I have created an image classification model (thanks to fastai team
) and now in implementation stage.
Created the program in such a way to load the weights(pkl file) only once and so to use the same instance for all predictions(This is for performance improvement)
Model Inference works perfectly .
But I ran into trouble when I set multithreading > 1 :
Below is the error I am getting when the prediction is invoked in parallel for 2 different threads:
Anyone please help on this.
Why we are not able to enable multithreading when the model is loaded only once!!!