How to run multiple keras programs on single gpu?

Question

I am working on a python project where i need to build multiple Keras models for each dataset. Here when i run a Keras model building the program is using 10% of my GPU(GTX

Answer 1

In the section of Advanced solution - Using the same model more than once in parallel to double speed, i found that my models could not accurately predict the result.

I want to parallel caculating multiple keras models with same input and output shape but different fedding data, these models share the same structure but different weight matrixs. Besides, how to set multiple loss function to different sub-models insted of 'mse'. By using your methods, there is only one squential layer inside the model, so i wonder will out2 and out3 share the same weight matrix in the model?