Commits

Branch selector

 0d23e7a

User selector

All users

Datepicker

All time

Commit History

Commits on May 1, 2019

• created imports

  dctian

  committed
  0d23e7a

• Initial check in for end to end self driving (Nvidia)

  dctian

  committed
  6462d4e

Commits on Apr 30, 2019

• remove this file from source control

  dctian

  committed
  0f7e523

• make red light stop more robust using a time.sleep. not elegant, but it seems to work. may be better to check for 20 frames of no object for stability.

  dctian

  committed
  065f4f4

• updated project files

  dctian

  committed
  ec09c91

• traffic sign detection and processing working for all 5 traffic signs and pedestrians.

  dctian

  committed
  6a61ac5

Commits on Apr 29, 2019

• traffic sign detection works in software only. need in car testing.

  dctian

  committed
  8df74d5

• object detection is working

  dctian

  committed
  90b260b

• do not include tmp folders

  dctian

  committed
  e912ad3

Commits on Apr 27, 2019

• calibrated so that lane following works with the ELP wide angle camera (the blue looks darker than the stock camera). Removed perspective view code.

  dctian

  committed
  8c016d9

Commits on Apr 20, 2019

• Add more images for documentation

  dctian

  committed
  9f3cf5d

• Add ssd cocoedge tpu test

  dctian

  committed
  f9f2e90

• fix more typos

  dctian

  committed
  149d537

• fix typo in opencv test

  dctian

  committed
  649e34f

• Add opencv test to check if installation went fine.

  dctian

  committed
  f5f89b0

Commits on Apr 19, 2019

• Created using Colaboratory

  dctian

  committed
  fa740fb

• this image not used in article

  dctian

  committed
  91eec7c

• 1st and 2nd DeepPiCar Medium Articles and its images (total of 6)

  dctian

  committed
  93414e8

Commits on Apr 16, 2019

• version 2 of road object detection model. This should be much more precise.

  dctian

  committed
  44d2bb2

• labeled images are jpg now

  dctian

  committed
  8bf7efe

• Add more labeled data

  dctian

  committed
  0d2601d

• More object images

  dctian

  committed
  b70e0a6

• convert perspetive view to top down view so that headings will be computed more correctly

  dctian

  committed
  cf9ffec

Commits on Apr 15, 2019

• Initial check in of traffic sign and person processor

  dctian

  committed
  12dd47f

• This model file is specifically compiled for edge tpu.

  dctian

  committed
  0c3de65

• got object detection script to run. inferences are very slow at 2fps, compared with model zoo pretrained model at 8fps...

  dctian

  committed
  065ad76

• detection model after 764 epochs

  dctian

  committed
  1c6ff54

• Updated the format of the label file to be used by the script.

  dctian

  committed
  383bbd1

• upload scripts to detect road signs in video

  dctian

  committed
  2bcda66

• detection model after 100 epochs

  dctian

  committed
  3b26b81

• Merge branch 'master' of https://github.com/dctian/DeepPiCar

  dctian

  committed
  c523a5b

• lane follower minor update

  dctian

  committed
  44ddb39

• tensorboard showing progress of the training. current detection results on test set images vs labels

  dctian

  committed
  bccb231

• jupyter notebook to do the transfer learning training on Google Colab (for free)

  dctian

  committed
  1ce9dc0

• 1st check in of the tutorial. It is very long and may not be formatted correctly. Just check it in I don't lose it

  dctian

  committed
  1af630d

Pagination

PreviousNext