How do I use a GPU on AI Platform Pipelines? My pipeline uses set_gpu_limit(1) in one of the ops but I end up getting a This step is in Pending state with this message: Unschedulable: 0/3 nodes are available: 3 Insufficient nvidia.com/gpu. error.
Got it a few mins later... I followed the normal Kubeflow on GPU instructions
export GPU_POOL_NAME=gpu-pool
export CLUSTER_NAME=cluster-1
gcloud container node-pools create ${GPU_POOL_NAME} \
--accelerator type=nvidia-tesla-k80,count=1 \
--zone us-central1-a --cluster ${CLUSTER_NAME} \
--num-nodes=0 --machine-type=n1-standard-4 --min-nodes=0 --max-nodes=1 --enable-autoscaling
kubectl apply -f https://raw.githubusercontent.com/GoogleCloudPlatform/container-engine-accelerators/master/nvidia-driver-installer/cos/daemonset-preloaded.yaml
Related
I'm trying to deploy a model to GCP AI Platform but not getting anywhere, I know similar question has been asked before, but I can't make out what I'm doing wrong, any help would be appreciated. I'm using jupyter notebook for development.
MODEL_NAME='CLASSIFIER'
VERSION_NAME='v1'
RUNTIME_VERSION='1.13'
REGION='europe-west1'
!gcloud ai-platform models create {MODEL_NAME} --regions {REGION}
The model is created in the global end point
!gcloud beta ai-platform versions create {VERSION_NAME} \
--model {MODEL_NAME} \
--origin gs://{BUCKET}/{MODEL_DIR} \
--python-version 3.7 \
--runtime-version {RUNTIME_VERSION} \
--package-uris gs://{BUCKET}/{PACKAGES_DIR}/sentiment_classifier-0.1.tar.gz \
--prediction-class=model_prediction.CustomPrediction \
--machine-type mls1-c4-m4
but when I try to create the version, it tries to create it in the regional end point and fails, or so I think
Using endpoint [https://us-central1-ml.googleapis.com/]
ERROR: (gcloud.beta.ai-platform.versions.create) INVALID_ARGUMENT: Machine type is not available on this endpoint.
I was able to run the command without issues. For it I use an existing sample to replicate the execution of the commands .
As a remark, both models point to https://ml.googleapis.com as this region accepts the execution of beta commands as the one used for this case. You can see details of it on this link under --region. As a note, you can choose different main regions but keep in mind that it should support the execution of beta commands and that both (the model and the version) points to the same endpoint.
model
MODEL_NAME = 'CensusPredictor'
VERSION_NAME = 'v1'
REGION = "global"
! gcloud ai-platform models create $MODEL_NAME \
--regions $REGION
output
WARNING: To specify a region where the model will deployed on the global endpoint, please use `--regions` and do not specify `--region`. Using [us-central1] by default on https://ml.googleapis.com. Please note that your model will be inaccessible from https://us-central1-ml.googelapis.com
Learn more about regional endpoints and see a list of available regions: https://cloud.google.com/ai-platform/prediction/docs/regional-endpoints
Using endpoint [https://ml.googleapis.com/]
Created ai platform model [projects/<my-project-id>/models/CensusPredictor].
beta ai-platform versions create
! gcloud beta ai-platform versions create $VERSION_NAME --model $MODEL_NAME \
--origin gs://$BUCKET_NAME/custom_pipeline_tutorial/model/ \
--runtime-version 1.13 \
--python-version 3.5 \
--framework SCIKIT_LEARN \
--package-uris gs://$BUCKET_NAME/custom_pipeline_tutorial/code/census_package-1.0.tar.gz \
--region $REGION \
--machine-type mls1-c1-m2
output
Using endpoint [https://ml.googleapis.com/]
Creating version (this might take a few minutes)......done.
I am trying to use TPU on Cloud ML Engine but I am at a loss as to how I should provide the tpu argument which TPUClusterResolver expects.
This is the environment I am using:
--python-version 3.5 \
--runtime-version 1.12 \
--region us-central1 \
--scale-tier BASIC_TPU
The job crashes with:
ValueError: Please provide a TPU Name to connect to.
As a separate issue - ML engine seems to be adding --master grpc://10.129.152.2:8470 on its own to my job which also crashes the job. As a workaround for it I just added an un-used master flag to my code.
this was a known issue for runtime 1.11 and 1.12 and it has been fixed. Now, the service won't append --master to your training application. You should continue using TpuClusterResolver.
I follow this tutorial to do the training on Google cloud ml-engine. I follow it step by step but I am facing error when submit the ml job to cloud. I ran this command.
sam#sam-VirtualBox:~/models/research$ gcloud ml-engine jobs submit training whoami_object_detection_date +%s --job-dir=gs://tf_testing/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://tf_testing/train --pipeline_config_path=gs://tf_testing/data/faster_rcnn_resnet101_pets.config
and I got this error.
ERROR: (gcloud.ml-engine.jobs.submit.training) FAILED_PRECONDITION: Field: package_uris Error: The provided GCS paths [gs://tf_testing/train/packages/8ec87a281aadb58d3d82462bbffafa9d7e521cc03025209704bc643eb9f3bc37/slim-0.1.tar.gz, gs://tf_testing/train/packages/8ec87a281aadb58d3d82462bbffafa9d7e521cc03025209704bc643eb9f3bc37/object_detection-0.1.tar.gz] cannot be read by service account service-499049193648#cloud-ml.google.com.iam.gserviceaccount.com. - '#type': type.googleapis.com/google.rpc.BadRequest fieldViolations: - description: The provided GCS paths [gs://tf_testing/train/packages/8ec87a281aadb58d3d82462bbffafa9d7e521cc03025209704bc643eb9f3bc37/slim-0.1.tar.gz, gs://tf_testing/train/packages/8ec87a281aadb58d3d82462bbffafa9d7e521cc03025209704bc643eb9f3bc37/object_detection-0.1.tar.gz] cannot be read by service account service-499049193648#cloud-ml.google.com.iam.gserviceaccount.com. field: package_uris
I saw this post and this post and tried the solution but it did not help. FYI, I did not change PATH_TO_BE_CONFIGURED when ran this command. Could it be the reason?
sed -i "s|PATH_TO_BE_CONFIGURED|"gs://${YOUR_GCS_BUCKET}"/data|g" \
object_detection/samples/configs/faster_rcnn_resnet101_pets.config
You need to to allow the service account to read/write to your bucket:
gsutil acl ch -u $SVCACCT:WRITE gs://$BUCKET/
gsutil defacl ch -u $SVCACCT:O gs://$BUCKET/
Alternately:
gcloud ml-engine init-project
Will add the service account as an editor on the project. Make sure to do this in the project that owns the bucket
I'm trying train a model using Tensorflow on the Google Cloud ml-engine. It seems that tensorflow can't get to the libcupti files on the cloud compute machine due to the LD_LIBRARY_PATH not pointing to the correct directory, as implied by the log entry below:
lineno: 126
message: "Couldn't open CUDA library libcupti.so.8.0.
LD_LIBRARY_PATH: /usr/local/cuda/lib64"
levelname: "INFO"
pathname: "tensorflow/stream_executor/dso_loader.cc"
created: 1491143889.84344
As far as I know, the libcupti files are all in /usr/local/cuda/extras/CUPTI/lib64, so I would need to append this to the LD_LIBRARY_PATH variable, but how would I do that when submitting a job via a gcloud ml-engine jobs submit training $JOB_NAME command? Or maybe there's an easier solution?
I tried to use GPU with tensorflow on google cloud and it works for me. In my code I didn't do any GPU specific setting (nor set anything with LD_LIBRARY_PATH)
I think you can try with just a simple and standard tensorflow code and with you submit the job you attach a config then the job should automatically use GPU to do the calculation for you.
Try add a file such as cloudml-gpu.yaml in your module with the following content:
trainingInput:
scaleTier: CUSTOM
# standard_gpu provides 1 GPU. Change to complex_model_m_gpu for 4
GPUs
masterType: standard_gpu
runtimeVersion: "1.0"
Then add a option called --config=trainer/cloudml-gpu.yaml (suppose your training module is in a folder called trainer). For example:
export BUCKET_NAME=tf-learn-simple-sentiment
export JOB_NAME="example_5_train_$(date +%Y%m%d_%H%M%S)"
export JOB_DIR=gs://$BUCKET_NAME/$JOB_NAME
export REGION=europe-west1
gcloud ml-engine jobs submit training $JOB_NAME \
--job-dir gs://$BUCKET_NAME/$JOB_NAME \
--runtime-version 1.0 \
--module-name trainer.example5-keras \
--package-path ./trainer \
--region $REGION \
--config=trainer/cloudml-gpu.yaml \
-- \
--train-file gs://tf-learn-simple-sentiment/sentiment_set.pickle
when I use the command pip install tensorflow the download is only 99% complete and terminated at that point. How can I install tensorflow using google cloud shell.
Instead of installing it by yourself you can use the machine learning api and use TensorFlow for training or inference. Just follow this guidelines: https://cloud.google.com/ml/docs/quickstarts/training
You can submit a TensorFlow job like this:
gcloud beta ml jobs submit training ${JOB_NAME} \
--package-path=trainer \
--module-name=trainer.task \
--staging-bucket="${TRAIN_BUCKET}" \
--region=us-central1 \
-- \
--train_dir="${TRAIN_PATH}/train"