Conflicts between libraries developed based on handtrackingGpu and other TensorFlow based libraries

[chensisi0730]

Have I written custom code (as opposed to using a stock example script provided in MediaPipe)

No

OS Platform and Distribution

ios

MediaPipe Tasks SDK version

0.10.1

Task name (e.g. Image classification, Gesture recognition etc.)

handtrackingGpu

Programming Language and version (e.g. C++, Python, Java)

oc

Describe the actual behavior

it runs correctly

Describe the expected behaviour

it runs correctly

Standalone code/steps you may have used to try to get what you need

#0    0x000000010979c864 in TfLiteStatus tflite::ops::builtin::strided_slice::Eval<(tflite::ops::builtin::strided_slice::KernelType)0>(TfLiteContext*, TfLiteNode*) ()

Other info / Complete Logs

No response

[kuaashish]

@chensisi0730,

Could you kindly provide further elaboration on your query, including pertinent details such as the operating system you are utilising and a comprehensive description of the steps you are following from our documentation to reproduce the issue on our end?

[chensisi0730]

@kuaashish css-intel-Pro-2 bazel % brew config HOMEBREW_VERSION: 4.1.11 ORIGIN: https://github.com/Homebrew/brew HEAD: 4afb8e5602f3ecc9edf67a44257d8eceeaa8a108 Last commit: 3 days ago Core tap origin: https://github.com/Homebrew/homebrew-core Core tap HEAD: d793577b4ad83222bb580364ff66b1c5fede26a6 Core tap last commit: 2 weeks ago Core tap branch: master Core tap JSON: 14 Sep 03:31 UTC HOMEBREW_PREFIX: /usr/local HOMEBREW_CASK_OPTS: [] HOMEBREW_MAKE_JOBS: 8 Homebrew Ruby: 2.6.10 => /usr/local/Homebrew/Library/Homebrew/vendor/portable-ruby/2.6.10_1/bin/ruby CPU: octa-core 64-bit icelake Clang: 14.0.0 build 1400 Git: 2.33.0 => /usr/local/bin/git Curl: 8.1.2 => /usr/bin/curl macOS: 13.5.1-x86_64 CLT: 14.2.0.0.1.1668646533 Xcode: 14.2 => /Users/zhiyun/Downloads/Xcode.app/Contents/Developer

Bazel 6.1.1
mediapipe 0.10.1

I followed the official steps to compile handtrackingGpu, refer to https://developers.google.com/mediapipe/framework/getting_started/ios

The library developed based on handtrackingGpu conflicts with another library based on tflite 2.2. Another library called Nama(By command : pod 'Nama', '7.3.2') is also based on TensorFlow. Running handtrackingGpu separately will not cause a crash. If Nama's module is run, it will cause a runtime error, with the following error message as shown above

Nama compatible with Mediapipe version (near data 2021-03) , how can i downgrade the version of TensorFlow with mediaPipe ?

[kuaashish]

Hello @schmidt-sebastian,

Could you please look into this issue. Thank you

[kuaashish]

Hi @chensisi0730,

We would like to let you know that we've introduced a new hand Landmarker Task API for iOS. You can find the documentation to help you build it into your iOS application. This new API is easier to use and doesn't much rely on additional dependencies like our old hand solution did (we've stopped maintaining that one).

We encourage you to check out our new hand Task API for iOS, especially if you have a specific use case in mind. If you need any support, please feel free to reach out.

Thank you

[github-actions[bot]]

This issue has been marked stale because it has no recent activity since 7 days. It will be closed if no further activity occurs. Thank you.

[github-actions[bot]]

This issue was closed due to lack of activity after being marked stale for past 7 days.