Package Details: libedgetpu-git 20240314-2

Git Clone URL: https://aur.archlinux.org/libedgetpu-git.git (read-only, click to copy)
Package Base: libedgetpu-git
Description: Google Coral Edge TPU library and headers
Upstream URL: https://github.com/google-coral/libedgetpu
Licenses: Apache 2.0
Conflicts: libedgetpu, libedgetpu-max, libedgetpu-std
Provides: libedgetpu
Submitter: petrbel
Maintainer: petrbel (honzov, tomascapek, floop)
Last Packager: petrbel
Votes: 1
Popularity: 0.020040
First Submitted: 2024-11-26 14:04 (UTC)
Last Updated: 2025-01-07 15:08 (UTC)

Latest Comments

tlunter commented on 2025-06-08 20:49 (UTC)

Hi folks, instead of reverting to the old 24.3 flatbuffers file in the central package repository, I modified the PKGBUILD to pull the required version of flatbuffers, and add that to the libedgetpu Makefile so it installs cleanly without needing to look anything up: https://gist.github.com/tlunter/37a19da8394c0d431df0ed59c54412b5

Hope this helps!

sectoidman commented on 2025-05-18 05:52 (UTC)

I am also unable to build, even after installing the version of flatbuffers (24.3.25-3-) listed below.

/home/sectoidman/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../driver/usb/usb_standard_commands.cc: In member function ‘platforms::darwinn::StatusOr<platforms::darwinn::driver::UsbStandardCommands::DeviceDescriptor> platforms::darwinn::driver::UsbStandardCommands::GetDeviceDescriptor()’:
/home/sectoidman/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../driver/usb/usb_standard_commands.cc:51:26: error: cannot convert ‘platforms::darwinn::gtl::MutableArraySlice<unsigned char>’ to ‘int’
   51 |                     gtl::MutableArraySlice<uint8>(descriptor_buffer,
      |                          ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
      |                          |
      |                          platforms::darwinn::gtl::MutableArraySlice<unsigned char>
   52 |                                                   sizeof(descriptor_buffer)),
      |                                                   ~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/sectoidman/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../port/default/status_macros.h:227:51: note: in definition of macro ‘RETURN_IF_ERROR’
  227 |     const ::platforms::darwinn::Status _status = (expr); \
      |                                                   ^~~~
/home/sectoidman/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../driver/usb/usb_standard_commands.h:164:58: note: initializing argument 3 of ‘platforms::darwinn::Status platforms::darwinn::driver::UsbStandardCommands::GetDescriptor(platforms::darwinn::driver::UsbDeviceInterface::DescriptorType, int, int, size_t*, const char*)’
  164 |                        uint8_t desc_index, MutableBuffer data_in,
      |                                            ~~~~~~~~~~~~~~^~~~~~~
/home/sectoidman/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../driver/usb/usb_standard_commands.cc:73:14: error: ‘struct platforms::darwinn::driver::UsbStandardCommands::DeviceDescriptor’ has no member named ‘manufacturer_name_index’
   73 |   descriptor.manufacturer_name_index = descriptor_buffer[14];
      |              ^~~~~~~~~~~~~~~~~~~~~~~
/home/sectoidman/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../driver/usb/usb_standard_commands.cc:74:14: error: ‘struct platforms::darwinn::driver::UsbStandardCommands::DeviceDescriptor’ has no member named ‘product_name_index’
   74 |   descriptor.product_name_index = descriptor_buffer[15];
      |              ^~~~~~~~~~~~~~~~~~
/home/sectoidman/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../driver/usb/usb_standard_commands.cc:75:14: error: ‘struct platforms::darwinn::driver::UsbStandardCommands::DeviceDescriptor’ has no member named ‘serial_number_index’
   75 |   descriptor.serial_number_index = descriptor_buffer[16];
      |              ^~~~~~~~~~~~~~~~~~~
/home/sectoidman/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../driver/usb/usb_standard_commands.cc: At global scope:
/home/sectoidman/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../driver/usb/usb_standard_commands.cc:85:1: error: no declaration matches ‘platforms::darwinn::StatusOr<platforms::darwinn::driver::UsbStandardCommands::ConfigurationDescriptor> platforms::darwinn::driver::UsbStandardCommands::GetConfigurationDescriptor(uint8_t, size_t)’
   85 | UsbStandardCommands::GetConfigurationDescriptor(uint8_t index,

petrbel commented on 2025-01-15 09:16 (UTC)

@TonyB I suggest you referring to ArchLinux wiki which is in my opinion an excelent source of knowledge. Specifically, I suggest reading the page describing the archive: https://wiki.archlinux.org/title/Arch_Linux_Archive

By following the wiki, you are able to navigate yourself to this overview of all archived flatbuffers versions: https://archive.archlinux.org/packages/f/flatbuffers/ Then, simple pick a version you'd like to install.

Finally, you should be able to install the flatbuffers e.g. by pacman -U https://archive.archlinux.org/packages/f/flatbuffers/flatbuffers-24.3.25-3-x86_64.pkg.tar.zst

Hope this helps!

TonyB commented on 2025-01-14 19:52 (UTC)

@petrbel I'm new to Arch. Could you provide the command to install flatbuffers to 24.3. I tried a few attempts and googled, but had no success.

paulbrause commented on 2025-01-07 23:41 (UTC)

@petbel The "ugly" workaround worked in the end. My Google Coral Accelerator is still not running, but I was able to work around the error with this package. Thank you very much. :)

petrbel commented on 2025-01-07 15:15 (UTC)

@paulbrause Thank you for the submission. I was able to reproduce it. Based on what I've found, blatbuffers has a new release 24.12 (previously 24.3). Unfortunatelly, TensorFlow asserts blatbuffers' version to 24.3 (not only in the used TF version but also in master branch).

I updated libedgetpu-git's requirements so that they reflect this fact and provide a more sensible error message during the package dependency check.

If I may, I suggest downgrading flatbuffers to 24.3 which should solve the problem. Of course, that is only an ugly workaround solution. Nevertheless, unless TensorFlow requirements are updated, I'm afraid I don't know what to do more. Please feel free to suggest any other/better solution. Thanks!

paulbrause commented on 2025-01-07 00:38 (UTC) (edited on 2025-01-07 00:47 (UTC) by paulbrause)

When trying to install this package, the building fails for me:

Compiling tensorflow-2.18.0/tensorflow/lite/core/c/common.cc
In Datei, eingebunden von tensorflow-2.18.0/tensorflow/lite/schema/schema_generated.h:18,
    von tensorflow-2.18.0/tensorflow/lite/util.cc:30:
tensorflow-2.18.0/tensorflow/compiler/mlir/lite/schema/schema_generated.h:26:41: Fehler: statische Assertion fehlgeschlagen: Non-compatible flatbuffers version included
    26 |               FLATBUFFERS_VERSION_MINOR == 3 &&
       |                                         ^
tensorflow-2.18.0/tensorflow/compiler/mlir/lite/schema/schema_generated.h:26:41: Anmerkung: der Vergleich wird zu »(12 == 3)« reduziert
Compiling tensorflow-2.18.0/tensorflow/lite/array.cc
Compiling /home/smart/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../tflite/edgetpu_context_direct.cc
Compiling /home/smart/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../tflite/edgetpu_manager_direct.cc
tensorflow-2.18.0/tensorflow/lite/util.cc: In Funktion »TfLiteStatus tflite::BytesRequired(TfLiteType, const int*, size_t, size_t*, TfLiteContext*)«:
tensorflow-2.18.0/tensorflow/lite/util.cc:200:21: Warnung: Vergleich von Ganzzahlausdrücken, von denen nur einer ein Vorzeichen hat: »int« und »size_t« {aka »long unsigned int«} [-Wsign-compare]
    200 |   for (int k = 0; k < dims_size; k++) {
        |                   ~~^~~~~~~~~~~
make: *** [makefile_build/Makefile:192: /home/smart/.cache/yay/libedgetpu-git/src/libedgetpu/makefile_build/../out/tensorflow-2.18.0/tensorflow/lite/util.o] Fehler 1
make: *** Es wird auf noch nicht beendete Prozesse gewartet …
==> FEHLER: Ein Fehler geschah in build().
Breche ab...
-> Fehler beim Erstellen: libedgetpu-git-exit status 4
-> Die folgenden Pakete konnten nicht installiert werden. Ein manueller Eingriff ist erforderlich:
libedgetpu-git - exit status 4

Any idea how to fix this?