Commits · b89d2c0e110e273832fc5907961267c78f676e5e · llvm / hpvm-release · GitLab

Snippets Groups Projects

Jul 05, 2021
- Adding instructions for compiling mini-era benchmark for NVDLA · b89d2c0e
  Hashim Sharif authored 3 years ago
  
  b89d2c0e
- Adding script to invoke hpvm-clang with mini-era compilation · 2d37151b
  Hashim Sharif authored 3 years ago
  
  2d37151b
- Updating LeNet source to be NVDLA consistent (reads fp16 weights) · e88214dd
  Hashim Sharif authored 3 years ago
  
  e88214dd
- testing images for MNIST · ae000c18
  Hashim Sharif authored 3 years ago
  
  ae000c18
- 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 03, 2021
- Removed cudnn targets in non-cudnn benchmarks · 3b35b08f
  Yifan Zhao authored 4 years ago
  
  3b35b08f
- Pass compiler def instead of generated MODEL_PARAM into file (less invasive) · 92aeb02c
  Yifan Zhao authored 4 years ago
  
  92aeb02c
- Removed outdated hpvm-c/include · e272386c
  Yifan Zhao authored 4 years ago
  
  e272386c
Mar 31, 2021
- Fixing buggy LeNet configuration · 6c6fe1eb
  Hashim Sharif authored 4 years ago
  
  6c6fe1eb
- Removed config.h.in to fix config.h collision (vs tensor_runtime/include) · 9e5b04e6
  Yifan Zhao authored 4 years ago
  
  9e5b04e6
Mar 26, 2021
- Remove copyInputBatch altogether · c1de4884
  Yifan Zhao authored 4 years ago
  
  c1de4884
Mar 25, 2021
- Added -c argument to hpvm-c code · 98f92667
  Yifan Zhao authored 4 years ago
  
  98f92667
- Fixed bugs in hpvm-c benchmarks · 03425866
  Yifan Zhao authored 4 years ago
  
  03425866
- Fix various problems with hpvm-c benchmarks · 466d5930
  Yifan Zhao authored 4 years ago
  
  466d5930
Mar 22, 2021
- Fixing input size to 5K in cudnn HPVM src · d22b2427
  Hashim Sharif authored 4 years ago
  
  d22b2427
Feb 02, 2021
- Add graphs for different benchmarks and fix batch sizes for alexnets · e3f54677
  Akash Kothari authored 4 years ago
  
  e3f54677
Jan 27, 2021
- Moved test benchmarks to make space for keras and pytorch sources · 9c11c880
  Yifan Zhao authored 4 years ago
  
  9c11c880