Skip to content

Microservice_Flower

Concept Key Subkey Values
Description
id "MSID_FLOWER"
name "Object Detection & Classification Network"
author "DFKI"
date
version "1.0.0"
description "Detection and classification of objects from Images"
classificationSchema "ML"
type ["neural network", "deep learning", "convolutional neural network", "CNN"]
Service
deploymentFormat "docker-compose"
deploymentData
configurationData
mountedSharedDirectories
Hardware Requirements
recommendedNumberOfGPUs
recommendedGPURAM
gpuType
hpcRequired
edgeType
recommendedRAM 2
recommendedCPUs 2
requiredDiskSpace 42GB
OS Requirements
osArch "x86_64"
osType "linux"
Data Resources
dataResource
Model
model_types SavedModel (Tensorflow)
model_recommendedAuthTools Tensorflow
Parameters
parameters
Metrics
metrics [, ]

Algorithm

Concept Key Subkey Values
Description
id "ALGID_FLOWER"
name "DFKI Object Classification Algorithm"
description "Algorithm for ML inferences based on tensorflow"
classificationSchema "ML"
type ["neural network", "deep learning", "convolutional neural network", "CNN"]
author "Valerie Poser"
date
version "1.0.0"
Algorithm
listOfMicroservices ["MSID_FLOWER"]
deploymentMapping

Model

Concept Key Subkey Values
Description
ID "MODELID_FLOWER"
Name Flower Classification
Version "1.0.0"
License Public Domain
Provider DFKI
Provider_contacts
AuthTool
Type Tensorflow
Fidelity SavedModel
REPOSITORY_URI "s3://dbrain-s3-test"
PATH "dfki/flower/model"
FILENAME "dfki_ml_flower-inference.zip"
State_depend
Parameters
In-slots
Outputs [detectedClass, confidenceScore]
CosimSolverInfo
Dependant FMUs
Dependencies
OS Requirements
osArch
osType
osDistribution
Hardware Requirements
recommendedNumberOfGPUCores
minimumNumberOfGPUCores
recommendedGPURAM
minimumGPURAM
recommendedRAM
minimumRAM
recommendedCPUs
minimumCPUs
requiredDiskSpace
Publishing
Description "Model is based on a general imagenet model (https://tfhub.dev/google/imagenet/mobilenet_v3_small_100_224/classification/5) and has been fine-tuned on flower dataset, to classifiy different flower types."

Data

Concept Key Subkey Values
Administrative data
ID "DATAID_FLOWER"
AUTHOR
PROVIDER
DATE
Description
DATA_NAME "DFKIFlowerStream"
DATA_VERSION "1.0.0"
DATA_DESC This stream acts as a communication channel between an end user and an ML inference Algorithm
DATA_DESC_URL
DATA_SIZE 1m
Search support
DATA_TAGS ["jpg", "rgb"]
Datakind specification
DATA_KIND "STREAM"
DATA_DIRECTION "BIDIRECTIONAL"
DATA_FORMAT ["text/plain", "application/octet-stream"]
DATA_STORE_TYPE "KAFKA"
Data access specification
DATA_URI "kafka:193.225.250.14:9093:dfki-flower-in"
DATA_AUTH ["tls_manual"]
Open parameters
DATA_CREDENTIALS
Further access clauses (extensible)
DATA_AUX_INFO
Data content semantics
DATA_SCHEMA
DATA_SCHEMA_URL

MA Pair

Concept Key Subkey Values
Definition
ID "MAID_FLOWER"
Created at
Version 1.0.0
Author
Licensor
Derivation
Name DFKI Flower Classification
Scope Classification of flowers in images
IP Family
Namespace
M Asset "MODELID_FLOWER"
A Asset "ALGID_FLOWER"
Rules
Frequency
Payload

DMA Tuple

Concept Key Subkey Values
Definition
ID "DMAID_FLOWER"
Created at
Version 1.0.0
Author
Licensor
Derivation
Name Flower Classification for Valerie
Scope This DMA tuple contains all assets forming a complete object detection and classification (machine learning) stack, including in- and output streams, a trained model, as well as the inference algorithm.
IP Instance 123e4567-e89b-12d3-a456-426614174000
Namespace
MA Pair "MAID_FLOWER"
Schedule
Payload
Data Assets Mapping
DataAssetsMapping
Deployments
Deployments

Deployment

Concept Key Subkey Values
Description
id "HOSTID_FLOWER"
name "DFKI object detection infrastructure"
author "Valerie Poser"
date
Deployments
type "cloudbroker"
cloudbroker
deployment_id
instance_type_id
key_pair_id
opened_port
endpoint
cloud_config
edge
endpoint

Supporting Metadata

Concept Key Subkey Values
Person
ID
Created at
Version
First name
Last name
Email
CV
Image
Affiliation
Legal Entity (Organization)
ID
Created at
Version
Author
Name
Description
Logo
URL
Location
Phone
VAT ID No.
IP Family
ID
Created at
Version
Author
Manager
Name
Image
Model No.
Description
Capacity
IP Man.
IP Instance
ID
Created at
Version
Author
Operator
IP Family
Name
Image
Serial No.
Description
Location

DataAssetsMapping

Concept Key Subkey Values
DataAssetsMapping
MICROSERVICE_ID "MSID_FLOWER"
DATA_SOURCE_ID "DATAID_FLOWER"

ConfigurationData

Concept Key Subkey Values
ConfigurationData
filePath /data/rclone.conf
fileContent [s3-server]\n access_key: 123abc
mountPropagation Bidirectional

Data Resources

Concept Key Subkey Values
Data Resources
DATA_RESOURCE_ID kafka_stream
DATA_KIND ["STREAM"]
DATA_DIRECTION ["BIDIRECTIONAL"]
DATA_FORMAT ["image/jpeg", "text/plain"]
DATA_SOURCE_TYPE ["KAFKA"]
DATA_AUTH_TYPE ["tls_manual"]
DATA_SCHEMA
DATA_AUX_INFO

Parameters

Concept Key Subkey Values
Parameters
name "detection_threshold"
type "Integer"
mandatory True
defaultValue 42
description "This parameter can be used to configure the included algorithm in a certain way, possible values: 'A', 'B', 'Z'."

Metrics

Concept Key Subkey Values
Metrics
name "detectedClass"
correspondingMeasurement "detectedClass"
function "ML inference"
unit "enum"
description "which classes has been detected in the input image"

In-slots

Concept Key Subkey Values
In-slots
Key
Name
Dimensions
Is-continuous
Units
Unit
Exponent
Offset
Scale
Default-value
Ranges

Outputs

Concept Key Subkey Values
Outputs
Key
Name
Dimensions
Is-continuous
Units
Unit
Exponent
Offset
Scale
Default-value
Ranges