Search code examples
tensorflowobject-detectiongoogle-cloud-ml-enginegoogle-cloud-sdk

Error when following google object detection API tutorial

While following the tutorial in here, I've reached the stage of training the model on the cloud. Unfortunately, the command 

gcloud ml-engine jobs submit training `whoami`_object_detection_`date +%s` \
    --job-dir=gs://${YOUR_GCS_BUCKET}/train \
    --packages dist/object_detection-0.1.tar.gz,slim/dist/slim-0.1.tar.gz \
    --module-name object_detection.train \
    --region us-central1 \
    --config object_detection/samples/cloud/cloud.yml \
    -- \
    --train_dir=gs://${YOUR_GCS_BUCKET}/train \
    --pipeline_config_path=gs://${YOUR_GCS_BUCKET}/data/faster_rcnn_resnet101_pets.config

yields the following error:

ERROR: (gcloud.ml-engine.jobs.submit.training) FAILED_PRECONDITION: Field: package_uris Error: The provided GCS paths [gs://pet-detector-test/train/packages/[REMOVED]/slim-0.1.tar.gz, gs://pet-detector-test/train/packages/[REMOVED]/object_detection-0.1.tar.gz] cannot be read by service account service-[REMOVED]@cloud-ml.google.com.iam.gserviceaccount.com.

Note: I have removed some specific identifiers from the error message and changed them to [REMOVED]

The bucket seems to contain all the necessary data, enter image description hereenter image description here

Any idea how to proceed?

Solution

  • The problem was due to skipping a step in here.
    Specifically - use the following commands:

    gcloud config set project [your-project-id]
gcloud auth application-default login