Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Jul 17, 2020

• removed stagnant development tasks

  tano297

  committed
  f0a2022

  Copy full SHA for f0a2022

Commits on Aug 9, 2019

• yaml.load -> yaml.safe_load

  tano297

  committed
  5368eed

  Copy full SHA for 5368eed

• fps for tables

  tano297

  committed
  8cf091b

  Copy full SHA for 8cf091b

Commits on Aug 8, 2019

• removed style html

  tano297

  committed
  eaa528b

  Copy full SHA for eaa528b

• zed sdk depth modes

  tano297

  committed
  e8dd967

  Copy full SHA for e8dd967

• jetson benchmark

  tano297

  committed
  1f2fc62

  Copy full SHA for 1f2fc62

• benchmark jetson

  tano297

  committed
  d2d798c

  Copy full SHA for d2d798c

• possibility to enable DLA for jetson xavier

  tano297

  committed
  cf61973

  Copy full SHA for cf61973

• verbosity in cpp

  tano297

  committed
  360c2df

  Copy full SHA for 360c2df

Commits on Aug 6, 2019

• inference without logs OR verbosity make no sense

  tano297

  committed
  32a1180

  Copy full SHA for 32a1180

• removed torch deps from trt stuff

  tano297

  committed
  c222d30

  Copy full SHA for c222d30

Commits on Jul 24, 2019

• extra check in train.py for path OR cfg file

  tano297

  committed
  5aed807

  Copy full SHA for 5aed807

• fix pytorch 1.1 and below

  tano297

  committed
  08694ef

  Copy full SHA for 08694ef

Commits on Jul 17, 2019

• more models and datasets

  tano297

  committed
  90d0ef7

  Copy full SHA for 90d0ef7

Commits on Jul 11, 2019

• fixed ROS key in dockerfile base

  tano297

  committed
  0ab883b

  Copy full SHA for 0ab883b

Commits on Jul 2, 2019

• minor bugfix infer_video

  tano297

  committed
  86cf3bd

  Copy full SHA for 86cf3bd

• contrib

  tano297

  committed
  80bf7d4

  Copy full SHA for 80bf7d4

Commits on Jun 27, 2019

• updated readme.md

  tano297

  committed
  b1fb66a

  Copy full SHA for b1fb66a

Commits on Jun 26, 2019

• updated readme.md

  tano297

  committed
  593e1e8

  Copy full SHA for 593e1e8

Commits on Jun 25, 2019

• more models and datasets

  tano297

  committed
  d6ed39c

  Copy full SHA for d6ed39c

• proper 1d onehot

  tano297

  committed
  9d39a49

  Copy full SHA for 9d39a49

Commits on Jun 12, 2019

• tensorrt asynchronous execution and pinned memory for semantic segmentation as well, following last merge format

  tano297

  committed
  f8ab00c

  Copy full SHA for f8ab00c

• Merge pull request #2 from ChandrahasJR/master

  Show description for 6ba15e5
  tano297

  authored
  6ba15e5

  Copy full SHA for 6ba15e5

• iou evaluator accumulating in longs and dividing in doubles

  tano297

  committed
  ef7e55c

  Copy full SHA for ef7e55c

Commits on Jun 11, 2019

• Make async inference for TensorRT backend

  ChandrahasJR

  committed
  0cd8c5a

  Copy full SHA for 0cd8c5a

• smarter iou

  tano297

  committed
  3f90978

  Copy full SHA for 3f90978

• Merge pull request #1 from ChandrahasJR/master

  Show description for 95cd16a
  tano297

  authored
  95cd16a

  Copy full SHA for 95cd16a

• Fix module import error

  ChandrahasJR

  committed
  178215c

  Copy full SHA for 178215c

Commits on Jun 5, 2019

• added int8 inference (works VERY fast)

  tano297

  committed
  76ceb26

  Copy full SHA for 76ceb26

• cleaner threading python video inference

  tano297

  committed
  15a532b

  Copy full SHA for 15a532b

• minor improvement to multigpu training to fix losses in each gpu and increase batch size

  tano297

  committed
  da6a3dc

  Copy full SHA for da6a3dc

• minor improvements to infer_video.py to avoid accidental closes

  tano297

  committed
  9516685

  Copy full SHA for 9516685

Commits on Jun 1, 2019

• updated base docker to reflect changes in tensorrt5.1.5 (bugfix that reported int32 sizes wrong is necessary to use argmax on gpu)

  tano297

  committed
  dfdb21a

  Copy full SHA for dfdb21a

• now tensorrt does argmax with gpu for segmentation (its expensive to copy to cpu and slower to run)

  tano297

  committed
  e38cb8a

  Copy full SHA for e38cb8a

• tracesaver must have cudnn fastest and benchmark set, to equal speed in c++ and python

  tano297

  committed
  a9aac74

  Copy full SHA for a9aac74

Pagination

Previous

Next