Intel® Distribution of OpenVINO™ Toolkit
Community assistance about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all aspects of computer vision-related on Intel® platforms.
6367 Discussions

if possible to add two or more model trained in just one?

NMitr2
Beginner
539 Views

Hi all,

I have a question, is possible to add two or more different model trained with different object to create just one model with all this objects??

 

so I explain me better:

for example I have three models trained separately with three different objects (three different frozen_inference files)

model 1 = face,

model 2 = lamp,

model 3 = chair....

 

my question is :

can I take this three different files and make just a file froze_inference (model_tot = model 1 + model 2+ model 3) without to re-training the dataset for to get the output model with face,lamp and chair?

 

many thanks for your help.....

 

Regards

Nicola

0 Kudos
5 Replies
RandallMan_B_Intel
393 Views

Hi, NMitr2,

 

Thank you for reaching out. 

 

Please let me check about this case, I will back to you a soon as possible.

  

Best regards,

Randall B.

0 Kudos
NMitr2
Beginner
393 Views

many thanks for you help Randall, I have my own idea but I would like to have other opinion, I hope you could answer as well to me.......

 

Regards

Nicola

0 Kudos
RandallMan_B_Intel
393 Views

Hi NMitr2,

 

Thank you for patience.

Unfortunately, it’s not possible to inference multiple models as you need with OpenVINO. We have a demo where multiple models are loaded to the device. However, the results of the first model are then fed as the input of the second model for inference. You can see the Security Barrier Camera Demo.

We recommend to you combine the image data and re-train one network with all the objects. 

 

Regards,

Randall B.

0 Kudos
NMitr2
Beginner
393 Views

Hi RandallB,

many thanks for your answer I thought the same thing....

 

Regards

Nicola

0 Kudos
RandallMan_B_Intel
393 Views

Hi NMitr2,

 

Great, let me know if you have additional questions. Have a nice day!

 

Regards,

Randall B.

0 Kudos
Reply