Vinjohn

training phase multiple gpu used

2

Hi, When I start to train, I noticed that apart from the gpu I assigned in train.lua, all other gpus also occupied by 200 to 300 mb memory. Say I...

Output dimension

3

Hi, Thanks for your work. I am about to use your pre-trained mlmodel to build an iOS demo. But I am new in pose estimation. The output dimension is 14*112*112....

Hi, I wonder which part of the code corresponds to the **adaptive roialign** mentioned in the paper Position Detection and Direction Prediction for Arbitrary-Oriented Ships via Multiscale Rotation Region Convolutional...

Will the training code be made available?

2

Your work is inspiring, is there a plan to publish the training code lately? Thanks!

training images of segmentation

1

Hi, After downloading your 500GB datasets, I found only two directories with image/mask pairs, deeplab_resnet_examples and segnet_examples. However, the mask was generated not annotated. In the README.rst, you mentioned training-images,...