Closed james-bjss closed 4 years ago
It seems like reverting to older drivers 2.13.0 (Jan 2020)
resolved this issue.
However I was simply following the setup guide https://coral.ai/docs/accelerator/get-started/ which links to the latest 2.14.1 drivers and suggests running this example, so others may fall foul of this.
Hey, 2.14.1
should have been the same as 2.14
with added window supports. This issue is not expected, but please allow me to check with my team. I guess the added supports for pcie must have caused this, another user have experienced the same: https://github.com/google-coral/tflite/issues/16#issuecomment-651078588
Could you try upgrading to 2.14
(without the .1
)?
Hey,
2.14.1
should have been the same as2.14
with added window supports. This issue is not expected, but please allow me to check with my team. I guess the added supports for pcie must have caused this, another user have experienced the same: #16 (comment)Could you try upgrading to
2.14
(without the.1
)?
Sure, will try this evening and report back.
@Namburger - I can confirm that 2.14.0
is also working. So is just 2.14.1
that is affected.
@james-bjss perfect, thanks for confirmation, I'll report this with our team
@james-bjss as it turned out, we uploaded the wrong zipfile for that link :)
We just updated it, everything should works now if you redownload the zip file for 1.14.1
!
@Namburger - Oops! Just tested with the new Zip and it's working fine. Think this one can be closed now. Thanks for your help.
Perfect! Thanks for confirmation :)
Hi, When testing out my USB Coral device on Windows10 executing the classification example results in the following error:
c:\Users\James\tampopo\tf\coral\tflite\python\examples\classification>python classify_image.py --model models/mobilenet_v2_1.0_224_inat_bird_quant_edgetpu.tflite --labels models/inat_bird_labels.txt --input images/parrot.jpg Traceback (most recent call last): File "classify_image.py", line 122, in <module> main() File "classify_image.py", line 100, in main interpreter.allocate_tensors() File "C:\Users\James\AppData\Local\Programs\Python\Python37\lib\site-packages\tflite_runtime\interpreter.py", line 243, in allocate_tensors return self._interpreter.AllocateTensors() File "C:\Users\James\AppData\Local\Programs\Python\Python37\lib\site-packages\tflite_runtime\tensorflow_wrap_interpreter_wrapper.py", line 110, in AllocateTensors return _tensorflow_wrap_interpreter_wrapper.InterpreterWrapper_AllocateTensors(self) RuntimeError: Internal: Unsupported data type in custom op handler: 0Node number 1 (EdgeTpuDelegateForCustomOp) failed to prepare.
Further details: Python Version: Python 3.7.8rc1 tflite Runtime: tflite_runtime-2.1.0.post1-cp37-cp37m-win_amd64.whl edgetpu wheel: edgetpu-2.14.0-cp37-cp37m-win_amd64.whl Driver Package: 2.14.1 (Jun 2020) edgetpu_runtime_20200624.zip
Relates to: #9