search

google-coral / libedgetpu

Source code for the userspace level runtime driver for Coral.ai devices.
Apache License 2.0
179 stars 60 forks source link

TENSORFLOW_COMMIT for libedgetpu v16.0 #31

Closed chunyeow closed 2 years ago

chunyeow commented 2 years ago

Description

What is the TENSORFLOW_COMMIT for the following libedgetpu (16) version: ii libedgetpu-dev 16.0 arm64 Development files for libedgetpu ii libedgetpu1-std:arm64 16.0 arm64 Support library for Edge TPU

Any source of reference?

Click to expand! ### Issue Type Build/Install ### Operating System Ubuntu ### Coral Device Dev Board Mini ### Other Devices _No response_ ### Programming Language Other ### Relevant Log Output ```shell ./mendelcam --model ../all_models/mobilenet_v2_1.0_224_quant_edgetpu.tflite --labels ../all_models/imagenet_labels.txt ERROR: Internal: Unsupported data type in custom op handler: 0 ERROR: Node number 1 (EdgeTpuDelegateForCustomOp) failed to prepare. Error at inferencewrapper.cc:60 ```
hjonnala commented 2 years ago

are you looking for this: https://github.com/google-coral/libedgetpu/blob/master/workspace.bzl#L9 ?

chunyeow commented 2 years ago

$ python3 -c 'print(import("tflite_runtime").version)' 2.5.0.post1

The libedgetpu-dev over my coral dev mini is upgraded to version 16.0. So which TENSORFLOW_COMMIT should I used for?

Based on my understanding below: "Clone the TensorFlow repo using the TENSORFLOW_COMMIT version specified here—that's the version used to build the libedgetpu.so library, so your TensorFlow version must match. "

So I am looking for TENSORFLOW_COMMIT version that I should use to build my TensorFlow.

hjonnala commented 2 years ago

please use this commit: https://github.com/google-coral/libedgetpu/blob/master/workspace.bzl#L9

google-coral-bot[bot] commented 2 years ago

Are you satisfied with the resolution of your issue? Yes No