flyingmrwang

video stream change after drawMask

2

Glad to be the first issue! Your code is pretty concise and works well as demo for tfjs. What I am doing is to change "drawBokehEffect" to "drawMask", the result...

BatchNorm1d in last layer of backbone cause tensorRT conversion failed

1

I am trying to convert model to tensorRT. In my exported onnx model, BatchNorm1d always comes along with unsqueeze before it, which is not supported by tensorRT5. I noticed there...

I notice there is make_weights_for_balanced_classes to generate samplers to balance classes. While we can also make the balance with focal loss. So my question is do we really need both...

Question about your way to measure joint velocities

1

Hi, I noticed that joint velocities is an important component of your feature to classify. While the input frame rate is not fixed. I mean, for example, given a set...

Hi, this is a great work. After study, I have several questions: 1. If I wanna improve my model with your augmentation, can I just download your augmented dataset and...

hls(.m3u8) as video input?

1

Hi, does Deepstream support m3u8 live stream video as input? If yes, how to build the pipeline to capture and decode input?

"75" should be "39", which stands for "right corner of left eye" rather than "left side of nose bridge" https://github.com/Hsintao/pfld_106_face_landmarks/blob/e150195b0c945a9fcfe91641e3b79682e0e421e6/data/prepare.py#L206 x for left nose should be positive https://github.com/Hsintao/pfld_106_face_landmarks/blob/e150195b0c945a9fcfe91641e3b79682e0e421e6/data/prepare.py#L50 those mistakes...

segmentation fault

1

Any idea for solving this? I just re-compiled caffe after merging your dw_conv. This shows when I use caffe timer on both cpu and gpu mode. *** Aborted at 1520306614...