GPU usage for fast RCNN object detector

I currently use MATLAB 20107a and a GTX1050 GPU. I am trying to use my GPU for training a faster RCNN object detector using the function trainFasterRCNNObjectDetector by setting Execution Environment to gpu in options. But I still notice that my GPU is not being engaged. How do I solve this and what is going wrong? I would really appreciate an answer as soon as possible because I am in a time crunch.

[akhil s anand]

Hello, I am facing the same problem, could u find any solution/reason???. This happens only with training faster rcnn.

Thanks

[Ian Benderitter]

Hi, same for me. Exactly. I tried using a parallel pool of GPUs, or only one at the time, same issue : it keeps running it on CPU. I'm out of ideas to force it to run on GPU.

[Ram Shtoltz]

Hi, I have the same issue. Were you able to solve this?

[Ian Benderitter]

Personally still not, been pulling my hair off for days on that issue. I might consider voodoo next.

[Ghada Khaled]

I also have this issue, how can I specify if I want the run to use the GPU instead of the CPU

[nakamura takehisa]

こんばんは。私は日本人です。 うまく翻訳されればいいのですが・・・

私もFaster-RCNN を使っています。 大学のPCなのでセッティングは特にしませんでしたが、実行するにはParallelComputingToolboxなどのToolboxが必要です。

そして、並列計算をするにはMATLAB画面の左下にあるこのボタン（添付の画像）を押し、「並列プールの起動」を押す必要が

あります。

期待した回答じゃないかもしれません。私が答えれるのはここまでです。 もう少し、あなたの行っている設定などを示すのも有効かもしれません。 うまくいくことを願っています。