Skip to content

Add inplace updating lrp! rule calls and reuse gradient code #38

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
Merged
merged 6 commits into from
Mar 16, 2022
Merged

Add inplace updating lrp! rule calls and reuse gradient code #38

merged 6 commits into from
Mar 16, 2022

Conversation

adrhill
Copy link
Member

@adrhill adrhill commented Mar 16, 2022

No description provided.

@adrhill adrhill added the run benchmark Run benchmark CI action label Mar 16, 2022
@github-actions
Copy link

Benchmark result

Judge result

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

Job Properties

  • Time of benchmarks:
    • Target: 16 Mar 2022 - 15:04
    • Baseline: 16 Mar 2022 - 15:08
  • Package commits:
    • Target: baa836
    • Baseline: c778bb
  • Julia commits:
    • Target: bf5349
    • Baseline: bf5349
  • 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", "EpsilonRule"] 1.08 (5%) ❌ 1.00 (1%)
["Layer", "Conv", "GammaRule"] 1.11 (5%) ❌ 1.00 (1%)
["Layer", "Conv", "ZBoxRule"] 1.07 (5%) ❌ 1.00 (1%)
["Layer", "Conv", "ZeroRule"] 1.06 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "EpsilonRule"] 1.08 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "ZeroRule"] 1.12 (5%) ❌ 1.00 (1%)
["Layer", "MaxPool", "EpsilonRule"] 1.32 (5%) ❌ 1.00 (1%)
["Layer", "MaxPool", "GammaRule"] 1.21 (5%) ❌ 1.00 (1%)
["Layer", "MaxPool", "ZBoxRule"] 1.31 (5%) ❌ 1.00 (1%)
["Layer", "MaxPool", "ZeroRule"] 1.06 (5%) ❌ 1.00 (1%)
["Layer", "WrappedDense", "EpsilonRule"] 1.10 (5%) ❌ 1.00 (1%)
["Layer", "WrappedDense", "GammaRule"] 1.08 (5%) ❌ 1.00 (1%)
["VGG", "Gradient", "analyze"] 0.18 (5%) ✅ 0.99 (1%) ✅
["VGG", "Gradient", "construct analyzer"] 0.75 (5%) ✅ 1.00 (1%)
["VGG", "InputTimesGradient", "analyze"] 0.04 (5%) ✅ 0.58 (1%) ✅
["VGG", "InputTimesGradient", "construct analyzer"] 1.32 (5%) ❌ 1.00 (1%)
["VGG", "LRPCustom", "analyze"] 0.20 (5%) ✅ 0.78 (1%) ✅
["VGG", "LRPCustom", "construct analyzer"] 0.52 (5%) ✅ 1.00 (1%)

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Target

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 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       2318 s          2 s        548 s       5025 s          0 s
       #2  2095 MHz       5243 s          0 s        444 s       2259 s          0 s
       
  Memory: 6.7845458984375 GB (1245.34375 MB free)
  Uptime: 797.78 sec
  Load Avg:  1.28  1.29  0.81
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 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       3962 s          2 s        652 s       5661 s          0 s
       #2  2095 MHz       6098 s          0 s        513 s       3715 s          0 s
       
  Memory: 6.7845458984375 GB (1416.875 MB free)
  Uptime: 1036.66 sec
  Load Avg:  1.34  1.25  0.91
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Target result

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

Job Properties

  • Time of benchmark: 16 Mar 2022 - 15:4
  • Package commit: baa836
  • Julia commit: bf5349
  • 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", "EpsilonRule"] 782.851 μs (5%) 1.59 MiB (1%) 364
["Layer", "Conv", "GammaRule"] 806.851 μs (5%) 1.59 MiB (1%) 376
["Layer", "Conv", "ZBoxRule"] 2.442 ms (5%) 4.77 MiB (1%) 1047
["Layer", "Conv", "ZeroRule"] 766.350 μs (5%) 1.59 MiB (1%) 364
["Layer", "Dense", "EpsilonRule"] 21.001 μs (5%) 5.08 KiB (1%) 9
["Layer", "Dense", "GammaRule"] 95.506 μs (5%) 592.61 KiB (1%) 18
["Layer", "Dense", "ZBoxRule"] 248.416 μs (5%) 1.01 MiB (1%) 413
["Layer", "Dense", "ZeroRule"] 21.601 μs (5%) 5.08 KiB (1%) 9
["Layer", "MaxPool", "EpsilonRule"] 98.906 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "GammaRule"] 99.106 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "ZBoxRule"] 97.907 μs (5%) 154.62 KiB (1%) 73
["Layer", "MaxPool", "ZeroRule"] 87.306 μs (5%) 154.62 KiB (1%) 73
["Layer", "WrappedDense", "EpsilonRule"] 53.503 μs (5%) 213.72 KiB (1%) 153
["Layer", "WrappedDense", "GammaRule"] 145.410 μs (5%) 801.25 KiB (1%) 162
["Layer", "WrappedDense", "ZBoxRule"] 243.816 μs (5%) 1.01 MiB (1%) 413
["Layer", "WrappedDense", "ZeroRule"] 53.503 μs (5%) 213.11 KiB (1%) 153
["VGG", "Gradient", "analyze"] 1.165 s (5%) 13.500 ms 1.10 GiB (1%) 3697
["VGG", "Gradient", "construct analyzer"] 15.401 μs (5%) 27.41 KiB (1%) 205
["VGG", "InputTimesGradient", "analyze"] 1.160 s (5%) 13.004 ms 1.10 GiB (1%) 3700
["VGG", "InputTimesGradient", "construct analyzer"] 15.201 μs (5%) 27.41 KiB (1%) 205
["VGG", "LRPCustom", "analyze"] 2.041 s (5%) 25.139 ms 2.68 GiB (1%) 4779
["VGG", "LRPCustom", "construct analyzer"] 20.801 μs (5%) 45.52 KiB (1%) 155
["VGG", "LRPZero", "analyze"] 1.232 s (5%) 10.544 ms 968.64 MiB (1%) 3973
["VGG", "LRPZero", "construct analyzer"] 13.901 μs (5%) 43.88 KiB (1%) 82

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 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       2318 s          2 s        548 s       5025 s          0 s
       #2  2095 MHz       5243 s          0 s        444 s       2259 s          0 s
       
  Memory: 6.7845458984375 GB (1245.34375 MB free)
  Uptime: 797.78 sec
  Load Avg:  1.28  1.29  0.81
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline result

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

Job Properties

  • Time of benchmark: 16 Mar 2022 - 15:8
  • Package commit: c778bb
  • Julia commit: bf5349
  • 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", "EpsilonRule"] 722.046 μs (5%) 1.59 MiB (1%) 364
["Layer", "Conv", "GammaRule"] 726.747 μs (5%) 1.59 MiB (1%) 376
["Layer", "Conv", "ZBoxRule"] 2.284 ms (5%) 4.77 MiB (1%) 1047
["Layer", "Conv", "ZeroRule"] 721.947 μs (5%) 1.59 MiB (1%) 364
["Layer", "Dense", "EpsilonRule"] 19.501 μs (5%) 5.08 KiB (1%) 9
["Layer", "Dense", "GammaRule"] 92.206 μs (5%) 592.61 KiB (1%) 18
["Layer", "Dense", "ZBoxRule"] 236.715 μs (5%) 1.01 MiB (1%) 413
["Layer", "Dense", "ZeroRule"] 19.302 μs (5%) 5.08 KiB (1%) 9
["Layer", "MaxPool", "EpsilonRule"] 74.805 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "GammaRule"] 81.705 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "ZBoxRule"] 75.005 μs (5%) 154.62 KiB (1%) 73
["Layer", "MaxPool", "ZeroRule"] 82.006 μs (5%) 154.62 KiB (1%) 73
["Layer", "WrappedDense", "EpsilonRule"] 48.503 μs (5%) 213.72 KiB (1%) 153
["Layer", "WrappedDense", "GammaRule"] 134.108 μs (5%) 801.25 KiB (1%) 162
["Layer", "WrappedDense", "ZBoxRule"] 238.415 μs (5%) 1.01 MiB (1%) 413
["Layer", "WrappedDense", "ZeroRule"] 53.404 μs (5%) 213.11 KiB (1%) 153
["VGG", "Gradient", "analyze"] 6.337 s (5%) 25.063 ms 1.12 GiB (1%) 329053
["VGG", "Gradient", "construct analyzer"] 20.601 μs (5%) 27.41 KiB (1%) 205
["VGG", "InputTimesGradient", "analyze"] 30.601 s (5%) 289.064 ms 1.91 GiB (1%) 15640502
["VGG", "InputTimesGradient", "construct analyzer"] 11.501 μs (5%) 27.41 KiB (1%) 205
["VGG", "LRPCustom", "analyze"] 10.063 s (5%) 213.922 ms 3.45 GiB (1%) 16511486
["VGG", "LRPCustom", "construct analyzer"] 39.902 μs (5%) 45.52 KiB (1%) 155
["VGG", "LRPZero", "analyze"] 1.207 s (5%) 10.336 ms 968.62 MiB (1%) 3994
["VGG", "LRPZero", "construct analyzer"] 13.301 μs (5%) 43.88 KiB (1%) 82

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 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       3962 s          2 s        652 s       5661 s          0 s
       #2  2095 MHz       6098 s          0 s        513 s       3715 s          0 s
       
  Memory: 6.7845458984375 GB (1416.875 MB free)
  Uptime: 1036.66 sec
  Load Avg:  1.34  1.25  0.91
  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:                         2095.193
BogoMIPS:                        4190.38
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 Spec store bypass: Vulnerable
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Full generic retpoline, 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/ExplainabilityMethods.jl/ExplainabilityMethods.jl

Job Properties

  • Time of benchmarks:
    • Target: 16 Mar 2022 - 15:06
    • Baseline: 16 Mar 2022 - 15:09
  • Package commits:
    • Target: baa836
    • Baseline: c778bb
  • Julia commits:
    • Target: bf5349
    • Baseline: bf5349
  • 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
["VGG", "Gradient", "analyze"] 0.21 (5%) ✅ 0.99 (1%) ✅
["VGG", "Gradient", "construct analyzer"] 0.78 (5%) ✅ 1.00 (1%)
["VGG", "InputTimesGradient", "analyze"] 0.04 (5%) ✅ 0.58 (1%) ✅
["VGG", "InputTimesGradient", "construct analyzer"] 1.25 (5%) ❌ 1.00 (1%)
["VGG", "LRPCustom", "analyze"] 0.20 (5%) ✅ 0.78 (1%) ✅
["VGG", "LRPCustom", "construct analyzer"] 0.31 (5%) ✅ 1.00 (1%)
["VGG", "LRPZero", "analyze"] 0.94 (5%) ✅ 1.00 (1%)
["VGG", "LRPZero", "construct analyzer"] 1.15 (5%) ❌ 1.00 (1%)

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Target

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2593 MHz       5458 s          1 s        380 s       1496 s          0 s
       #2  2593 MHz       1894 s          2 s        510 s       4918 s          0 s
       
  Memory: 6.7845458984375 GB (1642.640625 MB free)
  Uptime: 737.88 sec
  Load Avg:  1.35  1.27  0.77
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2593 MHz       6318 s          1 s        431 s       2851 s          0 s
       #2  2593 MHz       3413 s          2 s        619 s       5560 s          0 s
       
  Memory: 6.7845458984375 GB (1669.17578125 MB free)
  Uptime: 964.98 sec
  Load Avg:  1.14  1.18  0.85
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Target result

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

Job Properties

  • Time of benchmark: 16 Mar 2022 - 15:6
  • Package commit: baa836
  • Julia commit: bf5349
  • 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", "EpsilonRule"] 851.706 μs (5%) 1.59 MiB (1%) 364
["Layer", "Conv", "GammaRule"] 859.807 μs (5%) 1.59 MiB (1%) 376
["Layer", "Conv", "ZBoxRule"] 2.681 ms (5%) 4.77 MiB (1%) 1047
["Layer", "Conv", "ZeroRule"] 855.206 μs (5%) 1.59 MiB (1%) 364
["Layer", "Dense", "EpsilonRule"] 20.700 μs (5%) 5.08 KiB (1%) 9
["Layer", "Dense", "GammaRule"] 100.900 μs (5%) 592.61 KiB (1%) 18
["Layer", "Dense", "ZBoxRule"] 258.402 μs (5%) 1.01 MiB (1%) 413
["Layer", "Dense", "ZeroRule"] 21.001 μs (5%) 5.08 KiB (1%) 9
["Layer", "MaxPool", "EpsilonRule"] 88.800 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "GammaRule"] 88.200 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "ZBoxRule"] 88.601 μs (5%) 154.62 KiB (1%) 73
["Layer", "MaxPool", "ZeroRule"] 88.700 μs (5%) 154.62 KiB (1%) 73
["Layer", "WrappedDense", "EpsilonRule"] 50.800 μs (5%) 213.72 KiB (1%) 153
["Layer", "WrappedDense", "GammaRule"] 149.301 μs (5%) 801.25 KiB (1%) 162
["Layer", "WrappedDense", "ZBoxRule"] 256.802 μs (5%) 1.01 MiB (1%) 413
["Layer", "WrappedDense", "ZeroRule"] 50.400 μs (5%) 213.11 KiB (1%) 153
["VGG", "Gradient", "analyze"] 1.080 s (5%) 12.744 ms 1.10 GiB (1%) 3697
["VGG", "Gradient", "construct analyzer"] 15.200 μs (5%) 27.41 KiB (1%) 205
["VGG", "InputTimesGradient", "analyze"] 1.098 s (5%) 12.908 ms 1.10 GiB (1%) 3700
["VGG", "InputTimesGradient", "construct analyzer"] 15.000 μs (5%) 27.41 KiB (1%) 205
["VGG", "LRPCustom", "analyze"] 1.900 s (5%) 25.657 ms 2.68 GiB (1%) 4779
["VGG", "LRPCustom", "construct analyzer"] 19.600 μs (5%) 45.52 KiB (1%) 155
["VGG", "LRPZero", "analyze"] 1.146 s (5%) 11.567 ms 968.64 MiB (1%) 3973
["VGG", "LRPZero", "construct analyzer"] 13.400 μs (5%) 43.88 KiB (1%) 82

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2593 MHz       5458 s          1 s        380 s       1496 s          0 s
       #2  2593 MHz       1894 s          2 s        510 s       4918 s          0 s
       
  Memory: 6.7845458984375 GB (1642.640625 MB free)
  Uptime: 737.88 sec
  Load Avg:  1.35  1.27  0.77
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline result

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

Job Properties

  • Time of benchmark: 16 Mar 2022 - 15:9
  • Package commit: c778bb
  • Julia commit: bf5349
  • 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", "EpsilonRule"] 829.606 μs (5%) 1.59 MiB (1%) 364
["Layer", "Conv", "GammaRule"] 834.906 μs (5%) 1.59 MiB (1%) 376
["Layer", "Conv", "ZBoxRule"] 2.602 ms (5%) 4.77 MiB (1%) 1047
["Layer", "Conv", "ZeroRule"] 828.406 μs (5%) 1.59 MiB (1%) 364
["Layer", "Dense", "EpsilonRule"] 21.401 μs (5%) 5.08 KiB (1%) 9
["Layer", "Dense", "GammaRule"] 99.601 μs (5%) 592.61 KiB (1%) 18
["Layer", "Dense", "ZBoxRule"] 265.101 μs (5%) 1.01 MiB (1%) 413
["Layer", "Dense", "ZeroRule"] 21.600 μs (5%) 5.08 KiB (1%) 9
["Layer", "MaxPool", "EpsilonRule"] 86.401 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "GammaRule"] 84.501 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "ZBoxRule"] 85.100 μs (5%) 154.62 KiB (1%) 73
["Layer", "MaxPool", "ZeroRule"] 85.101 μs (5%) 154.62 KiB (1%) 73
["Layer", "WrappedDense", "EpsilonRule"] 49.701 μs (5%) 213.72 KiB (1%) 153
["Layer", "WrappedDense", "GammaRule"] 147.601 μs (5%) 801.25 KiB (1%) 162
["Layer", "WrappedDense", "ZBoxRule"] 250.102 μs (5%) 1.01 MiB (1%) 413
["Layer", "WrappedDense", "ZeroRule"] 51.901 μs (5%) 213.11 KiB (1%) 153
["VGG", "Gradient", "analyze"] 5.140 s (5%) 29.201 ms 1.12 GiB (1%) 329147
["VGG", "Gradient", "construct analyzer"] 19.600 μs (5%) 27.41 KiB (1%) 205
["VGG", "InputTimesGradient", "analyze"] 26.111 s (5%) 347.330 ms 1.91 GiB (1%) 15642429
["VGG", "InputTimesGradient", "construct analyzer"] 12.000 μs (5%) 27.41 KiB (1%) 205
["VGG", "LRPCustom", "analyze"] 9.289 s (5%) 273.443 ms 3.45 GiB (1%) 16513216
["VGG", "LRPCustom", "construct analyzer"] 62.600 μs (5%) 45.52 KiB (1%) 155
["VGG", "LRPZero", "analyze"] 1.222 s (5%) 25.505 ms 968.62 MiB (1%) 3994
["VGG", "LRPZero", "construct analyzer"] 11.700 μs (5%) 43.88 KiB (1%) 82

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2593 MHz       6318 s          1 s        431 s       2851 s          0 s
       #2  2593 MHz       3413 s          2 s        619 s       5560 s          0 s
       
  Memory: 6.7845458984375 GB (1669.17578125 MB free)
  Uptime: 964.98 sec
  Load Avg:  1.14  1.18  0.85
  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 8272CL CPU @ 2.60GHz
Stepping:                        7
CPU MHz:                         2593.906
BogoMIPS:                        5187.81
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 Spec store bypass: Vulnerable
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Full generic retpoline, 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 8272CL CPU @ 2.60GHz
Vendor :Intel
Architecture :Skylake
Model Family: 0x06, Model: 0x55, Stepping: 0x07, 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/ExplainabilityMethods.jl/ExplainabilityMethods.jl

Job Properties

  • Time of benchmarks:
    • Target: 16 Mar 2022 - 16:39
    • Baseline: 16 Mar 2022 - 16:43
  • Package commits:
    • Target: b95610
    • Baseline: c778bb
  • Julia commits:
    • Target: bf5349
    • Baseline: bf5349
  • 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", "Dense", "EpsilonRule"] 1.06 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "GammaRule"] 1.58 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "ZeroRule"] 1.09 (5%) ❌ 1.00 (1%)
["VGG", "Gradient", "construct analyzer"] 0.70 (5%) ✅ 1.00 (1%)
["VGG", "InputTimesGradient", "analyze"] 0.04 (5%) ✅ 0.58 (1%) ✅
["VGG", "InputTimesGradient", "construct analyzer"] 1.18 (5%) ❌ 1.00 (1%)
["VGG", "LRPCustom", "analyze"] 0.20 (5%) ✅ 0.78 (1%) ✅
["VGG", "LRPCustom", "construct analyzer"] 0.37 (5%) ✅ 1.00 (1%)

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Target

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2593 MHz       2503 s          1 s        536 s       4051 s          0 s
       #2  2593 MHz       4808 s          2 s        360 s       1943 s          0 s
       
  Memory: 6.7845458984375 GB (1785.1640625 MB free)
  Uptime: 714.98 sec
  Load Avg:  1.26  1.26  0.78
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2593 MHz       4512 s          1 s        603 s       4221 s          0 s
       #2  2593 MHz       5154 s          2 s        458 s       3739 s          0 s
       
  Memory: 6.7845458984375 GB (1625.0234375 MB free)
  Uptime: 939.67 sec
  Load Avg:  1.13  1.17  0.86
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Target result

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

Job Properties

  • Time of benchmark: 16 Mar 2022 - 16:39
  • Package commit: b95610
  • Julia commit: bf5349
  • 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", "EpsilonRule"] 835.207 μs (5%) 1.59 MiB (1%) 364
["Layer", "Conv", "GammaRule"] 845.906 μs (5%) 1.59 MiB (1%) 376
["Layer", "Conv", "ZBoxRule"] 2.631 ms (5%) 4.77 MiB (1%) 1047
["Layer", "Conv", "ZeroRule"] 841.406 μs (5%) 1.59 MiB (1%) 364
["Layer", "Dense", "EpsilonRule"] 20.900 μs (5%) 5.08 KiB (1%) 9
["Layer", "Dense", "GammaRule"] 105.701 μs (5%) 592.61 KiB (1%) 18
["Layer", "Dense", "ZBoxRule"] 254.302 μs (5%) 1.01 MiB (1%) 413
["Layer", "Dense", "ZeroRule"] 21.300 μs (5%) 5.08 KiB (1%) 9
["Layer", "MaxPool", "EpsilonRule"] 92.101 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "GammaRule"] 91.701 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "ZBoxRule"] 92.100 μs (5%) 154.62 KiB (1%) 73
["Layer", "MaxPool", "ZeroRule"] 92.001 μs (5%) 154.62 KiB (1%) 73
["Layer", "WrappedDense", "EpsilonRule"] 49.600 μs (5%) 213.72 KiB (1%) 153
["Layer", "WrappedDense", "GammaRule"] 148.801 μs (5%) 801.25 KiB (1%) 162
["Layer", "WrappedDense", "ZBoxRule"] 258.202 μs (5%) 1.01 MiB (1%) 413
["Layer", "WrappedDense", "ZeroRule"] 51.800 μs (5%) 213.11 KiB (1%) 153
["VGG", "Gradient", "analyze"] 1.093 s (5%) 14.179 ms 1.10 GiB (1%) 3697
["VGG", "Gradient", "construct analyzer"] 14.201 μs (5%) 27.41 KiB (1%) 205
["VGG", "InputTimesGradient", "analyze"] 1.088 s (5%) 13.971 ms 1.10 GiB (1%) 3700
["VGG", "InputTimesGradient", "construct analyzer"] 14.200 μs (5%) 27.41 KiB (1%) 205
["VGG", "LRPCustom", "analyze"] 1.861 s (5%) 25.354 ms 2.68 GiB (1%) 4795
["VGG", "LRPCustom", "construct analyzer"] 19.400 μs (5%) 45.52 KiB (1%) 155
["VGG", "LRPZero", "analyze"] 1.154 s (5%) 11.605 ms 968.64 MiB (1%) 3989
["VGG", "LRPZero", "construct analyzer"] 13.200 μs (5%) 43.88 KiB (1%) 82

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2593 MHz       2503 s          1 s        536 s       4051 s          0 s
       #2  2593 MHz       4808 s          2 s        360 s       1943 s          0 s
       
  Memory: 6.7845458984375 GB (1785.1640625 MB free)
  Uptime: 714.98 sec
  Load Avg:  1.26  1.26  0.78
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline result

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

Job Properties

  • Time of benchmark: 16 Mar 2022 - 16:43
  • Package commit: c778bb
  • Julia commit: bf5349
  • 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", "EpsilonRule"] 818.789 μs (5%) 1.59 MiB (1%) 364
["Layer", "Conv", "GammaRule"] 820.116 μs (5%) 1.59 MiB (1%) 376
["Layer", "Conv", "ZBoxRule"] 2.586 ms (5%) 4.77 MiB (1%) 1047
["Layer", "Conv", "ZeroRule"] 820.208 μs (5%) 1.59 MiB (1%) 364
["Layer", "Dense", "EpsilonRule"] 19.700 μs (5%) 5.08 KiB (1%) 9
["Layer", "Dense", "GammaRule"] 66.800 μs (5%) 592.61 KiB (1%) 18
["Layer", "Dense", "ZBoxRule"] 264.502 μs (5%) 1.01 MiB (1%) 413
["Layer", "Dense", "ZeroRule"] 19.600 μs (5%) 5.08 KiB (1%) 9
["Layer", "MaxPool", "EpsilonRule"] 91.405 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "GammaRule"] 91.401 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "ZBoxRule"] 92.300 μs (5%) 154.62 KiB (1%) 73
["Layer", "MaxPool", "ZeroRule"] 92.004 μs (5%) 154.62 KiB (1%) 73
["Layer", "WrappedDense", "EpsilonRule"] 50.901 μs (5%) 213.72 KiB (1%) 153
["Layer", "WrappedDense", "GammaRule"] 149.905 μs (5%) 801.25 KiB (1%) 162
["Layer", "WrappedDense", "ZBoxRule"] 256.804 μs (5%) 1.01 MiB (1%) 413
["Layer", "WrappedDense", "ZeroRule"] 50.700 μs (5%) 213.11 KiB (1%) 153
["VGG", "Gradient", "analyze"] 1.140 s (5%) 37.125 ms 1.10 GiB (1%) 3707
["VGG", "Gradient", "construct analyzer"] 20.200 μs (5%) 27.41 KiB (1%) 205
["VGG", "InputTimesGradient", "analyze"] 25.239 s (5%) 315.545 ms 1.91 GiB (1%) 15646739
["VGG", "InputTimesGradient", "construct analyzer"] 12.000 μs (5%) 27.41 KiB (1%) 205
["VGG", "LRPCustom", "analyze"] 9.309 s (5%) 252.670 ms 3.44 GiB (1%) 16507915
["VGG", "LRPCustom", "construct analyzer"] 52.700 μs (5%) 45.52 KiB (1%) 155
["VGG", "LRPZero", "analyze"] 1.179 s (5%) 10.334 ms 968.62 MiB (1%) 3994
["VGG", "LRPZero", "construct analyzer"] 13.300 μs (5%) 43.88 KiB (1%) 82

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 UTC 2022 x86_64 x86_64
  CPU: Intel(R) Xeon(R) Platinum 8272CL CPU @ 2.60GHz: 
              speed         user         nice          sys         idle          irq
       #1  2593 MHz       4512 s          1 s        603 s       4221 s          0 s
       #2  2593 MHz       5154 s          2 s        458 s       3739 s          0 s
       
  Memory: 6.7845458984375 GB (1625.0234375 MB free)
  Uptime: 939.67 sec
  Load Avg:  1.13  1.17  0.86
  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 8272CL CPU @ 2.60GHz
Stepping:                        7
CPU MHz:                         2593.906
BogoMIPS:                        5187.81
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 Spec store bypass: Vulnerable
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Full generic retpoline, 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 8272CL CPU @ 2.60GHz
Vendor :Intel
Architecture :Skylake
Model Family: 0x06, Model: 0x55, Stepping: 0x07, 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

@codecov-commenter
Copy link

Codecov Report

Merging #38 (9011485) into master (c778bba) will increase coverage by 6.65%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #38      +/-   ##
==========================================
+ Coverage   87.73%   94.38%   +6.65%     
==========================================
  Files          10        9       -1     
  Lines         212      196      -16     
==========================================
- Hits          186      185       -1     
+ Misses         26       11      -15
Impacted Files Coverage Δ
src/gradient.jl 100.00% <100.00%> (ø)
src/lrp.jl 96.15% <100.00%> (-0.40%) ⬇️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c778bba...9011485. Read the comment docs.

@github-actions
Copy link

Benchmark result

Judge result

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

Job Properties

  • Time of benchmarks:
    • Target: 16 Mar 2022 - 16:41
    • Baseline: 16 Mar 2022 - 16:46
  • Package commits:
    • Target: b95610
    • Baseline: c778bb
  • Julia commits:
    • Target: bf5349
    • Baseline: bf5349
  • 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", "EpsilonRule"] 1.10 (5%) ❌ 1.00 (1%)
["Layer", "Conv", "GammaRule"] 1.14 (5%) ❌ 1.00 (1%)
["Layer", "Conv", "ZBoxRule"] 1.07 (5%) ❌ 1.00 (1%)
["Layer", "Conv", "ZeroRule"] 1.07 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "EpsilonRule"] 1.10 (5%) ❌ 1.00 (1%)
["Layer", "Dense", "ZeroRule"] 1.06 (5%) ❌ 1.00 (1%)
["Layer", "MaxPool", "EpsilonRule"] 0.89 (5%) ✅ 1.00 (1%)
["Layer", "MaxPool", "GammaRule"] 1.14 (5%) ❌ 1.00 (1%)
["Layer", "MaxPool", "ZBoxRule"] 0.81 (5%) ✅ 1.00 (1%)
["Layer", "WrappedDense", "ZBoxRule"] 0.94 (5%) ✅ 1.00 (1%)
["VGG", "Gradient", "analyze"] 0.25 (5%) ✅ 0.99 (1%) ✅
["VGG", "Gradient", "construct analyzer"] 0.81 (5%) ✅ 1.00 (1%)
["VGG", "InputTimesGradient", "analyze"] 0.04 (5%) ✅ 0.58 (1%) ✅
["VGG", "InputTimesGradient", "construct analyzer"] 1.27 (5%) ❌ 1.00 (1%)
["VGG", "LRPCustom", "analyze"] 0.21 (5%) ✅ 0.78 (1%) ✅
["VGG", "LRPCustom", "construct analyzer"] 0.39 (5%) ✅ 1.00 (1%)

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Target

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 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       3020 s          2 s        494 s       3397 s          0 s
       #2  2095 MHz       5134 s          1 s        561 s       1249 s          0 s
       
  Memory: 6.7845458984375 GB (996.234375 MB free)
  Uptime: 699.12 sec
  Load Avg:  1.28  1.32  0.87
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 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       3284 s          2 s        606 s       5567 s          0 s
       #2  2095 MHz       7531 s          1 s        629 s       1339 s          0 s
       
  Memory: 6.7845458984375 GB (1833.26171875 MB free)
  Uptime: 954.72 sec
  Load Avg:  1.26  1.26  0.96
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Target result

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

Job Properties

  • Time of benchmark: 16 Mar 2022 - 16:41
  • Package commit: b95610
  • Julia commit: bf5349
  • 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", "EpsilonRule"] 972.756 μs (5%) 1.59 MiB (1%) 364
["Layer", "Conv", "GammaRule"] 976.856 μs (5%) 1.59 MiB (1%) 376
["Layer", "Conv", "ZBoxRule"] 3.066 ms (5%) 4.77 MiB (1%) 1047
["Layer", "Conv", "ZeroRule"] 962.655 μs (5%) 1.59 MiB (1%) 364
["Layer", "Dense", "EpsilonRule"] 22.302 μs (5%) 5.08 KiB (1%) 9
["Layer", "Dense", "GammaRule"] 104.406 μs (5%) 592.61 KiB (1%) 18
["Layer", "Dense", "ZBoxRule"] 282.417 μs (5%) 1.01 MiB (1%) 413
["Layer", "Dense", "ZeroRule"] 23.601 μs (5%) 5.08 KiB (1%) 9
["Layer", "MaxPool", "EpsilonRule"] 88.505 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "GammaRule"] 100.906 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "ZBoxRule"] 80.205 μs (5%) 154.62 KiB (1%) 73
["Layer", "MaxPool", "ZeroRule"] 91.205 μs (5%) 154.62 KiB (1%) 73
["Layer", "WrappedDense", "EpsilonRule"] 60.204 μs (5%) 213.72 KiB (1%) 153
["Layer", "WrappedDense", "GammaRule"] 160.309 μs (5%) 801.25 KiB (1%) 162
["Layer", "WrappedDense", "ZBoxRule"] 264.015 μs (5%) 1.01 MiB (1%) 413
["Layer", "WrappedDense", "ZeroRule"] 60.003 μs (5%) 213.11 KiB (1%) 153
["VGG", "Gradient", "analyze"] 1.257 s (5%) 14.582 ms 1.10 GiB (1%) 3697
["VGG", "Gradient", "construct analyzer"] 16.901 μs (5%) 27.41 KiB (1%) 205
["VGG", "InputTimesGradient", "analyze"] 1.233 s (5%) 14.107 ms 1.10 GiB (1%) 3700
["VGG", "InputTimesGradient", "construct analyzer"] 17.301 μs (5%) 27.41 KiB (1%) 205
["VGG", "LRPCustom", "analyze"] 2.286 s (5%) 27.177 ms 2.68 GiB (1%) 4795
["VGG", "LRPCustom", "construct analyzer"] 21.901 μs (5%) 45.52 KiB (1%) 155
["VGG", "LRPZero", "analyze"] 1.275 s (5%) 13.563 ms 968.64 MiB (1%) 3989
["VGG", "LRPZero", "construct analyzer"] 14.901 μs (5%) 43.88 KiB (1%) 82

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 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       3020 s          2 s        494 s       3397 s          0 s
       #2  2095 MHz       5134 s          1 s        561 s       1249 s          0 s
       
  Memory: 6.7845458984375 GB (996.234375 MB free)
  Uptime: 699.12 sec
  Load Avg:  1.28  1.32  0.87
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-12.0.1 (ORCJIT, skylake-avx512)

Baseline result

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

Job Properties

  • Time of benchmark: 16 Mar 2022 - 16:46
  • Package commit: c778bb
  • Julia commit: bf5349
  • 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", "EpsilonRule"] 884.151 μs (5%) 1.59 MiB (1%) 364
["Layer", "Conv", "GammaRule"] 857.649 μs (5%) 1.59 MiB (1%) 376
["Layer", "Conv", "ZBoxRule"] 2.876 ms (5%) 4.77 MiB (1%) 1047
["Layer", "Conv", "ZeroRule"] 900.152 μs (5%) 1.59 MiB (1%) 364
["Layer", "Dense", "EpsilonRule"] 20.301 μs (5%) 5.08 KiB (1%) 9
["Layer", "Dense", "GammaRule"] 100.306 μs (5%) 592.61 KiB (1%) 18
["Layer", "Dense", "ZBoxRule"] 281.016 μs (5%) 1.01 MiB (1%) 413
["Layer", "Dense", "ZeroRule"] 22.301 μs (5%) 5.08 KiB (1%) 9
["Layer", "MaxPool", "EpsilonRule"] 99.606 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "GammaRule"] 88.705 μs (5%) 155.34 KiB (1%) 73
["Layer", "MaxPool", "ZBoxRule"] 99.306 μs (5%) 154.62 KiB (1%) 73
["Layer", "MaxPool", "ZeroRule"] 88.005 μs (5%) 154.62 KiB (1%) 73
["Layer", "WrappedDense", "EpsilonRule"] 58.403 μs (5%) 213.72 KiB (1%) 153
["Layer", "WrappedDense", "GammaRule"] 156.109 μs (5%) 801.25 KiB (1%) 162
["Layer", "WrappedDense", "ZBoxRule"] 281.116 μs (5%) 1.01 MiB (1%) 413
["Layer", "WrappedDense", "ZeroRule"] 58.904 μs (5%) 213.11 KiB (1%) 153
["VGG", "Gradient", "analyze"] 4.985 s (5%) 27.296 ms 1.12 GiB (1%) 329044
["VGG", "Gradient", "construct analyzer"] 20.902 μs (5%) 27.41 KiB (1%) 205
["VGG", "InputTimesGradient", "analyze"] 30.930 s (5%) 324.206 ms 1.91 GiB (1%) 15646735
["VGG", "InputTimesGradient", "construct analyzer"] 13.600 μs (5%) 27.41 KiB (1%) 205
["VGG", "LRPCustom", "analyze"] 10.914 s (5%) 247.864 ms 3.45 GiB (1%) 16513321
["VGG", "LRPCustom", "construct analyzer"] 55.503 μs (5%) 45.52 KiB (1%) 155
["VGG", "LRPZero", "analyze"] 1.241 s (5%) 10.276 ms 968.62 MiB (1%) 3994
["VGG", "LRPZero", "construct analyzer"] 14.601 μs (5%) 43.88 KiB (1%) 82

Benchmark Group List

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

  • ["Layer", "Conv"]
  • ["Layer", "Dense"]
  • ["Layer", "MaxPool"]
  • ["Layer", "WrappedDense"]
  • ["VGG", "Gradient"]
  • ["VGG", "InputTimesGradient"]
  • ["VGG", "LRPCustom"]
  • ["VGG", "LRPZero"]

Julia versioninfo

Julia Version 1.7.2
Commit bf53498635 (2022-02-06 15:21 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
      Ubuntu 20.04.4 LTS
  uname: Linux 5.11.0-1028-azure #31~20.04.2-Ubuntu SMP Tue Jan 18 08:46:15 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       3284 s          2 s        606 s       5567 s          0 s
       #2  2095 MHz       7531 s          1 s        629 s       1339 s          0 s
       
  Memory: 6.7845458984375 GB (1833.26171875 MB free)
  Uptime: 954.72 sec
  Load Avg:  1.26  1.26  0.96
  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:                         2095.175
BogoMIPS:                        4190.35
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 Spec store bypass: Vulnerable
Vulnerability Spectre v1:        Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:        Mitigation; Full generic retpoline, 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

@adrhill
Copy link
Member Author

adrhill commented Mar 16, 2022

Locally, the inplace-updating variant lrp! saves a lot of allocations:

  Results
  =========

  A ratio greater than 1.0 denotes a possible regression (marked with :x:), while a ratio less than 1.0 denotes a possible improvement (marked with :whitecheckmark:). 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", "EpsilonRule"]                  0.99 (5%) 0.97 (1%) :whitecheckmark:
              ["Layer", "Conv", "GammaRule"]                  1.00 (5%) 0.97 (1%) :whitecheckmark:
               ["Layer", "Conv", "ZeroRule"]                  1.00 (5%) 0.97 (1%) :whitecheckmark:
           ["Layer", "Dense", "EpsilonRule"]                  0.96 (5%) 0.59 (1%) :whitecheckmark:
              ["Layer", "Dense", "ZeroRule"]                  0.96 (5%) 0.59 (1%) :whitecheckmark:
         ["Layer", "MaxPool", "EpsilonRule"]                  0.98 (5%) 0.69 (1%) :whitecheckmark:
           ["Layer", "MaxPool", "GammaRule"]                  1.00 (5%) 0.69 (1%) :whitecheckmark:
            ["Layer", "MaxPool", "ZBoxRule"] 0.91 (5%) :whitecheckmark: 0.69 (1%) :whitecheckmark:
            ["Layer", "MaxPool", "ZeroRule"]                  1.00 (5%) 0.69 (1%) :whitecheckmark:
              ["VGG", "Gradient", "analyze"] 0.84 (5%) :whitecheckmark:                  1.00 (1%)
    ["VGG", "InputTimesGradient", "analyze"] 0.89 (5%) :whitecheckmark: 0.99 (1%) :whitecheckmark:
             ["VGG", "LRPCustom", "analyze"] 0.89 (5%) :whitecheckmark:                  0.99 (1%)
  ["VGG", "LRPCustom", "construct analyzer"] 0.94 (5%) :whitecheckmark:                  1.00 (1%)
               ["VGG", "LRPZero", "analyze"]                  0.96 (5%) 0.96 (1%) :whitecheckmark:
    ["VGG", "LRPZero", "construct analyzer"] 0.93 (5%) :whitecheckmark:                  1.00 (1%)

@adrhill adrhill changed the title Faster LRP preallocation and gradient code reuse Add inplace updating lrp! rule calls and reuse gradient code Mar 16, 2022
@adrhill adrhill merged commit 53b9f95 into master Mar 16, 2022
@adrhill adrhill deleted the ah/update-backend branch March 16, 2022 16:58
@adrhill adrhill mentioned this pull request May 25, 2022
Closed
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
run benchmark Run benchmark CI action
Projects
None yet
Milestone
No milestone
Development

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