tensorflow / serving

A flexible, high-performance serving system for machine learning models
https://www.tensorflow.org/serving
Apache License 2.0
6.18k stars 2.19k forks source link

Docs for batching parameter `pad_variable_length_inputs` #1934

Open jeongukjae opened 2 years ago

jeongukjae commented 2 years ago

When I use TF-Serving with batching options and variable length inputs, I could get errors like Tensors with name 'example_feature:0' from different tasks have different shapes and padding is turned off.Set pad_variable_length_inputs to true, or ensure that all tensors with the same namehave equal dimensions starting with the first dim., and I could find codes related to that parameters.

But in batching guide(https://github.com/tensorflow/serving/blob/master/tensorflow_serving/batching/README.md), there is no description about parameter pad_variable_length_inputs.

Can you add descriptions for pad_variable_length_inputs in the batching guide?

denadai2 commented 2 years ago

+1

YoungXu06 commented 2 years ago

+1

singhniraj08 commented 1 year ago

@jeongukjae,

Descriptions on pad_variable_length_inputs batching parameter can be found in tensorflow_serving/batching/batching_options.h file.

Thank you!

jeongukjae commented 1 year ago

@singhniraj08 Thank you for your help :) I was able to find the related code and understand it, but I think it would be even more helpful if this argument was added to the "TensorFlow Serving Batching Guide" documentation.

singhniraj08 commented 1 year ago

@christisg,

Can we add pad_variable_length_inputs parameter description in TensorFlow Serving Batching Guide. I can create a PR to update the doc.

Thank you!