same typeCOMMENT
LOWtorch/_refs/__init__.py1847# TODO: consider refactoring this with add implCOMMENT
LOWtorch/_refs/special/__init__.py235# TODO: add docstringCOMMENT
LOWaten/src/ATen/LegacyBatchedTensorImpl.cpp85// TODO: implement proper contiguity on batched tensor, then putCOMMENT
LOWaten/src/ATen/TensorIndexing.h211 // TODO: implement negative stepCOMMENT
LOWaten/src/ATen/native/Embedding.cpp82 // TODO: implement scale_grad_by_freqCOMMENT
LOWaten/src/ATen/native/LinearAlgebra.cpp572// TODO: implement _out variant avoiding copy and using already allocated storage directlyCOMMENT
LOWaten/src/ATen/native/LinearAlgebra.cpp3408// TODO: implement _out variant avoiding copy and using already allocated storage directlyCOMMENT
LOWaten/src/ATen/native/LinearAlgebra.cpp3457// TODO: implement _out variant avoiding copy and using already allocated storage directlyCOMMENT
LOWaten/src/ATen/native/quantized/cudnn/Conv.cpp299 // TODO: implement this for groups > 1; should be straightforward since we're only padding a single dimensionCOMMENT
LOWaten/src/ATen/native/quantized/cudnn/ConvPrepack.cpp80 // TODO: implement this for groups > 1COMMENT
LOWaten/src/ATen/functorch/BatchedTensorImpl.cpp127// TODO: implement proper contiguity on batched tensor, then putCOMMENT
LOWtorchgen/model.py1933 # TODO: implement a proper parser if this gets more uglyCOMMENT
LOWc10/core/TensorImpl.h1332 // TODO: implement layout() as native function/method so thatCOMMENT
AI Response Leakage11 hits · 74 pts
SeverityFileLineSnippetContext
HIGHtest/test_decomp.py1058 # explicit clearing is necessary as I will create a fresh modeCOMMENT
HIGHtest/test_fx_experimental.py1030 # Based on the code from - transformers/models/granitemoe/modeling_granitemoe.pySTRING
HIGHtest/inductor/test_max_autotune.py1253 # In this example, AOTInductor's JIT-compile will fuse(buf1, buf2) dueCOMMENT
HIGHtest/dynamo/test_repros.py4938 # In this example, there were 4950 guards (roughly (# tensors) ^ 2 // 2),COMMENT
HIGHtest/jit/test_freezing.py988 # In this example, 'a' is not mutated. However, we do not track which subCOMMENT
HIGHtorch/_dynamo/variables/higher_order_ops.py882 # used. In this example, the capture graph looks like:COMMENT
HIGHtorch/_dynamo/variables/higher_order_ops.py1896 # In this example, list `z` is in outer scope and gets appendedCOMMENT
HIGHtorch/_dynamo/variables/builtin.py706 # In this example, A.__radd__() is not called -> nothing is printed, becauseCOMMENT
HIGHtorch/profiler/profiler.py950 # In this example with wait=1, warmup=1, active=2, repeat=1,STRING
HIGH…er_guide/torch_compiler/torch.compiler_aot_inductor.md62 # [Note] In this example we directly feed the exported module to aoti_compile_and_package.COMMENT
HIGHaten/src/ATen/native/mps/operations/RnnOps.mm819 // In this implementation these two biases are added together and used further.COMMENT
Example Usage Blocks16 hits · 23 pts
SeverityFileLineSnippetContext
LOWtools/vendoring/quack/vendor.sh8# Usage:COMMENT
LOWtorch/library.h961/// Example usage:COMMENT
LOWtorch/library.h1038/// Example usage:COMMENT
LOWtorch/csrc/api/include/torch/python.h198/// Example usage:COMMENT
LOWtorch/csrc/api/include/torch/python.h234/// Example usage:COMMENT
LOWtorch/csrc/api/include/torch/nn/modules/common.h6/// Example usage:COMMENT
LOWtorch/nn/attention/flex_attention.py48# Usage:COMMENT
LOWtorch/_inductor/config.py177# Example usage:COMMENT
LOWtorch/_inductor/fuzzer.py538 # Example usage:STRING
LOWtorch/testing/_internal/common_device_type.py1062 # Usage:COMMENT
LOWtorch/testing/_internal/common_device_type.py1074 # Usage:COMMENT
LOWtorch/testing/_internal/common_device_type.py1305# Example usage:COMMENT
LOWtorch/_dynamo/variables/higher_order_ops.py150# Example usage:COMMENT
LOWthird_party/perfetto/update.sh8# Usage:COMMENT
LOWaten/src/ATen/TensorMeta.h17// Example usage:COMMENT
LOWaten/src/ATen/TensorMeta.h49// Example usage:COMMENT
Synthetic Comment Markers3 hits · 22 pts
SeverityFileLineSnippetContext
HIGHtest/test_view_ops.py1110 # Test various configurations - one per line as requestedCOMMENT
HIGHtest/quantization/fx/test_quantize_fx.py8938 # reshape will be quantized to fp16 as requested by this qconfigCOMMENT
HIGHc10/util/Backtrace.cpp226 // Skip as many frames as requested.COMMENT