GPU Support

[FlorianRuen]

Hi,

I'm using this API to detect human pose on many images (5-10 FPS), but it's very difficult to keep this speed to analysis without GPU support. (I have a machine with Tesla M60 GPU and i can't use it with this API)

When do you plan to deploy GPU usage ? I will try on my side to find edits to adapt the code with GPU usage

Thanks for your help,

Kind regards, Florian

[MLnick]

Hi @FlorianRuen thanks for your interest in the Model Asset Exchange and this model.

Currently we aren't directly supporting GPU docker images, however it should be possible to edit the Dockerfile to be based off the relevant NVIDIA CUDA docker base image (see https://github.com/NVIDIA/nvidia-docker).

You would then however need to install the required dependencies of the max-base image yourself (see https://github.com/IBM/MAX-Base/blob/master/requirements.txt for details).

Then the code may need adapting - perhaps tensorflow-gpu would be required (refer https://github.com/IBM/MAX-Human-Pose-Estimator/blob/master/requirements.txt) at a minimum.

Note we have not tested using a Docker GPU image, so can't be sure things would work even with these changes. However we'll be happy to assist with questions as you go.