Wallaroo SDK Essentials Guide: Model Uploads and Registrations

How to create and manage Wallaroo Models Uploads through the Wallaroo SDK

Table of Contents

Models are uploaded or registered to a Wallaroo workspace depending on the model framework and version.

Wallaroo Engine Runtimes

Pipeline deployment configurations provide two runtimes to run models in the Wallaroo engine:

  • Native Runtimes: Models that are deployed “as is” with the Wallaroo engine. These are:

    • ONNX
    • Python step
    • Tensorflow 2.9.1 in SavedModel format
  • Containerized Runtimes: Containerized models such as MLFlow or Arbitrary Python. These are run in the Wallaroo engine in their containerized form.

  • Non-Native Runtimes: Models that when uploaded are either converted to a native Wallaroo runtime, or are containerized so they can be run in the Wallaroo engine. When uploaded, Wallaroo will attempt to convert it to a native runtime. If it can not be converted, then it will be packed in a Wallaroo containerized model based on its framework type.

    Pipeline Deployment Configurations

Pipeline configurations are dependent on whether the model is converted to the Native Runtime space, or Containerized Model Runtime space.

This model will always run in the native runtime space.

Native Runtime Pipeline Deployment Configuration Example

The following configuration allocates 0.25 CPU and 1 Gi RAM to the native runtime models for a pipeline.

deployment_config = DeploymentConfigBuilder()
                    .cpus(0.25)
                    .memory('1Gi')
                    .build()

Models and Runtimes

Supported Models

The following frameworks are supported. Frameworks fall under either Native or Containerized runtimes in the Wallaroo engine. For more details, see the specific framework what runtime a specific model framework runs in.

Runtime DisplayModel Runtime SpacePipeline Configuration
tensorflowNativeNative Runtime Configuration Methods
onnxNativeNative Runtime Configuration Methods
pythonNativeNative Runtime Configuration Methods
mlflowContainerizedContainerized Runtime Deployment

Please note the following.

Wallaroo natively supports Open Neural Network Exchange (ONNX) models into the Wallaroo engine.

ParameterDescription
Web Sitehttps://onnx.ai/
Supported LibrariesSee table below.
FrameworkFramework.ONNX aka onnx
RuntimeNative aka onnx

The following ONNX versions models are supported:

Wallaroo VersionONNX VersionONNX IR VersionONNX OPset VersionONNX ML Opset Version
2023.2.1 (July 2023)1.12.18173
2023.2 (May 2023)1.12.18173
2023.1 (March 2023)1.12.18173
2022.4 (December 2022)1.12.18173
After April 2022 until release 2022.4 (December 2022)1.10.*7152
Before April 20221.6.*7132

For the most recent release of Wallaroo 2023.2.1, the following native runtimes are supported:

  • If converting another ML Model to ONNX (PyTorch, XGBoost, etc) using the onnxconverter-common library, the supported DEFAULT_OPSET_NUMBER is 17.

Using different versions or settings outside of these specifications may result in inference issues and other unexpected behavior.

ONNX models always run in the native runtime space.

Data Schemas

ONNX models deployed to Wallaroo have the following data requirements.

  • Equal rows constraint: The number of input rows and output rows must match.
  • All inputs are tensors: The inputs are tensor arrays with the same shape.
  • Data Type Consistency: Data types within each tensor are of the same type.

Equal Rows Constraint

Inference performed through ONNX models are assumed to be in batch format, where each input row corresponds to an output row. This is reflected in the in fields returned for an inference. In the following example, each input row for an inference is related directly to the inference output.

df = pd.read_json('./data/cc_data_1k.df.json')
display(df.head())

result = ccfraud_pipeline.infer(df.head())
display(result)

INPUT

 tensor
0[-1.0603297501, 2.3544967095000002, -3.5638788326, 5.1387348926, -1.2308457019, -0.7687824608, -3.5881228109, 1.8880837663, -3.2789674274, -3.9563254554, 4.0993439118, -5.6539176395, -0.8775733373, -9.131571192000001, -0.6093537873, -3.7480276773, -5.0309125017, -0.8748149526000001, 1.9870535692, 0.7005485718000001, 0.9204422758, -0.1041491809, 0.3229564351, -0.7418141657, 0.0384120159, 1.0993439146, 1.2603409756, -0.1466244739, -1.4463212439]
1[-1.0603297501, 2.3544967095000002, -3.5638788326, 5.1387348926, -1.2308457019, -0.7687824608, -3.5881228109, 1.8880837663, -3.2789674274, -3.9563254554, 4.0993439118, -5.6539176395, -0.8775733373, -9.131571192000001, -0.6093537873, -3.7480276773, -5.0309125017, -0.8748149526000001, 1.9870535692, 0.7005485718000001, 0.9204422758, -0.1041491809, 0.3229564351, -0.7418141657, 0.0384120159, 1.0993439146, 1.2603409756, -0.1466244739, -1.4463212439]
2[-1.0603297501, 2.3544967095000002, -3.5638788326, 5.1387348926, -1.2308457019, -0.7687824608, -3.5881228109, 1.8880837663, -3.2789674274, -3.9563254554, 4.0993439118, -5.6539176395, -0.8775733373, -9.131571192000001, -0.6093537873, -3.7480276773, -5.0309125017, -0.8748149526000001, 1.9870535692, 0.7005485718000001, 0.9204422758, -0.1041491809, 0.3229564351, -0.7418141657, 0.0384120159, 1.0993439146, 1.2603409756, -0.1466244739, -1.4463212439]
3[-1.0603297501, 2.3544967095000002, -3.5638788326, 5.1387348926, -1.2308457019, -0.7687824608, -3.5881228109, 1.8880837663, -3.2789674274, -3.9563254554, 4.0993439118, -5.6539176395, -0.8775733373, -9.131571192000001, -0.6093537873, -3.7480276773, -5.0309125017, -0.8748149526000001, 1.9870535692, 0.7005485718000001, 0.9204422758, -0.1041491809, 0.3229564351, -0.7418141657, 0.0384120159, 1.0993439146, 1.2603409756, -0.1466244739, -1.4463212439]
4[0.5817662108, 0.09788155100000001, 0.1546819424, 0.4754101949, -0.19788623060000002, -0.45043448540000003, 0.016654044700000002, -0.0256070551, 0.0920561602, -0.2783917153, 0.059329944100000004, -0.0196585416, -0.4225083157, -0.12175388770000001, 1.5473094894000001, 0.2391622864, 0.3553974881, -0.7685165301, -0.7000849355000001, -0.1190043285, -0.3450517133, -1.1065114108, 0.2523411195, 0.0209441826, 0.2199267436, 0.2540689265, -0.0450225094, 0.10867738980000001, 0.2547179311]

OUTPUT

 timein.tensorout.dense_1check_failures
02023-11-17 20:34:17.005[-1.0603297501, 2.3544967095, -3.5638788326, 5.1387348926, -1.2308457019, -0.7687824608, -3.5881228109, 1.8880837663, -3.2789674274, -3.9563254554, 4.0993439118, -5.6539176395, -0.8775733373, -9.131571192, -0.6093537873, -3.7480276773, -5.0309125017, -0.8748149526, 1.9870535692, 0.7005485718, 0.9204422758, -0.1041491809, 0.3229564351, -0.7418141657, 0.0384120159, 1.0993439146, 1.2603409756, -0.1466244739, -1.4463212439][0.99300325]0
12023-11-17 20:34:17.005[-1.0603297501, 2.3544967095, -3.5638788326, 5.1387348926, -1.2308457019, -0.7687824608, -3.5881228109, 1.8880837663, -3.2789674274, -3.9563254554, 4.0993439118, -5.6539176395, -0.8775733373, -9.131571192, -0.6093537873, -3.7480276773, -5.0309125017, -0.8748149526, 1.9870535692, 0.7005485718, 0.9204422758, -0.1041491809, 0.3229564351, -0.7418141657, 0.0384120159, 1.0993439146, 1.2603409756, -0.1466244739, -1.4463212439][0.99300325]0
22023-11-17 20:34:17.005[-1.0603297501, 2.3544967095, -3.5638788326, 5.1387348926, -1.2308457019, -0.7687824608, -3.5881228109, 1.8880837663, -3.2789674274, -3.9563254554, 4.0993439118, -5.6539176395, -0.8775733373, -9.131571192, -0.6093537873, -3.7480276773, -5.0309125017, -0.8748149526, 1.9870535692, 0.7005485718, 0.9204422758, -0.1041491809, 0.3229564351, -0.7418141657, 0.0384120159, 1.0993439146, 1.2603409756, -0.1466244739, -1.4463212439][0.99300325]0
32023-11-17 20:34:17.005[-1.0603297501, 2.3544967095, -3.5638788326, 5.1387348926, -1.2308457019, -0.7687824608, -3.5881228109, 1.8880837663, -3.2789674274, -3.9563254554, 4.0993439118, -5.6539176395, -0.8775733373, -9.131571192, -0.6093537873, -3.7480276773, -5.0309125017, -0.8748149526, 1.9870535692, 0.7005485718, 0.9204422758, -0.1041491809, 0.3229564351, -0.7418141657, 0.0384120159, 1.0993439146, 1.2603409756, -0.1466244739, -1.4463212439][0.99300325]0
42023-11-17 20:34:17.005[0.5817662108, 0.097881551, 0.1546819424, 0.4754101949, -0.1978862306, -0.4504344854, 0.0166540447, -0.0256070551, 0.0920561602, -0.2783917153, 0.0593299441, -0.0196585416, -0.4225083157, -0.1217538877, 1.5473094894, 0.2391622864, 0.3553974881, -0.7685165301, -0.7000849355, -0.1190043285, -0.3450517133, -1.1065114108, 0.2523411195, 0.0209441826, 0.2199267436, 0.2540689265, -0.0450225094, 0.1086773898, 0.2547179311][0.0010916889]0

All Inputs Are Tensors

All inputs into an ONNX model must be tensors. This requires that the shape of each element is the same. For example, the following is a proper input:

t [
    [2.35, 5.75],
    [3.72, 8.55],
    [5.55, 97.2]
]
Standard tensor array

Another example is a 2,2,3 tensor, where the shape of each element is (3,), and each element has 2 rows.

t = [
        [2.35, 5.75, 19.2],
        [3.72, 8.55, 10.5]
    ],
    [
        [5.55, 7.2, 15.7],
        [9.6, 8.2, 2.3]
    ]

In this example each element has a shape of (2,). Tensors with elements of different shapes, known as ragged tensors, are not supported. For example:

t = [
    [2.35, 5.75],
    [3.72, 8.55, 10.5],
    [5.55, 97.2]
])

**INVALID SHAPE**
Ragged tensor array - unsupported

For models that require ragged tensor or other shapes, see other data formatting options such as Bring Your Own Predict models.

Data Type Consistency

All inputs into an ONNX model must have the same internal data type. For example, the following is valid because all of the data types within each element are float32.

t = [
    [2.35, 5.75],
    [3.72, 8.55],
    [5.55, 97.2]
]

The following is invalid, as it mixes floats and strings in each element:

t = [
    [2.35, "Bob"],
    [3.72, "Nancy"],
    [5.55, "Wani"]
]

The following inputs are valid, as each data type is consistent within the elements.

df = pd.DataFrame({
    "t": [
        [2.35, 5.75, 19.2],
        [5.55, 7.2, 15.7],
    ],
    "s": [
        ["Bob", "Nancy", "Wani"],
        ["Jason", "Rita", "Phoebe"]
    ]
})
df
 ts
0[2.35, 5.75, 19.2][Bob, Nancy, Wani]
1[5.55, 7.2, 15.7][Jason, Rita, Phoebe]
ParameterDescription
Web Sitehttps://www.tensorflow.org/
Supported Librariestensorflow==2.9.1
FrameworkFramework.TENSORFLOW aka tensorflow
RuntimeNative aka tensorflow
Supported File TypesSavedModel format as .zip file

TensorFlow File Format

TensorFlow models are .zip file of the SavedModel format. For example, the Aloha sample TensorFlow model is stored in the directory alohacnnlstm:

├── saved_model.pb
└── variables
    ├── variables.data-00000-of-00002
    ├── variables.data-00001-of-00002
    └── variables.index

This is compressed into the .zip file alohacnnlstm.zip with the following command:

zip -r alohacnnlstm.zip alohacnnlstm/

ML models that meet the Tensorflow and SavedModel format will run as Wallaroo Native runtimes by default.

See the SavedModel guide for full details.

ParameterDescription
Web Sitehttps://www.python.org/
Supported Librariespython==3.8
FrameworkFramework.PYTHON aka python
RuntimeNative aka python

Python models uploaded to Wallaroo are executed as a native runtime.

Note that Python models - aka “Python steps” - are standalone python scripts that use the python libraries natively supported by the Wallaroo platform. These are used for either simple model deployment (such as ARIMA Statsmodels), or data formatting such as the postprocessing steps. A Wallaroo Python model will be composed of one Python script that matches the Wallaroo requirements.

This is contrasted with Arbitrary Python models, also known as Bring Your Own Predict (BYOP) allow for custom model deployments with supporting scripts and artifacts. These are used with pre-trained models (PyTorch, Tensorflow, etc) along with whatever supporting artifacts they require. Supporting artifacts can include other Python modules, model files, etc. These are zipped with all scripts, artifacts, and a requirements.txt file that indicates what other Python models need to be imported that are outside of the typical Wallaroo platform.

Python Models Requirements

Python models uploaded to Wallaroo are Python scripts that must include the wallaroo_json method as the entry point for the Wallaroo engine to use it as a Pipeline step.

This method receives the results of the previous Pipeline step, and its return value will be used in the next Pipeline step.

If the Python model is the first step in the pipeline, then it will be receiving the inference request data (for example: a preprocessing step). If it is the last step in the pipeline, then it will be the data returned from the inference request.

In the example below, the Python model is used as a post processing step for another ML model. The Python model expects to receive data from a ML Model who’s output is a DataFrame with the column dense_2. It then extracts the values of that column as a list, selects the first element, and returns a DataFrame with that element as the value of the column output.

def wallaroo_json(data: pd.DataFrame):
    print(data)
    return [{"output": [data["dense_2"].to_list()[0][0]]}]

In line with other Wallaroo inference results, the outputs of a Python step that returns a pandas DataFrame or Arrow Table will be listed in the out. metadata, with all inference outputs listed as out.{variable 1}, out.{variable 2}, etc. In the example above, this results the output field as the out.output field in the Wallaroo inference result.

 timein.tensorout.outputcheck_failures
02023-06-20 20:23:28.395[0.6878518042, 0.1760734021, -0.869514083, 0.3..[12.886651039123535]0
ParameterDescription
Web Sitehttps://huggingface.co/models
Supported Libraries
  • transformers==4.27.0
  • diffusers==0.14.0
  • accelerate==0.18.0
  • torchvision==0.14.1
  • torch==1.13.1
FrameworksThe following Hugging Face pipelines are supported by Wallaroo.
  • Framework.HUGGING_FACE_FEATURE_EXTRACTION aka hugging-face-feature-extraction
  • Framework.HUGGING_FACE_IMAGE_CLASSIFICATION aka hugging-face-image-classification
  • Framework.HUGGING_FACE_IMAGE_SEGMENTATION aka hugging-face-image-segmentation
  • Framework.HUGGING_FACE_IMAGE_TO_TEXT aka hugging-face-image-to-text
  • Framework.HUGGING_FACE_OBJECT_DETECTION aka hugging-face-object-detection
  • Framework.HUGGING_FACE_QUESTION_ANSWERING aka hugging-face-question-answering
  • Framework.HUGGING_FACE_STABLE_DIFFUSION_TEXT_2_IMG aka hugging-face-stable-diffusion-text-2-img
  • Framework.HUGGING_FACE_SUMMARIZATION aka hugging-face-summarization
  • Framework.HUGGING_FACE_TEXT_CLASSIFICATION aka hugging-face-text-classification
  • Framework.HUGGING_FACE_TRANSLATION aka hugging-face-translation
  • Framework.HUGGING_FACE_ZERO_SHOT_CLASSIFICATION aka hugging-face-zero-shot-classification
  • Framework.HUGGING_FACE_ZERO_SHOT_IMAGE_CLASSIFICATION aka hugging-face-zero-shot-image-classification
  • Framework.HUGGING_FACE_ZERO_SHOT_OBJECT_DETECTION aka hugging-face-zero-shot-object-detection
  • Framework.HUGGING_FACE_SENTIMENT_ANALYSIS aka hugging-face-sentiment-analysis
  • Framework.HUGGING_FACE_TEXT_GENERATION aka hugging-face-text-generation
RuntimeContainerized aka tensorflow / mlflow

Hugging Face Schemas

Input and output schemas for each Hugging Face pipeline are defined below. Note that adding additional inputs not specified below will raise errors, except for the following:

  • Framework.HUGGING-FACE-IMAGE-TO-TEXT
  • Framework.HUGGING-FACE-TEXT-CLASSIFICATION
  • Framework.HUGGING-FACE-SUMMARIZATION
  • Framework.HUGGING-FACE-TRANSLATION

Additional inputs added to these Hugging Face pipelines will be added as key/pair value arguments to the model’s generate method. If the argument is not required, then the model will default to the values coded in the original Hugging Face model’s source code.

See the Hugging Face Pipeline documentation for more details on each pipeline and framework.

Wallaroo FrameworkReference
Framework.HUGGING-FACE-FEATURE-EXTRACTION

Schemas:

input_schema = pa.schema([
    pa.field('inputs', pa.string())
])
output_schema = pa.schema([
    pa.field('output', pa.list_(
        pa.list_(
            pa.float64(),
            list_size=128
        ),
    ))
])
Wallaroo FrameworkReference
Framework.HUGGING-FACE-IMAGE-CLASSIFICATION

Schemas:

input_schema = pa.schema([
    pa.field('inputs', pa.list_(
        pa.list_(
            pa.list_(
                pa.int64(),
                list_size=3
            ),
            list_size=100
        ),
        list_size=100
    )),
    pa.field('top_k', pa.int64()),
])

output_schema = pa.schema([
    pa.field('score', pa.list_(pa.float64(), list_size=2)),
    pa.field('label', pa.list_(pa.string(), list_size=2)),
])
Wallaroo FrameworkReference
Framework.HUGGING-FACE-IMAGE-SEGMENTATION

Schemas:

input_schema = pa.schema([
    pa.field('inputs', 
        pa.list_(
            pa.list_(
                pa.list_(
                    pa.int64(),
                    list_size=3
                ),
                list_size=100
            ),
        list_size=100
    )),
    pa.field('threshold', pa.float64()),
    pa.field('mask_threshold', pa.float64()),
    pa.field('overlap_mask_area_threshold', pa.float64()),
])

output_schema = pa.schema([
    pa.field('score', pa.list_(pa.float64())),
    pa.field('label', pa.list_(pa.string())),
    pa.field('mask', 
        pa.list_(
            pa.list_(
                pa.list_(
                    pa.int64(),
                    list_size=100
                ),
                list_size=100
            ),
    )),
])
Wallaroo FrameworkReference
Framework.HUGGING-FACE-IMAGE-TO-TEXT

Any parameter that is not part of the required inputs list will be forwarded to the model as a key/pair value to the underlying models generate method. If the additional input is not supported by the model, an error will be returned.

Schemas:

input_schema = pa.schema([
    pa.field('inputs', pa.list_(