Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fix bug in LRP rules #92

Merged
merged 10 commits into from
Aug 17, 2022
Merged

Fix bug in LRP rules #92

merged 10 commits into from
Aug 17, 2022

Conversation

adrhill
Copy link
Member

@adrhill adrhill commented Aug 14, 2022 •
edited
Loading

Calling modify_layer! modifies pullback functions generated by Zygote.pullback. It is therefore necessary to directly compute (partial) relevances with pullback functions.

Additionally, the pullback needs to be computed for the pre-activations, not the layer output.

@adrhill adrhill added the run benchmark Run benchmark CI action label Aug 14, 2022
@codecov-commenter
Copy link

codecov-commenter commented Aug 14, 2022 •
edited
Loading

Codecov Report

Merging #92 (10cce32) into master (bc06035) will increase coverage by 1.91%.
The diff coverage is 94.23%.

@@            Coverage Diff             @@
##           master      #92      +/-   ##
==========================================
+ Coverage   93.53%   95.44%   +1.91%     
==========================================
  Files          16       16              
  Lines         541      571      +30     
==========================================
+ Hits          506      545      +39     
+ Misses         35       26       -9
Impacted Files Coverage Δ
src/flux_utils.jl 89.36% <80.00%> (-4.39%) ⬇️
src/lrp/rules.jl 97.26% <100.00%> (+0.31%) ⬆️
src/lrp/canonize.jl 96.55% <0.00%> (+3.44%) ⬆️
src/lrp/lrp.jl 96.42% <0.00%> (+3.57%) ⬆️
src/input_augmentation.jl 95.91% <0.00%> (+4.08%) ⬆️
src/gradient.jl 100.00% <0.00%> (+4.16%) ⬆️
src/lrp/show.jl 100.00% <0.00%> (+5.26%) ⬆️
src/lrp/composite.jl 100.00% <0.00%> (+7.14%) ⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

@github-actions
Copy link

Benchmark result

Judge result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmarks:
    • Target: 14 Aug 2022 - 18:45
    • Baseline: 14 Aug 2022 - 18:49
  • Package commits:
    • Target: 79606c
    • Baseline: bc0603
  • Julia commits:
    • Target: 742b9a
    • Baseline: 742b9a
  • Julia command flags:
    • Target: None
    • Baseline: None
  • Environment variables:
    • Target: None
    • Baseline: None

Results

A ratio greater than 1.0 denotes a possible regression (marked with ❌), while a ratio less
than 1.0 denotes a possible improvement (marked with ✅). Only significant results - results
that indicate possible regressions or improvements - are shown below (thus, an empty table means that all
benchmark results remained invariant between builds).

ID time ratio memory ratio
["Layer", "Conv", "AlphaBetaRule"] 1.02 (5%) 0.96 (1%) ✅
["Layer", "Dense", "AlphaBetaRule"] 1.06 (5%) ❌ 0.99 (1%)
["Layer", "Dense", "FlatRule"] 1.14 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "WSquareRule"] 1.15 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "ZBoxRule"] 1.22 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "ZeroRule"] 1.07 (5%) ❌ 1.00 (1%)
["VGG", "Gradient", "construct analyzer"] 0.82 (5%) ✅ 1.00 (1%)
["VGG", "InputTimesGradient", "analyze"] 0.06 (5%) ✅ 0.64 (1%) ✅
["VGG", "IntegratedGradients", "construct analyzer"] 0.76 (5%) ✅ 1.00 (1%)
["VGG", "SmoothGrad", "construct analyzer"] 0.94 (5%) ✅ 1.00 (1%)

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Target

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2095 MHz       2510 s          1 s        478 s       5483 s          0 s
       #2  2095 MHz       6086 s          2 s        580 s       1843 s          0 s
       
  Memory: 6.780967712402344 GB (648.4921875 MB free)
  Uptime: 854.91 sec
  Load Avg:  1.13  1.24  0.83
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2095 MHz       2969 s          1 s        579 s       7560 s          0 s
       #2  2095 MHz       8439 s          2 s        664 s       2054 s          0 s
       
  Memory: 6.780967712402344 GB (1508.734375 MB free)
  Uptime: 1119.86 sec
  Load Avg:  1.08  1.15  0.9
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Target result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmark: 14 Aug 2022 - 18:45
  • Package commit: 79606c
  • Julia commit: 742b9a
  • Julia command flags: None
  • Environment variables: None

Results

Below is a table of this job's results, obtained by running the benchmarks.
The values listed in the ID column have the structure [parent_group, child_group, ..., key], and can be used to
index into the BaseBenchmarks suite to retrieve the corresponding benchmarks.
The percentages accompanying time and memory values in the below table are noise tolerances. The "true"
time/memory value for a given benchmark is expected to fall within this percentage of the reported value.
An empty cell means that the value was zero.

ID time GC time memory allocations
["Layer", "Conv", "AlphaBetaRule"] 2.501 ms (5%) 5.65 MiB (1%) 572
["Layer", "Conv", "EpsilonRule"] 582.649 μs (5%) 1.36 MiB (1%) 131
["Layer", "Conv", "FlatRule"] 588.349 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "GammaRule"] 581.449 μs (5%) 1.36 MiB (1%) 137
["Layer", "Conv", "WSquareRule"] 590.649 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "ZBoxRule"] 1.818 ms (5%) 4.36 MiB (1%) 430
["Layer", "Conv", "ZeroRule"] 585.849 μs (5%) 1.36 MiB (1%) 131
["Layer", "Dense", "AlphaBetaRule"] 649.554 μs (5%) 1.80 MiB (1%) 143
["Layer", "Dense", "EpsilonRule"] 34.102 μs (5%) 1.45 KiB (1%) 3
["Layer", "Dense", "FlatRule"] 67.206 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "GammaRule"] 88.807 μs (5%) 197.95 KiB (1%) 10
["Layer", "Dense", "WSquareRule"] 90.508 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "ZBoxRule"] 407.134 μs (5%) 1.20 MiB (1%) 101
["Layer", "Dense", "ZeroRule"] 36.503 μs (5%) 1.45 KiB (1%) 3
["VGG", "Gradient", "analyze"] 1.162 s (5%) 13.397 ms 1.10 GiB (1%) 2012
["VGG", "Gradient", "construct analyzer"] 14.401 μs (5%) 26.66 KiB (1%) 189
["VGG", "InputTimesGradient", "analyze"] 1.163 s (5%) 12.980 ms 1.10 GiB (1%) 2015
["VGG", "InputTimesGradient", "construct analyzer"] 13.101 μs (5%) 26.66 KiB (1%) 189
["VGG", "IntegratedGradients", "analyze"] 10.187 s (5%) 92.545 ms 10.34 GiB (1%) 16944
["VGG", "IntegratedGradients", "construct analyzer"] 12.601 μs (5%) 24.69 KiB (1%) 187
["VGG", "LRP", "analyze"] 1.101 s (5%) 10.687 ms 761.54 MiB (1%) 1816
["VGG", "LRP", "construct analyzer"] 13.001 μs (5%) 44.86 KiB (1%) 83
["VGG", "SmoothGrad", "analyze"] 10.477 s (5%) 92.731 ms 10.33 GiB (1%) 16854
["VGG", "SmoothGrad", "construct analyzer"] 14.401 μs (5%) 24.69 KiB (1%) 187

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2095 MHz       2510 s          1 s        478 s       5483 s          0 s
       #2  2095 MHz       6086 s          2 s        580 s       1843 s          0 s
       
  Memory: 6.780967712402344 GB (648.4921875 MB free)
  Uptime: 854.91 sec
  Load Avg:  1.13  1.24  0.83
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmark: 14 Aug 2022 - 18:49
  • Package commit: bc0603
  • Julia commit: 742b9a
  • Julia command flags: None
  • Environment variables: None

Results

Below is a table of this job's results, obtained by running the benchmarks.
The values listed in the ID column have the structure [parent_group, child_group, ..., key], and can be used to
index into the BaseBenchmarks suite to retrieve the corresponding benchmarks.
The percentages accompanying time and memory values in the below table are noise tolerances. The "true"
time/memory value for a given benchmark is expected to fall within this percentage of the reported value.
An empty cell means that the value was zero.

ID time GC time memory allocations
["Layer", "Conv", "AlphaBetaRule"] 2.450 ms (5%) 5.87 MiB (1%) 590
["Layer", "Conv", "EpsilonRule"] 564.847 μs (5%) 1.36 MiB (1%) 131
["Layer", "Conv", "FlatRule"] 573.648 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "GammaRule"] 566.747 μs (5%) 1.36 MiB (1%) 137
["Layer", "Conv", "WSquareRule"] 568.848 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "ZBoxRule"] 1.782 ms (5%) 4.36 MiB (1%) 430
["Layer", "Conv", "ZeroRule"] 564.347 μs (5%) 1.36 MiB (1%) 131
["Layer", "Dense", "AlphaBetaRule"] 615.452 μs (5%) 1.81 MiB (1%) 155
["Layer", "Dense", "EpsilonRule"] 34.103 μs (5%) 1.45 KiB (1%) 3
["Layer", "Dense", "FlatRule"] 58.905 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "GammaRule"] 87.907 μs (5%) 197.95 KiB (1%) 10
["Layer", "Dense", "WSquareRule"] 78.806 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "ZBoxRule"] 334.828 μs (5%) 1.20 MiB (1%) 101
["Layer", "Dense", "ZeroRule"] 34.202 μs (5%) 1.45 KiB (1%) 3
["VGG", "Gradient", "analyze"] 1.144 s (5%) 11.557 ms 1.10 GiB (1%) 2012
["VGG", "Gradient", "construct analyzer"] 17.502 μs (5%) 26.66 KiB (1%) 189
["VGG", "InputTimesGradient", "analyze"] 19.689 s (5%) 275.300 ms 1.74 GiB (1%) 12029327
["VGG", "InputTimesGradient", "construct analyzer"] 13.402 μs (5%) 26.66 KiB (1%) 189
["VGG", "IntegratedGradients", "analyze"] 10.511 s (5%) 142.850 ms 10.34 GiB (1%) 16944
["VGG", "IntegratedGradients", "construct analyzer"] 16.501 μs (5%) 24.69 KiB (1%) 187
["VGG", "LRP", "analyze"] 1.132 s (5%) 34.307 ms 761.54 MiB (1%) 1816
["VGG", "LRP", "construct analyzer"] 12.801 μs (5%) 44.86 KiB (1%) 83
["VGG", "SmoothGrad", "analyze"] 11.003 s (5%) 160.046 ms 10.42 GiB (1%) 1464280
["VGG", "SmoothGrad", "construct analyzer"] 15.301 μs (5%) 24.69 KiB (1%) 187

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2095 MHz       2969 s          1 s        579 s       7560 s          0 s
       #2  2095 MHz       8439 s          2 s        664 s       2054 s          0 s
       
  Memory: 6.780967712402344 GB (1508.734375 MB free)
  Uptime: 1119.86 sec
  Load Avg:  1.08  1.15  0.9
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Runtime information

Runtime Info
BLAS #threads 2
BLAS.vendor() openblas64
Sys.CPU_THREADS 2

lscpu output:

Architecture:                    x86_64
CPU op-mode(s):                  32-bit, 64-bit
Byte Order:                      Little Endian
Address sizes:                   46 bits physical, 48 bits virtual
CPU(s):                          2
On-line CPU(s) list:             0,1
Thread(s) per core:              1
Core(s) per socket:              2
Socket(s):                       1
NUMA node(s):                    1
Vendor ID:                       GenuineIntel
CPU family:                      6
Model:                           85
Model name:                      Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz
Stepping:                        4
CPU MHz:                         2095.214
BogoMIPS:                        4190.42
Hypervisor vendor:               Microsoft
Virtualization type:             full
L1d cache:                       64 KiB
L1i cache:                       64 KiB
L2 cache:                        2 MiB
L3 cache:                        35.8 MiB
NUMA node0 CPU(s):               0,1
Vulnerability Itlb multihit:     KVM: Mitigation: VMX unsupported
Vulnerability L1tf:              Mitigation; PTE Inversion
Vulnerability Mds:               Mitigation; Clear CPU buffers; SMT Host state unknown
Vulnerability Meltdown:          Mitigation; PTI
Vulnerability Mmio stale data:   Vulnerable: Clear CPU buffers attempted, no microcode; SMT Host state unknown
Vulnerability Spec store bypass: Vulnerable
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Retpolines, STIBP disabled, RSB filling
Vulnerability Srbds:             Not affected
Vulnerability Tsx async abort:   Mitigation; Clear CPU buffers; SMT Host state unknown
Flags:                           fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology cpuid pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single pti fsgsbase bmi1 hle avx2 smep bmi2 erms invpcid rtm mpx avx512f avx512dq rdseed adx smap clflushopt avx512cd avx512bw avx512vl xsaveopt xsavec xsaves md_clear
Cpu Property Value
Brand Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz
Vendor :Intel
Architecture :Skylake
Model Family: 0x06, Model: 0x55, Stepping: 0x04, Type: 0x00
Cores 2 physical cores, 2 logical cores (on executing CPU)
No Hyperthreading hardware capability detected
Clock Frequencies Not supported by CPU
Data Cache Level 1:3 : (32, 1024, 36608) kbytes
64 byte cache line size
Address Size 48 bits virtual, 46 bits physical
SIMD 512 bit = 64 byte max. SIMD vector size
Time Stamp Counter TSC is accessible via rdtsc
TSC increased at every clock cycle (non-invariant TSC)
Perf. Monitoring Performance Monitoring Counters (PMC) are not supported
Hypervisor Yes, Microsoft

@github-actions
Copy link

Benchmark result

Judge result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmarks:
    • Target: 14 Aug 2022 - 18:45
    • Baseline: 14 Aug 2022 - 18:50
  • Package commits:
    • Target: 79606c
    • Baseline: bc0603
  • Julia commits:
    • Target: 742b9a
    • Baseline: 742b9a
  • Julia command flags:
    • Target: None
    • Baseline: None
  • Environment variables:
    • Target: None
    • Baseline: None

Results

A ratio greater than 1.0 denotes a possible regression (marked with ❌), while a ratio less
than 1.0 denotes a possible improvement (marked with ✅). Only significant results - results
that indicate possible regressions or improvements - are shown below (thus, an empty table means that all
benchmark results remained invariant between builds).

ID time ratio memory ratio
["Layer", "Conv", "AlphaBetaRule"] 1.01 (5%) 0.96 (1%) ✅
["Layer", "Conv", "EpsilonRule"] 1.07 (5%) ❌ 1.00 (1%)
["Layer", "Conv", "FlatRule"] 1.08 (5%) ❌ 1.00 (1%)
["Layer", "Conv", "WSquareRule"] 1.05 (5%) ❌ 1.00 (1%)
["Layer", "Conv", "ZBoxRule"] 1.10 (5%) ❌ 1.00 (1%)
["Layer", "Conv", "ZeroRule"] 1.09 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "FlatRule"] 1.06 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "GammaRule"] 1.24 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "ZeroRule"] 0.93 (5%) ✅ 1.00 (1%)
["VGG", "InputTimesGradient", "analyze"] 0.06 (5%) ✅ 0.64 (1%) ✅
["VGG", "InputTimesGradient", "construct analyzer"] 1.12 (5%) ❌ 1.00 (1%)
["VGG", "IntegratedGradients", "construct analyzer"] 1.11 (5%) ❌ 1.00 (1%)
["VGG", "LRP", "construct analyzer"] 1.05 (5%) ❌ 1.00 (1%)
["VGG", "SmoothGrad", "analyze"] 0.93 (5%) ✅ 0.99 (1%)
["VGG", "SmoothGrad", "construct analyzer"] 1.25 (5%) ❌ 1.00 (1%)

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Target

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2095 MHz       4120 s          2 s        577 s       3604 s          0 s
       #2  2095 MHz       4690 s          1 s        484 s       3151 s          0 s
       
  Memory: 6.780967712402344 GB (390.484375 MB free)
  Uptime: 836.42 sec
  Load Avg:  1.13  1.28  0.9
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2095 MHz       4457 s          2 s        672 s       5866 s          0 s
       #2  2095 MHz       7220 s          1 s        571 s       3237 s          0 s
       
  Memory: 6.780967712402344 GB (1615.421875 MB free)
  Uptime: 1106.68 sec
  Load Avg:  1.1  1.19  0.97
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Target result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmark: 14 Aug 2022 - 18:45
  • Package commit: 79606c
  • Julia commit: 742b9a
  • Julia command flags: None
  • Environment variables: None

Results

Below is a table of this job's results, obtained by running the benchmarks.
The values listed in the ID column have the structure [parent_group, child_group, ..., key], and can be used to
index into the BaseBenchmarks suite to retrieve the corresponding benchmarks.
The percentages accompanying time and memory values in the below table are noise tolerances. The "true"
time/memory value for a given benchmark is expected to fall within this percentage of the reported value.
An empty cell means that the value was zero.

ID time GC time memory allocations
["Layer", "Conv", "AlphaBetaRule"] 2.831 ms (5%) 5.65 MiB (1%) 572
["Layer", "Conv", "EpsilonRule"] 638.758 μs (5%) 1.36 MiB (1%) 131
["Layer", "Conv", "FlatRule"] 654.559 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "GammaRule"] 614.956 μs (5%) 1.36 MiB (1%) 137
["Layer", "Conv", "WSquareRule"] 649.159 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "ZBoxRule"] 2.121 ms (5%) 4.36 MiB (1%) 430
["Layer", "Conv", "ZeroRule"] 649.159 μs (5%) 1.36 MiB (1%) 131
["Layer", "Dense", "AlphaBetaRule"] 703.264 μs (5%) 1.80 MiB (1%) 143
["Layer", "Dense", "EpsilonRule"] 39.304 μs (5%) 1.45 KiB (1%) 3
["Layer", "Dense", "FlatRule"] 70.507 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "GammaRule"] 114.310 μs (5%) 197.95 KiB (1%) 10
["Layer", "Dense", "WSquareRule"] 92.408 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "ZBoxRule"] 443.640 μs (5%) 1.20 MiB (1%) 101
["Layer", "Dense", "ZeroRule"] 39.403 μs (5%) 1.45 KiB (1%) 3
["VGG", "Gradient", "analyze"] 1.181 s (5%) 11.838 ms 1.10 GiB (1%) 2012
["VGG", "Gradient", "construct analyzer"] 18.401 μs (5%) 26.66 KiB (1%) 189
["VGG", "InputTimesGradient", "analyze"] 1.176 s (5%) 12.129 ms 1.10 GiB (1%) 2015
["VGG", "InputTimesGradient", "construct analyzer"] 14.301 μs (5%) 26.66 KiB (1%) 189
["VGG", "IntegratedGradients", "analyze"] 10.391 s (5%) 75.013 ms 10.34 GiB (1%) 16944
["VGG", "IntegratedGradients", "construct analyzer"] 16.702 μs (5%) 24.69 KiB (1%) 187
["VGG", "LRP", "analyze"] 1.147 s (5%) 8.870 ms 761.54 MiB (1%) 1816
["VGG", "LRP", "construct analyzer"] 12.202 μs (5%) 44.86 KiB (1%) 83
["VGG", "SmoothGrad", "analyze"] 10.561 s (5%) 85.750 ms 10.33 GiB (1%) 16854
["VGG", "SmoothGrad", "construct analyzer"] 14.501 μs (5%) 24.69 KiB (1%) 187

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2095 MHz       4120 s          2 s        577 s       3604 s          0 s
       #2  2095 MHz       4690 s          1 s        484 s       3151 s          0 s
       
  Memory: 6.780967712402344 GB (390.484375 MB free)
  Uptime: 836.42 sec
  Load Avg:  1.13  1.28  0.9
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmark: 14 Aug 2022 - 18:50
  • Package commit: bc0603
  • Julia commit: 742b9a
  • Julia command flags: None
  • Environment variables: None

Results

Below is a table of this job's results, obtained by running the benchmarks.
The values listed in the ID column have the structure [parent_group, child_group, ..., key], and can be used to
index into the BaseBenchmarks suite to retrieve the corresponding benchmarks.
The percentages accompanying time and memory values in the below table are noise tolerances. The "true"
time/memory value for a given benchmark is expected to fall within this percentage of the reported value.
An empty cell means that the value was zero.

ID time GC time memory allocations
["Layer", "Conv", "AlphaBetaRule"] 2.803 ms (5%) 5.87 MiB (1%) 590
["Layer", "Conv", "EpsilonRule"] 594.453 μs (5%) 1.36 MiB (1%) 131
["Layer", "Conv", "FlatRule"] 607.855 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "GammaRule"] 630.856 μs (5%) 1.36 MiB (1%) 137
["Layer", "Conv", "WSquareRule"] 616.155 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "ZBoxRule"] 1.922 ms (5%) 4.36 MiB (1%) 430
["Layer", "Conv", "ZeroRule"] 596.554 μs (5%) 1.36 MiB (1%) 131
["Layer", "Dense", "AlphaBetaRule"] 681.361 μs (5%) 1.81 MiB (1%) 155
["Layer", "Dense", "EpsilonRule"] 39.603 μs (5%) 1.45 KiB (1%) 3
["Layer", "Dense", "FlatRule"] 66.706 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "GammaRule"] 92.008 μs (5%) 197.95 KiB (1%) 10
["Layer", "Dense", "WSquareRule"] 90.508 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "ZBoxRule"] 442.640 μs (5%) 1.20 MiB (1%) 101
["Layer", "Dense", "ZeroRule"] 42.203 μs (5%) 1.45 KiB (1%) 3
["VGG", "Gradient", "analyze"] 1.151 s (5%) 11.173 ms 1.10 GiB (1%) 2012
["VGG", "Gradient", "construct analyzer"] 18.402 μs (5%) 26.66 KiB (1%) 189
["VGG", "InputTimesGradient", "analyze"] 20.301 s (5%) 243.428 ms 1.74 GiB (1%) 12025699
["VGG", "InputTimesGradient", "construct analyzer"] 12.801 μs (5%) 26.66 KiB (1%) 189
["VGG", "IntegratedGradients", "analyze"] 10.270 s (5%) 143.062 ms 10.34 GiB (1%) 16944
["VGG", "IntegratedGradients", "construct analyzer"] 15.002 μs (5%) 24.69 KiB (1%) 187
["VGG", "LRP", "analyze"] 1.153 s (5%) 33.240 ms 761.54 MiB (1%) 1816
["VGG", "LRP", "construct analyzer"] 11.601 μs (5%) 44.86 KiB (1%) 83
["VGG", "SmoothGrad", "analyze"] 11.321 s (5%) 157.460 ms 10.42 GiB (1%) 1456740
["VGG", "SmoothGrad", "construct analyzer"] 11.601 μs (5%) 24.69 KiB (1%) 187

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2095 MHz       4457 s          2 s        672 s       5866 s          0 s
       #2  2095 MHz       7220 s          1 s        571 s       3237 s          0 s
       
  Memory: 6.780967712402344 GB (1615.421875 MB free)
  Uptime: 1106.68 sec
  Load Avg:  1.1  1.19  0.97
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Runtime information

Runtime Info
BLAS #threads 2
BLAS.vendor() openblas64
Sys.CPU_THREADS 2

lscpu output:

Architecture:                    x86_64
CPU op-mode(s):                  32-bit, 64-bit
Byte Order:                      Little Endian
Address sizes:                   46 bits physical, 48 bits virtual
CPU(s):                          2
On-line CPU(s) list:             0,1
Thread(s) per core:              1
Core(s) per socket:              2
Socket(s):                       1
NUMA node(s):                    1
Vendor ID:                       GenuineIntel
CPU family:                      6
Model:                           85
Model name:                      Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz
Stepping:                        4
CPU MHz:                         2095.214
BogoMIPS:                        4190.42
Hypervisor vendor:               Microsoft
Virtualization type:             full
L1d cache:                       64 KiB
L1i cache:                       64 KiB
L2 cache:                        2 MiB
L3 cache:                        35.8 MiB
NUMA node0 CPU(s):               0,1
Vulnerability Itlb multihit:     KVM: Mitigation: VMX unsupported
Vulnerability L1tf:              Mitigation; PTE Inversion
Vulnerability Mds:               Mitigation; Clear CPU buffers; SMT Host state unknown
Vulnerability Meltdown:          Mitigation; PTI
Vulnerability Mmio stale data:   Vulnerable: Clear CPU buffers attempted, no microcode; SMT Host state unknown
Vulnerability Spec store bypass: Vulnerable
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Retpolines, STIBP disabled, RSB filling
Vulnerability Srbds:             Not affected
Vulnerability Tsx async abort:   Mitigation; Clear CPU buffers; SMT Host state unknown
Flags:                           fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology cpuid pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single pti fsgsbase bmi1 hle avx2 smep bmi2 erms invpcid rtm avx512f avx512dq rdseed adx smap clflushopt avx512cd avx512bw avx512vl xsaveopt xsavec xsaves md_clear
Cpu Property Value
Brand Intel(R) Xeon(R) Platinum 8171M CPU @ 2.60GHz
Vendor :Intel
Architecture :Skylake
Model Family: 0x06, Model: 0x55, Stepping: 0x04, Type: 0x00
Cores 2 physical cores, 2 logical cores (on executing CPU)
No Hyperthreading hardware capability detected
Clock Frequencies Not supported by CPU
Data Cache Level 1:3 : (32, 1024, 36608) kbytes
64 byte cache line size
Address Size 48 bits virtual, 46 bits physical
SIMD 512 bit = 64 byte max. SIMD vector size
Time Stamp Counter TSC is accessible via rdtsc
TSC increased at every clock cycle (non-invariant TSC)
Perf. Monitoring Performance Monitoring Counters (PMC) are not supported
Hypervisor Yes, Microsoft

@adrhill adrhill removed the run benchmark Run benchmark CI action label Aug 14, 2022
@adrhill adrhill added the run benchmark Run benchmark CI action label Aug 15, 2022
@github-actions
Copy link

Benchmark result

Judge result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmarks:
    • Target: 15 Aug 2022 - 14:18
    • Baseline: 15 Aug 2022 - 14:22
  • Package commits:
    • Target: 79606c
    • Baseline: bc0603
  • Julia commits:
    • Target: 742b9a
    • Baseline: 742b9a
  • Julia command flags:
    • Target: None
    • Baseline: None
  • Environment variables:
    • Target: None
    • Baseline: None

Results

A ratio greater than 1.0 denotes a possible regression (marked with ❌), while a ratio less
than 1.0 denotes a possible improvement (marked with ✅). Only significant results - results
that indicate possible regressions or improvements - are shown below (thus, an empty table means that all
benchmark results remained invariant between builds).

ID time ratio memory ratio
["Layer", "Conv", "AlphaBetaRule"] 1.02 (5%) 0.96 (1%) ✅
["VGG", "Gradient", "construct analyzer"] 0.94 (5%) ✅ 1.00 (1%)
["VGG", "InputTimesGradient", "analyze"] 0.12 (5%) ✅ 0.64 (1%) ✅
["VGG", "InputTimesGradient", "construct analyzer"] 1.18 (5%) ❌ 1.00 (1%)
["VGG", "IntegratedGradients", "construct analyzer"] 0.92 (5%) ✅ 1.00 (1%)
["VGG", "SmoothGrad", "construct analyzer"] 0.89 (5%) ✅ 1.00 (1%)

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Target

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8370C CPU @ 2.80GHz: 
              speed         user         nice          sys         idle          irq
       #1  2793 MHz       5501 s          1 s        369 s       1740 s          0 s
       #2  2793 MHz       4252 s          1 s        413 s       2957 s          0 s
       
  Memory: 6.780967712402344 GB (1191.51953125 MB free)
  Uptime: 765.03 sec
  Load Avg:  1.24  1.35  0.91
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, icelake-server)

Baseline

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8370C CPU @ 2.80GHz: 
              speed         user         nice          sys         idle          irq
       #1  2793 MHz       8082 s          1 s        434 s       1795 s          0 s
       #2  2793 MHz       4890 s          1 s        487 s       4948 s          0 s
       
  Memory: 6.780967712402344 GB (1447.109375 MB free)
  Uptime: 1035.17 sec
  Load Avg:  1.23  1.3  1.0
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, icelake-server)

Target result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmark: 15 Aug 2022 - 14:18
  • Package commit: 79606c
  • Julia commit: 742b9a
  • Julia command flags: None
  • Environment variables: None

Results

Below is a table of this job's results, obtained by running the benchmarks.
The values listed in the ID column have the structure [parent_group, child_group, ..., key], and can be used to
index into the BaseBenchmarks suite to retrieve the corresponding benchmarks.
The percentages accompanying time and memory values in the below table are noise tolerances. The "true"
time/memory value for a given benchmark is expected to fall within this percentage of the reported value.
An empty cell means that the value was zero.

ID time GC time memory allocations
["Layer", "Conv", "AlphaBetaRule"] 2.613 ms (5%) 5.65 MiB (1%) 572
["Layer", "Conv", "EpsilonRule"] 625.899 μs (5%) 1.36 MiB (1%) 131
["Layer", "Conv", "FlatRule"] 629.499 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "GammaRule"] 625.000 μs (5%) 1.36 MiB (1%) 137
["Layer", "Conv", "WSquareRule"] 628.199 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "ZBoxRule"] 1.919 ms (5%) 4.36 MiB (1%) 430
["Layer", "Conv", "ZeroRule"] 624.600 μs (5%) 1.36 MiB (1%) 131
["Layer", "Dense", "AlphaBetaRule"] 741.399 μs (5%) 1.80 MiB (1%) 143
["Layer", "Dense", "EpsilonRule"] 49.699 μs (5%) 1.45 KiB (1%) 3
["Layer", "Dense", "FlatRule"] 74.700 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "GammaRule"] 111.300 μs (5%) 197.95 KiB (1%) 10
["Layer", "Dense", "WSquareRule"] 97.799 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "ZBoxRule"] 485.399 μs (5%) 1.20 MiB (1%) 101
["Layer", "Dense", "ZeroRule"] 49.600 μs (5%) 1.45 KiB (1%) 3
["VGG", "Gradient", "analyze"] 2.370 s (5%) 9.165 ms 1.10 GiB (1%) 2012
["VGG", "Gradient", "construct analyzer"] 13.400 μs (5%) 26.66 KiB (1%) 189
["VGG", "InputTimesGradient", "analyze"] 2.384 s (5%) 9.580 ms 1.10 GiB (1%) 2015
["VGG", "InputTimesGradient", "construct analyzer"] 12.700 μs (5%) 26.66 KiB (1%) 189
["VGG", "IntegratedGradients", "analyze"] 23.648 s (5%) 63.856 ms 10.34 GiB (1%) 16944
["VGG", "IntegratedGradients", "construct analyzer"] 12.600 μs (5%) 24.69 KiB (1%) 187
["VGG", "LRP", "analyze"] 2.292 s (5%) 3.611 ms 761.54 MiB (1%) 1816
["VGG", "LRP", "construct analyzer"] 12.500 μs (5%) 44.86 KiB (1%) 83
["VGG", "SmoothGrad", "analyze"] 23.695 s (5%) 66.614 ms 10.33 GiB (1%) 16854
["VGG", "SmoothGrad", "construct analyzer"] 11.000 μs (5%) 24.69 KiB (1%) 187

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8370C CPU @ 2.80GHz: 
              speed         user         nice          sys         idle          irq
       #1  2793 MHz       5501 s          1 s        369 s       1740 s          0 s
       #2  2793 MHz       4252 s          1 s        413 s       2957 s          0 s
       
  Memory: 6.780967712402344 GB (1191.51953125 MB free)
  Uptime: 765.03 sec
  Load Avg:  1.24  1.35  0.91
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, icelake-server)

Baseline result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmark: 15 Aug 2022 - 14:22
  • Package commit: bc0603
  • Julia commit: 742b9a
  • Julia command flags: None
  • Environment variables: None

Results

Below is a table of this job's results, obtained by running the benchmarks.
The values listed in the ID column have the structure [parent_group, child_group, ..., key], and can be used to
index into the BaseBenchmarks suite to retrieve the corresponding benchmarks.
The percentages accompanying time and memory values in the below table are noise tolerances. The "true"
time/memory value for a given benchmark is expected to fall within this percentage of the reported value.
An empty cell means that the value was zero.

ID time GC time memory allocations
["Layer", "Conv", "AlphaBetaRule"] 2.573 ms (5%) 5.87 MiB (1%) 590
["Layer", "Conv", "EpsilonRule"] 617.800 μs (5%) 1.36 MiB (1%) 131
["Layer", "Conv", "FlatRule"] 626.100 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "GammaRule"] 618.499 μs (5%) 1.36 MiB (1%) 137
["Layer", "Conv", "WSquareRule"] 624.700 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "ZBoxRule"] 1.896 ms (5%) 4.36 MiB (1%) 430
["Layer", "Conv", "ZeroRule"] 618.899 μs (5%) 1.36 MiB (1%) 131
["Layer", "Dense", "AlphaBetaRule"] 728.902 μs (5%) 1.81 MiB (1%) 155
["Layer", "Dense", "EpsilonRule"] 49.600 μs (5%) 1.45 KiB (1%) 3
["Layer", "Dense", "FlatRule"] 75.400 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "GammaRule"] 111.199 μs (5%) 197.95 KiB (1%) 10
["Layer", "Dense", "WSquareRule"] 96.900 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "ZBoxRule"] 480.499 μs (5%) 1.20 MiB (1%) 101
["Layer", "Dense", "ZeroRule"] 49.500 μs (5%) 1.45 KiB (1%) 3
["VGG", "Gradient", "analyze"] 2.369 s (5%) 12.370 ms 1.10 GiB (1%) 2012
["VGG", "Gradient", "construct analyzer"] 14.300 μs (5%) 26.66 KiB (1%) 189
["VGG", "InputTimesGradient", "analyze"] 19.243 s (5%) 222.373 ms 1.74 GiB (1%) 12023683
["VGG", "InputTimesGradient", "construct analyzer"] 10.800 μs (5%) 26.66 KiB (1%) 189
["VGG", "IntegratedGradients", "analyze"] 23.754 s (5%) 106.462 ms 10.34 GiB (1%) 16944
["VGG", "IntegratedGradients", "construct analyzer"] 13.700 μs (5%) 24.69 KiB (1%) 187
["VGG", "LRP", "analyze"] 2.297 s (5%) 23.561 ms 761.54 MiB (1%) 1816
["VGG", "LRP", "construct analyzer"] 12.700 μs (5%) 44.86 KiB (1%) 83
["VGG", "SmoothGrad", "analyze"] 24.447 s (5%) 115.755 ms 10.42 GiB (1%) 1452626
["VGG", "SmoothGrad", "construct analyzer"] 12.400 μs (5%) 24.69 KiB (1%) 187

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1014-azure #17~20.04.1-Ubuntu SMP Thu Jun 23 20:01:51 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8370C CPU @ 2.80GHz: 
              speed         user         nice          sys         idle          irq
       #1  2793 MHz       8082 s          1 s        434 s       1795 s          0 s
       #2  2793 MHz       4890 s          1 s        487 s       4948 s          0 s
       
  Memory: 6.780967712402344 GB (1447.109375 MB free)
  Uptime: 1035.17 sec
  Load Avg:  1.23  1.3  1.0
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, icelake-server)

Runtime information

Runtime Info
BLAS #threads 2
BLAS.vendor() openblas64
Sys.CPU_THREADS 2

lscpu output:

Architecture:                    x86_64
CPU op-mode(s):                  32-bit, 64-bit
Byte Order:                      Little Endian
Address sizes:                   46 bits physical, 48 bits virtual
CPU(s):                          2
On-line CPU(s) list:             0,1
Thread(s) per core:              1
Core(s) per socket:              2
Socket(s):                       1
NUMA node(s):                    1
Vendor ID:                       GenuineIntel
CPU family:                      6
Model:                           106
Model name:                      Intel(R) Xeon(R) Platinum 8370C CPU @ 2.80GHz
Stepping:                        6
CPU MHz:                         2793.438
BogoMIPS:                        5586.87
Hypervisor vendor:               Microsoft
Virtualization type:             full
L1d cache:                       96 KiB
L1i cache:                       64 KiB
L2 cache:                        2.5 MiB
L3 cache:                        48 MiB
NUMA node0 CPU(s):               0,1
Vulnerability Itlb multihit:     KVM: Mitigation: VMX unsupported
Vulnerability L1tf:              Mitigation; PTE Inversion
Vulnerability Mds:               Mitigation; Clear CPU buffers; SMT Host state unknown
Vulnerability Meltdown:          Mitigation; PTI
Vulnerability Mmio stale data:   Vulnerable: Clear CPU buffers attempted, no microcode; SMT Host state unknown
Vulnerability Spec store bypass: Vulnerable
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Retpolines, STIBP disabled, RSB filling
Vulnerability Srbds:             Not affected
Vulnerability Tsx async abort:   Mitigation; Clear CPU buffers; SMT Host state unknown
Flags:                           fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology cpuid pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single pti fsgsbase bmi1 hle avx2 smep bmi2 erms invpcid rtm avx512f avx512dq rdseed adx smap clflushopt avx512cd avx512bw avx512vl xsaveopt xsavec xsaves md_clear
Cpu Property Value
Brand Intel(R) Xeon(R) Platinum 8370C CPU @ 2.80GHz
Vendor :Intel
Architecture :UnknownIntel
Model Family: 0x06, Model: 0x6a, Stepping: 0x06, Type: 0x00
Cores 2 physical cores, 2 logical cores (on executing CPU)
No Hyperthreading hardware capability detected
Clock Frequencies Not supported by CPU
Data Cache Level 1:3 : (48, 1280, 49152) kbytes
64 byte cache line size
Address Size 48 bits virtual, 46 bits physical
SIMD 512 bit = 64 byte max. SIMD vector size
Time Stamp Counter TSC is accessible via rdtsc
TSC increased at every clock cycle (non-invariant TSC)
Perf. Monitoring Performance Monitoring Counters (PMC) are not supported
Hypervisor Yes, Microsoft

@github-actions
Copy link

Benchmark result

Judge result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmarks:
    • Target: 15 Aug 2022 - 14:18
    • Baseline: 15 Aug 2022 - 14:22
  • Package commits:
    • Target: 7423d8
    • Baseline: bc0603
  • Julia commits:
    • Target: 742b9a
    • Baseline: 742b9a
  • Julia command flags:
    • Target: None
    • Baseline: None
  • Environment variables:
    • Target: None
    • Baseline: None

Results

A ratio greater than 1.0 denotes a possible regression (marked with ❌), while a ratio less
than 1.0 denotes a possible improvement (marked with ✅). Only significant results - results
that indicate possible regressions or improvements - are shown below (thus, an empty table means that all
benchmark results remained invariant between builds).

ID time ratio memory ratio
["Layer", "Conv", "AlphaBetaRule"] 1.03 (5%) 0.95 (1%) ✅
["Layer", "Conv", "EpsilonRule"] 1.08 (5%) ❌ 1.00 (1%)
["Layer", "Conv", "ZBoxRule"] 1.02 (5%) 0.96 (1%) ✅
["Layer", "Dense", "AlphaBetaRule"] 1.24 (5%) ❌ 1.21 (1%) ❌
["Layer", "Dense", "ZBoxRule"] 1.17 (5%) ❌ 1.16 (1%) ❌
["VGG", "Gradient", "construct analyzer"] 0.80 (5%) ✅ 1.00 (1%)
["VGG", "InputTimesGradient", "analyze"] 0.06 (5%) ✅ 0.64 (1%) ✅
["VGG", "InputTimesGradient", "construct analyzer"] 1.25 (5%) ❌ 1.00 (1%)
["VGG", "IntegratedGradients", "construct analyzer"] 0.81 (5%) ✅ 1.00 (1%)
["VGG", "LRP", "analyze"] 0.94 (5%) ✅ 1.00 (1%)
["VGG", "LRP", "construct analyzer"] 1.23 (5%) ❌ 1.00 (1%)
["VGG", "SmoothGrad", "analyze"] 0.93 (5%) ✅ 0.99 (1%)
["VGG", "SmoothGrad", "construct analyzer"] 1.14 (5%) ❌ 1.00 (1%)

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Target

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1017-azure #20~20.04.1-Ubuntu SMP Fri Aug 5 12:16:53 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) CPU E5-2673 v3 @ 2.40GHz: 
              speed         user         nice          sys         idle          irq
       #1  2394 MHz       3564 s          0 s        419 s       3314 s          0 s
       #2  2394 MHz       5484 s          0 s        536 s       1317 s          0 s
       
  Memory: 6.781246185302734 GB (149.46875 MB free)
  Uptime: 739.72 sec
  Load Avg:  1.13  1.28  0.9
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, haswell)

Baseline

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1017-azure #20~20.04.1-Ubuntu SMP Fri Aug 5 12:16:53 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) CPU E5-2673 v3 @ 2.40GHz: 
              speed         user         nice          sys         idle          irq
       #1  2394 MHz       4245 s          0 s        526 s       5255 s          0 s
       #2  2394 MHz       7746 s          0 s        595 s       1729 s          0 s
       
  Memory: 6.781246185302734 GB (1741.74609375 MB free)
  Uptime: 1013.26 sec
  Load Avg:  1.09  1.18  0.96
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, haswell)

Target result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmark: 15 Aug 2022 - 14:18
  • Package commit: 7423d8
  • Julia commit: 742b9a
  • Julia command flags: None
  • Environment variables: None

Results

Below is a table of this job's results, obtained by running the benchmarks.
The values listed in the ID column have the structure [parent_group, child_group, ..., key], and can be used to
index into the BaseBenchmarks suite to retrieve the corresponding benchmarks.
The percentages accompanying time and memory values in the below table are noise tolerances. The "true"
time/memory value for a given benchmark is expected to fall within this percentage of the reported value.
An empty cell means that the value was zero.

ID time GC time memory allocations
["Layer", "Conv", "AlphaBetaRule"] 3.236 ms (5%) 5.59 MiB (1%) 582
["Layer", "Conv", "EpsilonRule"] 779.305 μs (5%) 1.36 MiB (1%) 131
["Layer", "Conv", "FlatRule"] 790.004 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "GammaRule"] 778.805 μs (5%) 1.36 MiB (1%) 137
["Layer", "Conv", "WSquareRule"] 788.505 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "ZBoxRule"] 2.380 ms (5%) 4.17 MiB (1%) 426
["Layer", "Conv", "ZeroRule"] 779.905 μs (5%) 1.36 MiB (1%) 131
["Layer", "Dense", "AlphaBetaRule"] 1.127 ms (5%) 2.20 MiB (1%) 161
["Layer", "Dense", "EpsilonRule"] 67.800 μs (5%) 1.45 KiB (1%) 3
["Layer", "Dense", "FlatRule"] 115.601 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "GammaRule"] 146.202 μs (5%) 197.95 KiB (1%) 10
["Layer", "Dense", "WSquareRule"] 140.301 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "ZBoxRule"] 721.432 μs (5%) 1.39 MiB (1%) 107
["Layer", "Dense", "ZeroRule"] 67.901 μs (5%) 1.45 KiB (1%) 3
["VGG", "Gradient", "analyze"] 1.246 s (5%) 16.357 ms 1.10 GiB (1%) 2012
["VGG", "Gradient", "construct analyzer"] 15.200 μs (5%) 26.66 KiB (1%) 189
["VGG", "InputTimesGradient", "analyze"] 1.246 s (5%) 17.366 ms 1.10 GiB (1%) 2015
["VGG", "InputTimesGradient", "construct analyzer"] 15.900 μs (5%) 26.66 KiB (1%) 189
["VGG", "IntegratedGradients", "analyze"] 12.096 s (5%) 124.144 ms 10.34 GiB (1%) 16944
["VGG", "IntegratedGradients", "construct analyzer"] 14.600 μs (5%) 24.69 KiB (1%) 187
["VGG", "LRP", "analyze"] 1.184 s (5%) 13.490 ms 761.54 MiB (1%) 1816
["VGG", "LRP", "construct analyzer"] 19.800 μs (5%) 44.86 KiB (1%) 83
["VGG", "SmoothGrad", "analyze"] 12.051 s (5%) 119.338 ms 10.33 GiB (1%) 16854
["VGG", "SmoothGrad", "construct analyzer"] 12.900 μs (5%) 24.69 KiB (1%) 187

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1017-azure #20~20.04.1-Ubuntu SMP Fri Aug 5 12:16:53 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) CPU E5-2673 v3 @ 2.40GHz: 
              speed         user         nice          sys         idle          irq
       #1  2394 MHz       3564 s          0 s        419 s       3314 s          0 s
       #2  2394 MHz       5484 s          0 s        536 s       1317 s          0 s
       
  Memory: 6.781246185302734 GB (149.46875 MB free)
  Uptime: 739.72 sec
  Load Avg:  1.13  1.28  0.9
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, haswell)

Baseline result

Benchmark Report for /home/runner/work/ExplainableAI.jl/ExplainableAI.jl

Job Properties

  • Time of benchmark: 15 Aug 2022 - 14:22
  • Package commit: bc0603
  • Julia commit: 742b9a
  • Julia command flags: None
  • Environment variables: None

Results

Below is a table of this job's results, obtained by running the benchmarks.
The values listed in the ID column have the structure [parent_group, child_group, ..., key], and can be used to
index into the BaseBenchmarks suite to retrieve the corresponding benchmarks.
The percentages accompanying time and memory values in the below table are noise tolerances. The "true"
time/memory value for a given benchmark is expected to fall within this percentage of the reported value.
An empty cell means that the value was zero.

ID time GC time memory allocations
["Layer", "Conv", "AlphaBetaRule"] 3.130 ms (5%) 5.87 MiB (1%) 590
["Layer", "Conv", "EpsilonRule"] 724.105 μs (5%) 1.36 MiB (1%) 131
["Layer", "Conv", "FlatRule"] 793.905 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "GammaRule"] 785.204 μs (5%) 1.36 MiB (1%) 137
["Layer", "Conv", "WSquareRule"] 788.109 μs (5%) 1.41 MiB (1%) 140
["Layer", "Conv", "ZBoxRule"] 2.340 ms (5%) 4.36 MiB (1%) 430
["Layer", "Conv", "ZeroRule"] 781.405 μs (5%) 1.36 MiB (1%) 131
["Layer", "Dense", "AlphaBetaRule"] 906.006 μs (5%) 1.81 MiB (1%) 155
["Layer", "Dense", "EpsilonRule"] 67.800 μs (5%) 1.45 KiB (1%) 3
["Layer", "Dense", "FlatRule"] 115.100 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "GammaRule"] 146.201 μs (5%) 197.95 KiB (1%) 10
["Layer", "Dense", "WSquareRule"] 135.900 μs (5%) 200.02 KiB (1%) 11
["Layer", "Dense", "ZBoxRule"] 615.804 μs (5%) 1.20 MiB (1%) 101
["Layer", "Dense", "ZeroRule"] 67.901 μs (5%) 1.45 KiB (1%) 3
["VGG", "Gradient", "analyze"] 1.264 s (5%) 14.889 ms 1.10 GiB (1%) 2012
["VGG", "Gradient", "construct analyzer"] 19.000 μs (5%) 26.66 KiB (1%) 189
["VGG", "InputTimesGradient", "analyze"] 20.173 s (5%) 265.158 ms 1.74 GiB (1%) 12026008
["VGG", "InputTimesGradient", "construct analyzer"] 12.700 μs (5%) 26.66 KiB (1%) 189
["VGG", "IntegratedGradients", "analyze"] 12.157 s (5%) 178.227 ms 10.34 GiB (1%) 16944
["VGG", "IntegratedGradients", "construct analyzer"] 18.000 μs (5%) 24.69 KiB (1%) 187
["VGG", "LRP", "analyze"] 1.258 s (5%) 47.851 ms 761.54 MiB (1%) 1816
["VGG", "LRP", "construct analyzer"] 16.100 μs (5%) 44.86 KiB (1%) 83
["VGG", "SmoothGrad", "analyze"] 12.957 s (5%) 193.617 ms 10.42 GiB (1%) 1457276
["VGG", "SmoothGrad", "construct analyzer"] 11.300 μs (5%) 24.69 KiB (1%) 187

Benchmark Group List

Here's a list of all the benchmark groups executed by this job:

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "IntegratedGradients"]
  • ["VGG", "LRP"]
  • ["VGG", "SmoothGrad"]

Julia versioninfo

Julia Version 1.7.3
Commit 742b9abb4d (2022-05-06 12:58 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.15.0-1017-azure #20~20.04.1-Ubuntu SMP Fri Aug 5 12:16:53 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) CPU E5-2673 v3 @ 2.40GHz: 
              speed         user         nice          sys         idle          irq
       #1  2394 MHz       4245 s          0 s        526 s       5255 s          0 s
       #2  2394 MHz       7746 s          0 s        595 s       1729 s          0 s
       
  Memory: 6.781246185302734 GB (1741.74609375 MB free)
  Uptime: 1013.26 sec
  Load Avg:  1.09  1.18  0.96
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, haswell)

Runtime information

Runtime Info
BLAS #threads 2
BLAS.vendor() openblas64
Sys.CPU_THREADS 2

lscpu output:

Architecture:                    x86_64
CPU op-mode(s):                  32-bit, 64-bit
Byte Order:                      Little Endian
Address sizes:                   46 bits physical, 48 bits virtual
CPU(s):                          2
On-line CPU(s) list:             0,1
Thread(s) per core:              1
Core(s) per socket:              2
Socket(s):                       1
NUMA node(s):                    1
Vendor ID:                       GenuineIntel
CPU family:                      6
Model:                           63
Model name:                      Intel(R) Xeon(R) CPU E5-2673 v3 @ 2.40GHz
Stepping:                        2
CPU MHz:                         2394.455
BogoMIPS:                        4788.91
Hypervisor vendor:               Microsoft
Virtualization type:             full
L1d cache:                       64 KiB
L1i cache:                       64 KiB
L2 cache:                        512 KiB
L3 cache:                        30 MiB
NUMA node0 CPU(s):               0,1
Vulnerability Itlb multihit:     KVM: Mitigation: VMX unsupported
Vulnerability L1tf:              Mitigation; PTE Inversion
Vulnerability Mds:               Mitigation; Clear CPU buffers; SMT Host state unknown
Vulnerability Meltdown:          Mitigation; PTI
Vulnerability Mmio stale data:   Vulnerable: Clear CPU buffers attempted, no microcode; SMT Host state unknown
Vulnerability Retbleed:          Not affected
Vulnerability Spec store bypass: Vulnerable
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Retpolines, STIBP disabled, RSB filling
Vulnerability Srbds:             Not affected
Vulnerability Tsx async abort:   Not affected
Flags:                           fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology cpuid pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm invpcid_single pti fsgsbase bmi1 avx2 smep bmi2 erms invpcid xsaveopt md_clear
Cpu Property Value
Brand Intel(R) Xeon(R) CPU E5-2673 v3 @ 2.40GHz
Vendor :Intel
Architecture :Haswell
Model Family: 0x06, Model: 0x3f, Stepping: 0x02, Type: 0x00
Cores 2 physical cores, 2 logical cores (on executing CPU)
No Hyperthreading hardware capability detected
Clock Frequencies Not supported by CPU
Data Cache Level 1:3 : (32, 256, 30720) kbytes
64 byte cache line size
Address Size 48 bits virtual, 46 bits physical
SIMD 256 bit = 32 byte max. SIMD vector size
Time Stamp Counter TSC is accessible via rdtsc
TSC increased at every clock cycle (non-invariant TSC)
Perf. Monitoring Performance Monitoring Counters (PMC) are not supported
Hypervisor Yes, Microsoft

@adrhill adrhill removed the run benchmark Run benchmark CI action label Aug 15, 2022
@adrhill adrhill changed the title Fix bug in AlphaBetaRule, ZPlusRule & ZBoxRule Fix bug in LRP rules Aug 16, 2022
@adrhill adrhill merged commit 239deae into master Aug 17, 2022
@adrhill adrhill deleted the ah/fix-alphabeta branch August 17, 2022 08:43
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@adrhill @codecov-commenter