simple_tensorflow_serving
simple_tensorflow_serving copied to clipboard
batching feature
Is there a feature of batch merge calculation to improve GPU utilization? There is a batch merge in tensorflow-serving, which is very helpful for development and deployment. thanks