I have multiple inputs as src, f1, f2, f3, f4.
src is the image to be processed.f1 - f4 are the initial input for RNN 1-4 ,their channnels are more than 16 and their initial values are 0.0f, so I can't input them with a all-zero image with snpe-net-run(input_list),because channel number of image must <=4.
So what should I do to input the r1-r4 into model with snpe-net-run.
Opinions expressed in the content posted here are the personal opinions of the original authors, and do not necessarily reflect those of Qualcomm Incorporated or its subsidiaries (“Qualcomm”). The content is provided for informational purposes only and is not meant to be an endorsement or representation by Qualcomm or any other party. This site may also provide links or references to non-Qualcomm sites and resources. Qualcomm makes no representations, warranties, or other commitments whatsoever about any non-Qualcomm sites or third-party resources that may be referenced, accessible from, or linked to this site.
Dear customer,
Not very clear on your question.
Do you mean , you have 5 inputs for your model?
We support multi-inputs for the given model.
Thanks.
Hi,
I did mean that I have 5 inputs, like you said. But it is distinct from 5 image inputs, 4 input of these 5 inputs is feature map, eg. (1, 16, 512, 512), snpe-net-run can support it? if it can't, what should I do?
Can you understand my words?
Dear customer,
The mulitple input is determined by your model input. If your model just trained on input, you need to fill with same dimension input to list.
Otherwise, your model have multiple inputs after trained, you can speicific muliple input like as below into your list file.
input1:=data/1.raw input2:=data/2.raw <more if needs>
BR.
Wei