tensorflow / tensorflow/models

Android Tensorflow ObjectDetection lib_task_api default number of threads?

Open
#10,462 0 comments 0 reactions 3 assignees View on GitHub

@pkulzc is already working on this.

Since Jan 18, 2022.

models:research:odapi type:docs
Dominant language
Python
Stars
77.7k
Forks
44.8k
PR merge metrics
No merged PRs in 30d

Description

Hello,
We are working on implementing Tensorflow object detection based on the example provided here:
https://github.com/tensorflow/examples/tree/eb925e460f761f5ed643d17f0c449e040ac2ac45/lite/examples/object_detection/android
It seems that there are 2 implementations for that one:

(1) lib_task_api that leverages the out-of-box API from the TensorFlow Lite Task Library;
(2) lib_interpreter that creates the custom inference pipleline using the TensorFlow Lite Interpreter Java API.

We are using lib_task_api and we don't see the way how we can set the number of threads with it?
For the lib_interpreter - there is a variable that is being set up for that.
https://github.com/tensorflow/examples/blob/a228a3460f3fdd8edee9e8b061a08ffc92629907/lite/examples/object_detection/android/lib_interpreter/src/main/java/org/tensorflow/lite/examples/detection/tflite/TFLiteObjectDetectionAPIModel.java#L138
On the other hand - for lib_task_api i see that this method is deprecated and it is not set up
https://github.com/tensorflow/examples/blob/eb925e460f761f5ed643d17f0c449e040ac2ac45/lite/examples/object_detection/android/lib_task_api/src/main/java/org/tensorflow/lite/examples/detection/tflite/TFLiteObjectDetectionAPIModel.java#L87

My question being - what is the default number of threads for lib_task_api?
I am asking this one because i would like to tune it up for different set of devices.

Best Regards

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.