Open MerlinPCarson opened 3 years ago
The Spokestack framework expects an encode and detect model. Should these be trained as 1 model end to end and then split into 2 different TFLite models after training?
This can be defined as one sequence but you can create different model by using different (input output) combination
The Spokestack framework expects an encode and detect model. Should these be trained as 1 model end to end and then split into 2 different TFLite models after training?