how to get string as true value of tfx.orchestration.data_types.RuntimeParameter during execution pipeline?
Hi,
I'm defining a runtime parameter like data_root = tfx.orchestration.data_types.RuntimeParameter(name='data-root', ptype=str) for a base path, from which I define many subfolders for various components like str(data_root)+'/model' for model serving path in tfx.components.Pusher().
It was working like a charm before I moved to tfx==1.12.0: str(data_root) is now providing a json dump.
To overcome that, i tried to define a runtime parameter for model path like model_root = tfx.orchestration.data_types.RuntimeParameter(name='model-root', ptype=str) and then feed the Pusher component the way I saw in many tutotrials:
pusher = Pusher(model=trainer.outputs['model'],
model_blessing=evaluator.outputs['blessing'],
push_destination=tfx.proto.PushDestination(
filesystem=tfx.proto.PushDestination.Filesystem(base_directory=model_root)))
but I get a TypeError saying tfx.proto.PushDestination.Filesystem does not accept Runtime parameter.
It completely breaks the existing setup as i received those parameters from external client for each kubeflow run.
Thanks a lot for any help.
I was able to fix it.
First of all, the docstring is not clear regarding which parameter of Pusher can be a RuntimeParameter or not.
I finally went to __init__ code definition of component Pusher to see that only the parameter push_destination can be a RuntimeParameter:
def __init__(
self,
model: Optional[types.BaseChannel] = None,
model_blessing: Optional[types.BaseChannel] = None,
infra_blessing: Optional[types.BaseChannel] = None,
push_destination: Optional[Union[pusher_pb2.PushDestination,
data_types.RuntimeParameter]] = None,
custom_config: Optional[Dict[str, Any]] = None,
custom_executor_spec: Optional[executor_spec.ExecutorSpec] = None):
Then I defined the component consequently, using my RuntimeParameter
model_root = tfx.orchestration.data_types.RuntimeParameter(name='model-serving-location', ptype=str)
pusher = Pusher(model=trainer.outputs['model'],
model_blessing=evaluator.outputs['blessing'],
push_destination=model_root)
As push_destination parameter is supposed to be message proto tfx.proto.pusher_pb2.PushDestination, you have then to respect the associated schema when instantiating and running a pipeline execution, meaning the value should be like:
{'type': 'model-serving-location': 'value': '{"filesystem": {"base_directory": "path/to/model/serving/for/the/run"}}'}
Regards
Related
I've added a Lambda Step as the first step in my Sagemaker Pipeline. It processes some data and creates 2 files as part of the output like so:
from sagemaker.workflow.lambda_step import LambdaStep, Lambda, LambdaOutput, LambdaOutputTypeEnum
# lamb_preprocess = LambdaStep(func_arn="")
output_param_1 = LambdaOutput(output_name="status", output_type=LambdaOutputTypeEnum.Integer)
output_param_2 = LambdaOutput(output_name="file_name_a_c_drop", output_type=LambdaOutputTypeEnum.String)
output_param_3 = LambdaOutput(output_name="file_name_q_c_drop", output_type=LambdaOutputTypeEnum.String)
step_lambda = LambdaStep(
name="ProcessingLambda",
lambda_func=Lambda(
function_arn="arn:aws:lambda:us-east-1:xxxxxxxx:function:xxxxx"
),
inputs={
"input_data": input_data,
"input_file": trigger_file,
"input_bucket": trigger_bucket
},
outputs = [
output_param_1, output_param_2, output_param_3
]
)
In my next step, I want to trigger a Processing Job for which I need to pass in the above Lambda function's outputs as it's inputs. I'm trying to do it like so:
inputs = [
ProcessingInput(source=step_lambda.properties.Outputs["file_name_q_c_drop"], destination="/opt/ml/processing/input"),
ProcessingInput(source=step_lambda.properties.Outputs["file_name_a_c_drop"], destination="/opt/ml/processing/input"),
]
However, when the processing step is trying to get created, I get a validation message saying
Object of type Properties is not JSON serializable
I followed the data dependency docs here: https://sagemaker.readthedocs.io/en/stable/amazon_sagemaker_model_building_pipeline.html#lambdastep and tried accessing step_lambda.OutputParameters["file_name_a_c_drop"] too but it errored out saying 'LambdaStep' object has no attribute 'OutputParameters'
How do I properly access the return value of a LambdaStep in a Sagemaker pipeline ?
You can access the output as follows - step_lambda.OutputParameters["output1"]. You don't need to add .properties
To access a LambdaStep output in another step you can do this:
step_lambda.properties.Outputs["file_name_a_c_drop"]
Try this
steplambda.properties.ProcessingOutputConfig.Outputs["file_name_q_c_drop"].S3Output.S3Uri
I'm trying to deploy my custom-trained model using a custom-container, i.e. create an endpoint from a model that I created.
I'm doing the same thing with AI Platform (same model & container) and it works fine there.
At the first try I deployed the model successfully, but ever since whenever I try to create an endpoint it says "deploying" for 1+ hours and then it fails with the following error:
google.api_core.exceptions.FailedPrecondition: 400 Error: model server never became ready. Please validate that your model file or container configuration are valid. Model server logs can be found at (link)
The log shows the following:
* Running on all addresses (0.0.0.0)
WARNING: This is a development server. Do not use it in a production deployment.
* Running on http://127.0.0.1:8080
[05/Jul/2022 12:00:37] "[33mGET /v1/endpoints/1/deployedModels/2025850174177280000 HTTP/1.1[0m" 404 -
[05/Jul/2022 12:00:38] "[33mGET /v1/endpoints/1/deployedModels/2025850174177280000 HTTP/1.1[0m" 404 -
Where the last line is being spammed until it ultimately fails.
My flask app is as follows:
import base64
import os.path
import pickle
from typing import Dict, Any
from flask import Flask, request, jsonify
from streamliner.models.general_model import GeneralModel
class Predictor:
def __init__(self, model: GeneralModel):
self._model = model
def predict(self, instance: str) -> Dict[str, Any]:
decoded_pickle = base64.b64decode(instance)
features_df = pickle.loads(decoded_pickle)
prediction = self._model.predict(features_df).tolist()
return {"prediction": prediction}
app = Flask(__name__)
with open('./model.pkl', 'rb') as model_file:
model = pickle.load(model_file)
predictor = Predictor(model=model)
#app.route("/predict", methods=['POST'])
def predict() -> Any:
if request.method == "POST":
instance = request.get_json()
instance = instance['instances'][0]
predictions = predictor.predict(instance)
return jsonify(predictions)
#app.route("/health")
def health() -> str:
return "ok"
if __name__ == '__main__':
port = int(os.environ.get("PORT", 8080))
app.run(host='0.0.0.0', port=port)
The deployment code which I do through Python is irrelevant because the problem persists when I deploy through GCP's UI.
The model creation code is as follows:
def upload_model(self):
model = {
"name": self.model_name_on_platform,
"display_name": self.model_name_on_platform,
"version_aliases": ["default", self.run_id],
"container_spec": {
"image_uri": f'{REGION}-docker.pkg.dev/{GCP_PROJECT_ID}/{self.repository_name}/{self.run_id}',
"predict_route": "/predict",
"health_route": "/health",
},
}
parent = self.model_service_client.common_location_path(project=GCP_PROJECT_ID, location=REGION)
model_path = self.model_service_client.model_path(project=GCP_PROJECT_ID,
location=REGION,
model=self.model_name_on_platform)
upload_model_request_specifications = {'parent': parent, 'model': model,
'model_id': self.model_name_on_platform}
try:
print("trying to get model")
self.get_model(model_path=model_path)
except NotFound:
print("didn't find model, creating a new one")
else:
print("found an existing model, creating a new version under it")
upload_model_request_specifications['parent_model'] = model_path
upload_model_request = model_service.UploadModelRequest(upload_model_request_specifications)
response = self.model_service_client.upload_model(request=upload_model_request, timeout=1800)
print("Long running operation:", response.operation.name)
upload_model_response = response.result(timeout=1800)
print("upload_model_response:", upload_model_response)
My problem is very close to this one with the difference that I do have a health check.
Why would it work on the first deployment and fail ever since? Why would it work on AI Platform but fail on Vertex AI?
This issue could be due to different reasons:
Validate the container configuration port, it should use port 8080.
This configuration is important because Vertex AI sends liveness
checks, health checks, and prediction requests to this port on the
container. You can see this document about containers, and this
other about custom containers.
Another possible reason is quota limits, which could need to be increased. You will be able to verify this using this document to do it
In the health and predict route use the MODEL_NAME you are using.
Like this example
"predict_route": "/v1/models/MODEL_NAME:predict",
"health_route": "/v1/models/MODEL_NAME",
Validate that the account you are using has enough permissions to
read your project's GCS bucket.
Validate the Model location, should be the correct path.
If any of the suggestions above work, it’s a requirement to contact GCP Support by creating a Support Case to fix it. It’s impossible for the community to troubleshoot it without using internal GCP resources
In case you haven't yet found a solution you can try out custom prediction routines. They are really helpful as they strip away the necessity to write the server part of the code and allows us to focus solely on the logic of our ml model and any kind of pre or post processing. Here is the link to help you out https://codelabs.developers.google.com/vertex-cpr-sklearn#0. Hope this helps.
I am trying to use a global class that I've defined in a shared library to help organise job parameters. It's not working, and I'm not even sure if it is possible.
My job looks something like this:
pipelineJob('My-Job') {
definition {
// Job definition goes here
}
parameters {
choiceParam('awsAccount', awsAccount.ALL)
}
}
In a file in /vars/awsAccount.groovy I have the following code:
class awsAccount implements Serializable {
final String SANDPIT = "sandpit",
final String DEV = "dev",
final String PROD = "prod"
static String[] ALL = [SANDPIT, DEV, PROD]
}
Global pipeline libraries are configured to load implicitly from the my repository's master branch.
When attempting to update the DSL scripts I receive the error:
ERROR: (myJob.groovy, line 67) No such property: awsAccount for class: javaposse.jobdsl.dsl.helpers.BuildParametersContext
Why does it not find the class, and is it even possible to use shared library classes like this in pipeline job?
Disclaimer: I know it works using Jenkinsfile. Unfortunatelly, not tested usng Declarative Pipelines - but no answers yet, so it may be worth a try
Regarding your first question: there are some reasons why a class from your shared-lib could not be found. Starting from the library import, the library syntax, etc. But they definitvely work for DSL. To be more precise about it, additional information would be great. But be sure that:
You have your groovy class definition using exactly the directory structure as described in the documentation (https://www.jenkins.io/doc/book/pipeline/shared-libraries/)
Give a name to the shared-lib in jenkins as you configure it and be sure is exactly the name you use in the import
Use the import as described in the documentation (under Using Libraries)
Regarding your second question (the one that names this SO question): yes, you can include parameter jobs from information in your shared-lib. At least, using Jenkinsfiles. You can even define properties to be included in the pipelie. I got it working with a tricky syntax due to different problems.
Again, I am using Jenkinsfile and this is what worked for me:
In my shared-lib class, I added a static function that introduces the build parameters. Notice the input parameters that function needs and its usage:
class awsAccount implements Serializable {
//
static giveMeParameters (script) {
return [
// Some parms
script.string(defaultValue: '', description: 'A default parameter', name: 'textParm'),
script.booleanParam(defaultValue: false, description: 'If set to True, do whatever you need - otherwise, do not do it', name: 'boolOption'),
]
}
}
To introduce those parameters in the pipeline, you need to place the returned value of the function into the parameters array
properties (
parameters (
awsAccount.giveMeParameters (this)
)
Again, notice the syntax when calling the function. Similar to this, you can also define functions in the shared-lib that return properties and use them in multiple jobs (disableConcurrentBuilds, buildDiscarder, etc)
I am using a dataflow template to run cloud dataflow
I am providing some default values and calling template. Dataflow shows the pipeline options correctly in the dataflow pipeline summary. but it's not taking the runtime values.
class Mypipeoptions(PipelineOptions):
#classmethod
def _add_argparse_args(cls, parser):
parser.add_value_provider_argument(
'--preprocess_indir',
help='GCS path of the data to be preprocessed',
required=False,
default='gs://default/dataset/'
)
parser.add_value_provider_argument(
'--output_dir_train',
help='GCS path of the preprocessed train data',
required=False,
default='gs://default/train/'
)
parser.add_value_provider_argument(
'--output_dir_test',
help='GCS path of the preprocessed test data',
required=False,
default='gs://default/test/'
)
parser.add_value_provider_argument(
'--output_dir_validate',
help='GCS path of the preprocessed validate data',
required=False,
default='gs://default/validate/'
)
Then I am checking the values are accessible
p = beam.Pipeline(options=args)
if args.preprocess_indir.is_accessible():
input_dir = args.preprocess_indir
else:
input_dir = getValObj(args.preprocess_indir)
if args.output_dir_train.is_accessible():
output_train = args.output_dir_train
else:
output_train = getValObj(args.output_dir_train)
if args.output_dir_test.is_accessible():
output_test = args.output_dir_test
else:
output_test = getValObj(args.output_dir_test)
if args.output_dir_validate.is_accessible():
output_validate = args.output_dir_validate
else:
output_validate = getValObj(args.output_dir_validate)
Now when calling the template, I could see the values I wanted being passed as (Mypipeoptions)pipeine option parmater but its not used in the actual run, Instead using default options given
I think I found the solution, I was assigning runtime parameters to variables and then passing it to the input or output.
When I directly passed the runtime parameters to source or sink it worked. Like the one below
'Write train dataset to destination' >> beam.io.tfrecordio.WriteToTFRecord(
file_path_prefix=args.output_dir_train
)
I believe that the part I missed was that when the template is created it builds the graph and only the runtime parameters can be plugged into its runtime. Other computations are already done when building a graph.
Please correct me if I am wrong
I'm using Dask to distribute work to a cluster. I'm creating a cluster and calling .submit() to submit a function to the scheduler. It returns a Futures object. I'm trying to figure out how to obtain the input arguments to that future object once it's been completed.
For example:
from dask.distributed import Client
from dask_yarn import YarnCluster
def somefunc(a,b,c ..., n ):
# do something
return
cluster = YarnCluster.from_specification(spec)
client = Client(cluster)
future = client.submit(somefunc, arg1, arg2, ..., argn)
# ^^^ how do I obtain the input arguments for this future object?
# `future.args` doesn't work
Futures don't hold onto their inputs. You can do this yourself though.
futures = {}
future = client.submit(func, *args)
futures[future] = args
A future only knows the key by which it is uniquely known on the scheduler. At the time of submission, if it has dependencies, these are transiently found and sent to the scheduler but no copy if kept locally.
The pattern you are after sounds more like delayed, which keeps hold of its graph, and indeed client.compute(delayed_thing) returns a future.
d = delayed(somefunc)(a, b, c)
future = client.compute(d)
dict(d.dask) # graph of things needed by d
You could communicate directly with the scheduler to find the dependencies of some key, which will in general also be keys, and so reverse-engineer the graph, but that does not sound like a great path, so I won't try to describe it here.