wasi_nn_openvino.c: implement multiple models per instance

[yamt]

tested with two models: (with #4379)

--load-graph=id=graph1,file=public/license-plate-recognition-barrier-0007/FP32/license-plate-recognition-barrier-0007.xml,file=public/license-plate-recognition-barrier-0007/FP32/license-plate-recognition-barrier-0007.bin \
--load-graph=id=graph2,file=classify/model.xml,file=classify/model.bin \
--init-execution-context=id=exec1,graph-id=graph1 \
--init-execution-context=id=exec2,graph-id=graph2 \
--set-input=context-id=exec1,dim=1,dim=24,dim=94,dim=3,file=out.bin \
--set-input=context-id=exec2,file=classify/banana-3x224x224-bgr.bin,dim=1,dim=3,dim=224,dim=224 \
--compute=context-id=exec1 \
--compute=context-id=exec2 \
--get-output=context-id=exec1,file=exec1-result.bin \
--get-output=context-id=exec2,file=exec2-result.bin

[lum1n0us]

Would you mind providing detailed instructions so people, like me, can reproduce it locally?

[yamt]

Would you mind providing detailed instructions so people, like me, can reproduce it locally?

the classification model/input are taken from WasmEdge-WASINN-examples/openvino-mobilenet-raw.
i guess you are familiar with it.
(i think i transposed the input tensor manually to match the shape. but it isn't essential to test this PR.)

you can prepare license-plate-recognition-barrier-0007 model by

git clone https://github.com/openvinotoolkit/open_model_zoo

virtualenv venv
. ./venv/bin/activate

pip install openvino-dev
pip install -r open_model_zoo/tools/model_tools/requirements-tensorflow.in
pip install open_model_zoo/tools/model_tools

omz_downloader --name license-plate-recognition-barrier-0007
omz_converter --name license-plate-recognition-barrier-0007

i don't remember how i prepared out.bin.
i guess i cropped https://github.com/openvinotoolkit/open_model_zoo/blob/master/models/public/license-plate-recognition-barrier-0007/assets/license-plate-recognition-barrier-0007.png manually and converted it with a one-time python script using opencv and numpy.