how to enable mkldnn on paddle engine

[ruhengChen]

Description

I tried .optOptions(Collections.singletonMap("enableMKLDNN", "true")) to create Criteria and find the options of PpModel added indeead, but it did not work. how to fix this?

Expected Behavior

enable mkldnn on paddle engine

Environment Info

os.name: Windows 10 amd64

------------------ CUDA -----------------
[DEBUG] - No cudart library found in path.
GPU Count: 0

----------------- Engines ---------------
DJL version: 0.18.0
Default Engine: MXNet
[DEBUG] - Using cache dir: C:\Users\CRH\.djl.ai\mxnet\1.9.0-mkl-win-x86_64
[INFO ] - Downloading libgcc_s_seh-1.dll ...
[INFO ] - Downloading libgfortran-3.dll ...
[INFO ] - Downloading libopenblas.dll ...
[INFO ] - Downloading libquadmath-0.dll ...
[INFO ] - Downloading mxnet.dll ...
[DEBUG] - Loading mxnet library from: C:\Users\CRH\.djl.ai\mxnet\1.9.0-mkl-win-x86_64\mxnet.dll
Default Device: cpu()
PyTorch: 2
MXNet: 0
XGBoost: 10
TensorFlow: 3

[mymagicpower]

For paddlepaddle this feature is not enabled (i think). Contact DJL team to get further/deep engine support. For other engines, refer to below link: http://docs.djl.ai/docs/development/inference_performance_optimization.html