Open jwnz opened 4 years ago
Loading the model for predictions takes a long time.
This is problematic when handling data in small batches as the model has to be loaded for every batch.
Is there a way of preloading the model for faster inference of small batches when using the Python API?
This may be a duplicate of #298
Loading the model for predictions takes a long time.
This is problematic when handling data in small batches as the model has to be loaded for every batch.
Is there a way of preloading the model for faster inference of small batches when using the Python API?
This may be a duplicate of #298