pytorch / pytorch
File Size

The distribution of size of files (measured in lines of code).

Intro
  • File size measurements show the distribution of size of files.
  • Files are classified in four categories based on their size (lines of code): 1-100 (very small files), 101-200 (small files), 201-500 (medium size files), 501-1000 (long files), 1001+(very long files).
  • It is a good practice to keep files small. Long files may become "bloaters", code that have increased to such gargantuan proportions that they are hard to work with.
Learn more...
File Size Overall
  • There are 6,667 files with 1,190,840 lines of code.
    • 146 very long files (341,906 lines of code)
    • 348 long files (238,854 lines of code)
    • 975 medium size files (303,055 lines of codeclsfd_ftr_w_mp_ins)
    • 1,080 small files (153,861 lines of code)
    • 4,118 very small files (153,164 lines of code)
28% | 20% | 25% | 12% | 12%
Legend:
1001+
501-1000
201-500
101-200
1-100


explore: zoomable circles | sunburst | 3D view
File Size per Extension
1001+
501-1000
201-500
101-200
1-100
cpp37% | 22% | 24% | 8% | 6%
py26% | 18% | 25% | 14% | 15%
h19% | 15% | 25% | 16% | 22%
cc19% | 21% | 26% | 17% | 15%
hpp87% | 2% | 4% | 1% | 3%
yaml98% | 0% | 0% | 0% | 1%
cu10% | 31% | 34% | 13% | 9%
c22% | 30% | 24% | 14% | 8%
cmake33% | 10% | 28% | 7% | 20%
cuh23% | 7% | 48% | 13% | 6%
pb100% | 0% | 0% | 0% | 0%
mm33% | 7% | 19% | 19% | 19%
bzl81% | 0% | 0% | 0% | 18%
metal100% | 0% | 0% | 0% | 0%
S0% | 51% | 47% | 0% | <1%
in0% | 54% | 11% | 16% | 18%
pyi0% | 24% | 13% | 27% | 34%
js0% | 100% | 0% | 0% | 0%
java0% | 0% | 83% | 0% | 16%
proto0% | 0% | 59% | 0% | 40%
glsl0% | 0% | 0% | 11% | 88%
yy0% | 0% | 0% | 100% | 0%
html0% | 0% | 0% | 85% | 14%
rb0% | 0% | 0% | 56% | 43%
css0% | 0% | 0% | 0% | 100%
ll0% | 0% | 0% | 0% | 100%
m0% | 0% | 0% | 0% | 100%
pt0% | 0% | 0% | 0% | 100%
ps10% | 0% | 0% | 0% | 100%
toml0% | 0% | 0% | 0% | 100%
lds0% | 0% | 0% | 0% | 100%
File Size per Logical Decomposition
primary
1001+
501-1000
201-500
101-200
1-100
torch28% | 22% | 24% | 11% | 12%
aten33% | 21% | 25% | 10% | 8%
caffe226% | 17% | 24% | 15% | 15%
c1028% | 15% | 26% | 16% | 14%
tools27% | 11% | 25% | 18% | 16%
cmake32% | 10% | 27% | 9% | 19%
benchmarks0% | 6% | 47% | 26% | 19%
ROOT0% | 86% | 0% | 0% | 13%
binaries0% | 14% | 39% | 23% | 22%
android0% | 23% | 46% | 21% | 8%
modules0% | 0% | 18% | 36% | 44%
scripts0% | 0% | 24% | 52% | 23%
ios0% | 0% | 0% | 40% | 59%
mypy_plugins0% | 0% | 0% | 0% | 100%
Longest Files (Top 50)
File# lines# units
SobolEngineOpsUtils.cpp
in aten/src/ATen/native
 42413 -
vk_mem_alloc.h
in aten/src/ATen/native/vulkan/api
 12820 484
common_methods_invocations.py
in torch/testing/_internal
 12259 383
cl.hpp
in caffe2/contrib/opencl/OpenCL
 10263 215
cl.hpp
in caffe2/mobile/contrib/libopencl-stub/include/CL
 9959 208
native_functions.yaml
in aten/src/ATen/native
 8466 -
cuda_to_hip_mappings.py
in torch/utils/hipify
 8181 -
common_nn.py
in torch/testing/_internal
 5816 118
vulkan.h
in caffe2/mobile/contrib/libvulkan-stub/include/vulkan
 5488 -
ir_emitter.cpp
in torch/csrc/jit/frontend
 4328 122
external_functions_codegen.cpp
in torch/csrc/jit/tensorexpr
 3297 155
FunctionsManual.cpp
in torch/csrc/autograd
 3181 188
cc
embedding_lookup_idx_avx2.cc
in caffe2/perfkernels
 2972 18
QuantizedOpKernels.cpp
in aten/src/ATen/native/quantized/cpu/kernels
 2972 25
cc
embedding_lookup_fused_8bit_rowwise_idx_avx2.cc
in caffe2/perfkernels
 2968 18
register_prim_ops.cpp
in torch/csrc/jit/runtime
 2958 23
cc
embedding_lookup_avx2.cc
in caffe2/perfkernels
 2912 18
cc
embedding_lookup_fused_8bit_rowwise_avx2.cc
in caffe2/perfkernels
 2902 18
BatchLinearAlgebra.cpp
in aten/src/ATen/native
 2819 116
cu
math_gpu.cu
in caffe2/utils
 2801 -
functional.py
in torch/nn
 2645 118
BatchLinearAlgebra.cpp
in aten/src/ATen/native/cuda
 2525 102
loopnest.cpp
in torch/csrc/jit/tensorexpr
 2502 134
parser.cpp
in torch/csrc/jit/codegen/cuda
 2466 54
cc
math_cpu.cc
in caffe2/utils
 2441 65
symbolic_opset9.py
in torch/onnx
 2371 260
pb
concat_elim_test_net.pb
in caffe2/opt/custom
 2367 -
mobile_bytecode_generated.h
in torch/csrc/jit/serialization
 2339 241
ops.cpp
in torch/csrc/jit/runtime/static
 2339 41
variant.h
in c10/util
 2317 93
LinearAlgebra.cpp
in aten/src/ATen/native
 2281 91
ProcessGroupGloo.cpp
in torch/csrc/distributed/c10d
 2263 104
mpscnn.mm
in caffe2/mobile/contrib/ios/mpscnn
 2213 38
ir_simplifier.cpp
in torch/csrc/jit/tensorexpr
 2198 67
cu
pool_op.cu
in caffe2/operators
 2151 -
fusion_segmenter.cpp
in torch/csrc/jit/codegen/cuda
 2148 72
index_compute.cpp
in torch/csrc/jit/codegen/cuda
 2108 37
vec256_zarch.h
in aten/src/ATen/cpu/vec/vec256/zarch
 2070 81
order_preserving_flat_hash_map.h
in c10/util
 2068 302
common_utils.py
in torch/testing/_internal
 2066 162
TensorShape.cpp
in aten/src/ATen/native
 2064 123
shape_type_inference.cpp
in torch/csrc/jit/passes/onnx
 2036 50
flat_hash_map.h
in c10/util
 2001 294
script_init.cpp
in torch/csrc/jit/python
 1978 40
ir.cpp
in torch/csrc/jit/ir
 1885 87
optimizer.py
in caffe2/python
 1845 79
cmake
FindCUDA.cmake
in cmake/Modules_CUDA_fix/upstream
 1839 -
shape_analysis.cpp
in torch/csrc/jit/passes
 1821 15
cmake
Dependencies.cmake
in cmake
 1814 -
libopencl.c
in caffe2/mobile/contrib/libopencl-stub/src
 1810 4
Files With Most Units (Top 20)
File# lines# units
vk_mem_alloc.h
in aten/src/ATen/native/vulkan/api
 12820 484
common_methods_invocations.py
in torch/testing/_internal
 12259 383
order_preserving_flat_hash_map.h
in c10/util
 2068 302
flat_hash_map.h
in c10/util
 2001 294
symbolic_opset9.py
in torch/onnx
 2371 260
mobile_bytecode_generated.h
in torch/csrc/jit/serialization
 2339 241
cl.hpp
in caffe2/contrib/opencl/OpenCL
 10263 215
cl.hpp
in caffe2/mobile/contrib/libopencl-stub/include/CL
 9959 208
ivalue_inl.h
in aten/src/ATen/core
 1803 206
common_quantization.py
in torch/testing/_internal
 1664 202
FunctionsManual.cpp
in torch/csrc/autograd
 3181 188
tree_views.h
in torch/csrc/jit/frontend
 1021 188
jit_type.h
in aten/src/ATen/core
 1616 181
kernel_ir.h
in torch/csrc/jit/codegen/cuda
 1221 179
common_utils.py
in torch/testing/_internal
 2066 162
external_functions_codegen.cpp
in torch/csrc/jit/tensorexpr
 3297 155
transforms.py
in torch/distributions
 708 136
loopnest.cpp
in torch/csrc/jit/tensorexpr
 2502 134
TensorShape.cpp
in aten/src/ATen/native
 2064 123
ir_emitter.cpp
in torch/csrc/jit/frontend
 4328 122
Files With Long Lines (Top 20)

There are 774 files with lines longer than 120 characters. In total, there are 6052 long lines.

File# lines# units# long lines
native_functions.yaml
in aten/src/ATen/native
 8466 - 521
pb
concat_elim_test_net.pb
in caffe2/opt/custom
 2367 - 470
vulkan.h
in caffe2/mobile/contrib/libvulkan-stub/include/vulkan
 5488 - 233
derivatives.yaml
in tools/autograd
 1722 - 212
libvulkan-stub.c
in caffe2/mobile/contrib/libvulkan-stub/src
 515 11 145
autocast_mode.cpp
in aten/src/ATen
 590 20 142
BatchLinearAlgebra.cpp
in aten/src/ATen/native
 2819 116 133
library.cpp
in aten/src/ATen/native/quantized
 190 - 119
cuh
Math.cuh
in aten/src/ATen/native/cuda
 1509 - 105
vk_mem_alloc.h
in aten/src/ATen/native/vulkan/api
 12820 484 105
Math.h
in aten/src/ATen/native
 1620 30 100
ForeachOpsKernels.cpp
in aten/src/ATen/native
 214 2 75
mobile_bytecode_generated.h
in torch/csrc/jit/serialization
 2339 241 70
common_methods_invocations.py
in torch/testing/_internal
 12259 383 65
FunctionsManual.h
in torch/csrc/autograd
 440 2 52
cu
LossCTC.cu
in aten/src/ATen/native/cuda
 702 - 46
cu
ForeachPointwiseOp.cu
in aten/src/ATen/native/cuda
 172 - 44
FunctionsManual.cpp
in torch/csrc/autograd
 3181 188 43
cu
DistanceKernel.cu
in aten/src/ATen/native/cuda
 339 - 42
BatchLinearAlgebra.cpp
in aten/src/ATen/native/cuda
 2525 102 41