OCI Python SDKを使用したモデルの作成および保存

# Create a default config using DEFAULT profile in default location
# Refer to
# https://docs.cloud.oracle.com/en-us/iaas/Content/API/Concepts/sdkconfig.htm#SDK_and_CLI_Configuration_File
# for more info
 
import oci
from oci.data_science.models import CreateModelDetails, Metadata, CreateModelProvenanceDetails, UpdateModelDetails, UpdateModelProvenanceDetails
config = oci.config.from_file()
data_science_client = oci.data_science.DataScienceClient(config=config)
 
# Initialize service client with user principal (config file)
config = oci.config.from_file()
data_science_client = oci.data_science.DataScienceClient(config=config)
 
# Alternatively initialize service client with resource principal (for example in a notebook session)
# auth = oci.auth.signers.get_resource_principals_signer()
# data_science_client = oci.data_science.DataScienceClient({}, signer=auth)

provenance_details = CreateModelProvenanceDetails(repository_url="EXAMPLE-repositoryUrl-Value",
                                                  git_branch="EXAMPLE-gitBranch-Value",
                                                  git_commit="EXAMPLE-gitCommit-Value",
                                                  script_dir="EXAMPLE-scriptDir-Value",
                                                  # OCID of the ML job Run or Notebook session on which this model was
                                                  # trained
                                                  training_id="<Notebooksession or ML Job Run OCID>"
                                                  )

# create the list of defined metadata around model taxonomy:
defined_metadata_list = [
    Metadata(key="UseCaseType", value="image_classification"),
    Metadata(key="Framework", value="keras"),
    Metadata(key="FrameworkVersion", value="0.2.0"),
    Metadata(key="Algorithm",value="ResNet"),
    Metadata(key="hyperparameters",value="{\"max_depth\":\"5\",\"learning_rate\":\"0.08\",\"objective\":\"gradient descent\"}")
]

# Adding your own custom metadata:
custom_metadata_list = [
    Metadata(key="Image Accuracy Limit", value="70-90%", category="Performance",
             description="Performance accuracy accepted"),
    Metadata(key="Pre-trained environment",
             value="https://blog.floydhub.com/guide-to-hyperparameters-search-for-deep-learning-models/",
             category="Training environment", description="Environment link for pre-trained model"),
    Metadata(key="Image Sourcing", value="https://lionbridge.ai/services/image-data/", category="other",
             description="Source for image training data")
]

import json
from json import load
# Declare input/output schema for our model - this is optional
# It must be a valid json or yaml string
# Schema like model artifact is immutable hence it is allowed only at the model creation time and cannot be updated
# Schema json sample in appendix
input_schema = load(open('SR_input_schema.json','rb'))
input_schema_str= json.dumps(input_schema)
output_schema = load(open('SR_output_schema.json','rb'))
output_schema_str= json.dumps(output_schema)

import oci
 
config = oci.config.from_file()
data_science_client = oci.data_science.DataScienceClient(config=config)
# Change the timeout value to 1800 sec (30 mins)
data_science_client.base_client.timeout =  30 * 60

import zipfile
import os
     
def zipdir(target_zip_path, ziph, source_artifact_directory):
    ''' Creates a zip archive of a model artifact directory.
     
    Parameters:
     
    - target_zip_path: the path where you want to store the zip archive of your artifact
    - ziph: a zipfile.ZipFile object
    - source_artifact_directory: the path to the artifact directory.
 
    Returns a zip archive in the target_zip_path you specify.    
 
    '''
    for root, dirs, files in os.walk(source_artifact_directory):
        for file in files:
            ziph.write(os.path.join(root, file),
                       os.path.relpath(os.path.join(root,file),
                                       os.path.join(target_zip_path,'.')))
       
zipf = zipfile.ZipFile('<relpath-to-artifact-directory>.zip', 'w', zipfile.zip_DEFLATED)
zipdir('.', zipf, "<relpath-to-artifact-directory>")
zipf.close()

# creating a model details object:
model_details = CreateModelDetails(
    compartment_id='<compartment-ocid-of-model>',
    project_id='<project-ocid>',
    display_name='<display-name-of-model>',
    description='<description-of-model>',
    custom_metadata_list=custom_metadata_list,
    defined_metadata_list=defined_metadata_list,
    input_schema=input_schema_str,
    output_schema=output_schema_str)
 
# creating the model object:
model = data_science_client.create_model(model_details)
# adding the provenance:
data_science_client.create_model_provenance(model.data.id, provenance_details)
# adding the artifact:
with open('<relpath-to-artifact-directory>.zip','rb') as artifact_file:
    artifact_bytes = artifact_file.read()
    data_science_client.create_model_artifact(model.data.id, artifact_bytes, content_disposition='attachment; filename="<relpath-to-artifact-directory>.zip"')