Can't get correct output when running in NVIDIA TensorRT about mobilenet-ssd HOT 8 OPEN

Yes, your WeChat account?

from mobilenet-ssd.

Any progress?
I am also working on this.
If tensorRT+mobilenet works, please post your instruction here.

from mobilenet-ssd.

any update?

from mobilenet-ssd.

我最近也在研究用TensorRT来优化你的模型，能否加你的qq或者微信进行交流请教呀。谢谢 @chuanqi305 @toxic2m

from mobilenet-ssd.

@toxic2m Hi, what layers did you have to implement using the custom layer API in TensorRT? Are you able to provide the source for this?

from mobilenet-ssd.

@toxic2m @chuanqi305 can you guys elaborate on the status? share the implementation on TensorRT

from mobilenet-ssd.

Hi @chuanqi305 and all,

Sorry for my delay response, I don't login github frequently.
Could you please share your email?
You can touch me at [email protected].
BTW, I have already implemented your code in TensorRT 3.0. But unfortunately I can't share the source code.
You need to implement Depthwise layer with TensorRT plugin.
Thanks.

from mobilenet-ssd.

@toxic2m how to get the values of the convolution weights and bias in tensorRT?
Thank you

from mobilenet-ssd.