Skip to content

cannot build TFRA-enabled TF-serving binary for TFRA 0.8.0 #486

Open
@huangenyan

Description

@huangenyan

When trying to complie the TFTA-enabled serving binary, I following the instructions in README, I use tensorflow/serving.git branch r2.16 according to the compatibility matrix:

## If enable GPU OPs
export SERVING_WITH_GPU=1 

## Specifiy the branch of TFRA
export TFRA_BRANCH="master" # The `master` and `r0.6` are available.

## Create workspace, modify the directory as you prefer to.
export TFRA_SERVING_WORKSPACE=~/tfra_serving_workspace/
mkdir -p $TFRA_SERVING_WORKSPACE && cd $TFRA_SERVING_WORKSPACE

## Clone the release branches of serving and TFRA according to `Compatibility Matrix`.
git clone -b r2.16 https://github.com/tensorflow/serving.git
git clone -b $TFRA_BRANCH https://github.com/tensorflow/recommenders-addons.git

## Run config shell script
cd $TFRA_SERVING_WORKSPACE/recommenders-addons/tools
bash config_tfserving.sh $TFRA_BRANCH $TFRA_SERVING_WORKSPACE/serving $SERVING_WITH_GPU

## Build serving with TFRA OPs.
cd $TFRA_SERVING_WORKSPACE/serving
./tools/run_in_docker.sh bazel build tensorflow_serving/model_servers:tensorflow_model_server

and encounter the following error:

Status: Image is up to date for tfra/serving:2.8.3-devel
docker.io/tfra/serving:2.8.3-devel
== Running cmd: sh -c 'cd /home/ec2-user/tfra_serving_workspace/serving; TEST_TMPDIR=.cache bazel build tensorflow_serving/model_servers:tensorflow_model_server'
$TEST_TMPDIR defined: output root default is '/home/ec2-user/tfra_serving_workspace/serving/.cache' and max_idle_secs default is '15'.
Starting local Bazel server and connecting to it...
ERROR: Traceback (most recent call last):
        File "/home/ec2-user/tfra_serving_workspace/serving/WORKSPACE", line 40, column 15, in <toplevel>
                versions.check("6.4.0")
        File "/home/ec2-user/tfra_serving_workspace/serving/.cache/_bazel_ec2-user/c41472d4aa52dc503a9f5911f5d7f7bc/external/bazel_skylib/lib/versions.bzl", line 105, column 13, in _check_bazel_version
                fail("Current Bazel version is {}; expected at least {}".format(
Error in fail: Current Bazel version is 3.7.2; expected at least 6.4.0
ERROR: error loading package '': Encountered error while reading extension file 'python/repositories.bzl': no such package
'@rules_python//python': error loading package 'external': Could not load //external package
INFO: Elapsed time: 7.755s
INFO: 0 processes.
FAILED: Build did NOT complete successfully (0 packages loaded)

I check the code in tools/config_tfserving.sh and find it uses a docker image on Dockerhub tfra/serving:2.8.3-devel as the compilation enviroment, but the environment doesn't support building 2.16 TF-serving app. Is it possible to add other images to support different versions?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions