Implementing TFlite models

General Discussion
, ,
#1

I have recently run a model on a custom dataset. Deployed it in the android app. When I run it in the android app it prints out random integers like 2.99231231,0.2742424,-0.9953453,-0.775656.
Can someone tell me what is meaning of this?
How can I know what tflite.model outputs.
Please answer in details.
Thank You.

#3

Hi @Varun_Singh

Please explain what sort of model is this so we can assess the “integer” outputs. Also upload it somewhere so we can view it on netron.app

Best

#5

@George_Soloupis
it’s mobilenet v2 model.
Ok I will try and let you know

#7

@George_Soloupis The output is in the form StatefulPartitionedCall:0
can you tell me how can I access this ?

#8

@George_Soloupis similar model i trained
Identity 224
name: Identity
type: float32[1,4]
location: 224
can you tell me how I can access this also?