Commits · c47b6408722050730f6d814db3346abfad3b690a · llvm / hpvm-release · GitLab

Snippets Groups Projects

Jul 05, 2021
- Adding working FP16 quantization and Tranpose Dense weights scripts (tested on simulator) · c47b6408
  Hashim Sharif authored 3 years ago
  
  c47b6408
Jul 02, 2021
- Adding weight manipulation scripts (FP16 + Transpose) · 48b62996
  Hashim Sharif authored 3 years ago
  
  48b62996
- Backing up Shubham's new scales · 46bdcda8
  Hashim Sharif authored 3 years ago
  
  46bdcda8
Jun 30, 2021
- Adding FP16 weight quantization script - NVDLA needs FP16 weights · bd9f17a7
  Hashim Sharif authored 3 years ago
  
  bd9f17a7
- Compiling AlexNet with INT8 -- pushing calib values shared by Shubham · 1e9c7b89
  Hashim Sharif authored 3 years ago
  
  1e9c7b89
- Adding few test images for mini-era CNN · efda3996
  Hashim Sharif authored 3 years ago
  
  efda3996
- Moving location of calib table file · 16015d99
  Hashim Sharif authored 3 years ago
  
  16015d99
- HPVM NVDLA backend INT8 mode working with mini-era CNN · 49f39179
  Hashim Sharif authored 3 years ago
  
  49f39179
Jun 28, 2021
- Porting Mini-era CNN to HPVM-9 -- compiles with ported NVDLA pass · fd51ccbd
  Hashim Sharif authored 3 years ago
  
  fd51ccbd
Jun 24, 2021
- Modifying readTrainedWeights calls in Alexnet to read global constants · ddd7a4d2
  Hashim Sharif authored 3 years ago
  
  ddd7a4d2
Jun 09, 2021
- Removing config.h header and adding iostream header · 87dcf884
  Hashim Sharif authored 3 years ago
  
  87dcf884
Apr 26, 2021
- Moved pipeline benchmark to work on OpenCV 3/4 · 6e961cf8
  Yifan Zhao authored 3 years ago
  
  6e961cf8
Apr 08, 2021
- Doc fixes from proofreading · 9933bc26
  Yifan Zhao authored 3 years ago
  
  9933bc26
- Use noexcept instead of -fno-exceptions · 99424f70
  Yifan Zhao authored 3 years ago
  
  99424f70
- Small improvements to the doc · f6e1bf5c
  Yifan Zhao authored 3 years ago
  
  f6e1bf5c
- Pick up predtuner updates and updated related document · 1d52149a
  Yifan Zhao authored 3 years ago
  
  1d52149a
Apr 06, 2021
- Fixed a bug in profile plotting and added test case for it · 831efeff
  Yifan Zhao authored 3 years ago
  
  831efeff
- CUDA is now optional; not having CUDA will disable all ApproxHPVM stuff · 17c8d30d
  Yifan Zhao authored 3 years ago
  
  17c8d30d
- Added hpvm-tensor-rt unit tests · 86b14625
  Yifan Zhao authored 3 years ago
  
  86b14625
- Added tests for Keras frontend · aede13c7
  Yifan Zhao authored 3 years ago
  
  aede13c7
Apr 05, 2021
- Adding argparse options to Keras frontend test script · da71c5c7
  Hashim Sharif authored 3 years ago
  
  da71c5c7
- PyTorch frontend and tuner jointly supports knobs-on-devices · 771a9e21
  Yifan Zhao authored 3 years ago
  
  771a9e21
Apr 04, 2021
- Updated test readme · b1225ff2
  Yifan Zhao authored 3 years ago
  
  b1225ff2
- Moved profiling test cases to test frontend + profiling · 424a32a2
  Yifan Zhao authored 3 years ago
  
  424a32a2
- Moved example code generation into dnn package · 922eeaee
  Yifan Zhao authored 3 years ago
  
  922eeaee
- Moved jetson clock script (use it in the future) · 612f1cba
  Yifan Zhao authored 3 years ago
  
  612f1cba
- Bump predtuner version (bugfix) and changes to test script · 4014c9f5
  Yifan Zhao authored 3 years ago
  
  4014c9f5
- Added test cases for tuning (coming from torch frontend) · 475b44d2
  Yifan Zhao authored 3 years ago
  
  475b44d2
- Fixed some problems in HPVM benchmark 'pipeline' · d8d55c3e
  Yifan Zhao authored 3 years ago
  
  d8d55c3e
- Wrapping up Keras frontend automated test script · 65434337
  Hashim Sharif authored 3 years ago
  
  65434337
- Pushing Keras test script changes · 11068a0e
  Hashim Sharif authored 3 years ago
  
  11068a0e
- Proper parens in previous change · c49249e1
  leonkm2 authored 3 years ago and Yifan Zhao committed 3 years ago
  
  c49249e1
- Make hpvm-cava allocate file paths dynamically · 6e835b56
  leonkm2 authored 3 years ago and Yifan Zhao committed 3 years ago
  
  6e835b56
Apr 03, 2021
- Removed cudnn targets in non-cudnn benchmarks · 3b35b08f
  Yifan Zhao authored 3 years ago
  
  3b35b08f
- Pass compiler def instead of generated MODEL_PARAM into file (less invasive) · 92aeb02c
  Yifan Zhao authored 3 years ago
  
  92aeb02c
- Added accuracy checking for frontend-generated code · 13c1799f
  Yifan Zhao authored 3 years ago
  
  13c1799f
- Removed outdated hpvm-c/include · e272386c
  Yifan Zhao authored 3 years ago
  
  e272386c
Apr 02, 2021
- Created test case for pytorch frontend · 717e55dc
  Yifan Zhao authored 3 years ago
  
  717e55dc
- Only test first 5 configs in check-hpvm-profiler (should take 3 mins overall) · dd1bff2e
  Yifan Zhao authored 3 years ago
  
  dd1bff2e
Apr 01, 2021
- Only provide gpu variants of old HPVM bench when OpenCL exists · dea38432
  Yifan Zhao authored 3 years ago
  
  dea38432