AutoML Services

Clarifai enables developers, data scientists, and machine learning engineers as well as non-technical users, to automatically build scalable machine learning models.

automl-hero

The easiest deep learning AI ecosystem for developers, data scientists, and no-code operators

Train high-quality custom machine learning models with minimal effort and machine learning expertise — build and train your own models in minutes.

 

 

Fastest End-to-End

The world’s fastest path to actionable insights and value from your unstructured image, video, text, and audio data.

Deploy Everywhere

Clarifai offers powerful APIs and clients for every programming language that can be deployed anywhere you operate from any public cloud, private cloud, bare metal, air-gapped, and edge deployments.

Ready-to-Use, Pre-Trained AI Models

Our extensive library of pre-trained models address the most important use cases across industries. They're available out-of-the-box and require no specialized knowledge in data science. 

Learn more

Low Code / No Code Solutions

Use powerful tools and an easy-to-use UI to quickly build AI apps that require no development skills. Improve project turnaround time to solve business problems quickly.

Spend less time on low level details and more time on using ML to improve business outcomes

Use Clarifai's AutoML tools to take care of sourcing and preparing data, engineering features, training and tuning models, deploying models, and ongoing model monitoring and updating.

Data Preparation

Use machine learning to prepare datasets to train models

With Clarifai Scribe Label, you can use AI to automate the process of labeling data for multiple labeling tasks. Data augmentation algorithms are incorporated into how data is ingested. With Clarifai's Mesh Workflows, automatically sort, filter, and annotate data. And with Clarifai Spacetime Search, use AI-powered search with built-in class, concept, and visual similarity search capabilities. With Scribe LabelForce easily scale your data labeling with Clarifai's fully managed data labeling services for creating high-quality training datasets.

automl-data-prep
Feature Engineering

Comprehensive tool set for transforming unstructured data into features to improve model accuracy

Use data preparation workflows to automatically annotate inputs based on automatic accuracy criteria, reducing the time it takes to annotate new data. Combine pre-trained and custom-trained embedding, clustering, concept, and detect region models to index inputs for search to manage unstructured data at scale. Using Knowledge Graph's link model concepts to other people's concepts within Clarifai's AI Lake. With Clarifai Armada automate the detection and classification of features. Help users generate new features from existing ones with our complete set of APIs and UIs. With Scribe Label, ML-powered data labeling tool, use pre-trained, custom-trained, or 3rd party models to auto annotate data for scale.

automl-feature-engineering
Automated Model Building

Automate custom model building for unstructured datasets

With Clarifai Enlight customize modular model architectures to accelerate continuous model improvements. Automate data analysis and recommend or select the potential algorithms to optimize model performance. Use model training templates to automatically optimize hyperparameters. Easily evaluate model performance to facilitate model interpretability. Use quick training and transfer learning  to train models by using the understanding of other pre-trained model embeddings.

automl-model-building
Model Management

Automatically evaluate models for model prediction performance

Clarifai Enlight's robust set of tools can easily evaluate model performance to facilitate model interoperability. Customize modular model architectures to accelerate continuous model improvements. With Enlight Versioning leverage model, workflow, and dataset versioning capabilities to provide automatic snapshots every time models are trained. With Enlight Evaluations easily identify high-performing and low-performing concepts. Use Clarifai AI Lake, a repository of AI resources within one ecosystem, to easily share AI Apps across organizations.

automl-model-managment
Bias Detection

Powerful tools to identify bias

Use Clarifai's pre-trained models to identify bias and unbalanced samples due to oversampling, the inclusion of data that is prohibited due to regulatory constraints, and other sources in model output. Clarifai's Demographic models and workflows automatically identify age, gender, and multi-cultural appearance to help you discover class imbalances.

automl-bias-detection
Model Interpretability

Gauge model performance at varying levels of granularity

Clarifai Enlight Model Training Services allow you to request an evaluation by specifying a dataset with trusted annotations to compare model predictions against. View evaluation results within in-browser dashboards. Easily identify potential problem areas that to improve subsequent retraining efforts.

automl-model-intrepretability
Model Explainability

Understand model characteristics before deployment to debug predictions

Clarifai Enlight Evaluation Services makes it easy to train various task-specific model-types each with a set of preset training templates. Use inline tooltips and default values for customizable training parameters to configure model training for optimal performance.

automl-model-explainability

Start building high-performance models today

Get started with 1,000 free operations each month. Request a free API key and start building today.

response = stub.PostModelOutputs(
    service_pb2.PostModelOutputsRequest(
        model_id="{THE_MODEL_ID}",
        inputs=[
            resources_pb2.Input(
                data=resources_pb2.Data(
                    image=resources_pb2.Image(
                        url="https://samples.clarifai.com/metro-north.jpg"
                    )
                )
            )
        ]
    ),
    metadata=metadata
)
print("Predicted concepts:")
for concept in response.outputs[0].data.concepts:
    print(concept.name + " " + str(concept.value))
curl -X POST
    -H 'Authorization: Key YOUR_API_KEY'
    -H "Content-Type: application/json"
    -d '
    {
      "inputs": [
        {
          "data": {
            "image": {
              "url": "https://samples.clarifai.com/metro-north.jpg"
            }
          }
        }
      ]
    }'
    https://api.clarifai.com/v2/models/e0be3b9d6a454f0493ac3a30784001ff/outputs

 

MultiOutputResponse response = stub.postModelOutputs(
PostModelOutputsRequest.newBuilder()
.setModelId("aaa03c23b3724a16a56b629203edc62c")
.addInputs(
Input.newBuilder().setData(
Data.newBuilder().setImage(
Image.newBuilder().setUrl("YOUR_IMAGE_URL")
)
)
)
.build()
);
const request = new service.PostModelOutputsRequest();
request.setModelId("aaa03c23b3724a16a56b629203edc62c");
request.addInputs(
    new resources.Input()
        .setData(
            new resources.Data()
                .setImage(
                    new resources.Image()
                        .setUrl("https://samples.clarifai.com/dog2.jpeg")
                )
        )
)
[$response, $status] = $client->PostModelOutputs(
new PostModelOutputsRequest([
'model_id' => 'aaa03c23b3724a16a56b629203edc62c',
'inputs' => [
new Input([
'data' => new Data([
'image' => new Image([
'url' => 'https://samples.clarifai.com/dog2.jpeg'
])
])
])
]
]),
$metadata
)->wait();
var GeneralModelId = "aaa03c23b3724a16a56b629203edc62c"
response, err := client.PostModelOutputs(
ctx,
&api.PostModelOutputsRequest{
ModelId: GeneralModelId,
Inputs: []*api.Input{
{
Data: &api.Data{
Image: &api.Image{
Url: "https://samples.clarifai.com/dog2.jpeg",
},
},
},
},
},
)
const GENERAL_MODEL_ID: &str = "aaa03c23b3724a16a56b629203edc62c";

let request = service::PostModelOutputsRequest {
model_id: GENERAL_MODEL_ID.to_string(),
inputs: RepeatedField::from(vec![resources::Input {
data: SingularPtrField::some(resources::Data {
image: SingularPtrField::some(resources::Image {
url: "https://samples.clarifai.com/dog2.jpeg".to_string(),
..Default::default()
}),
..Default::default()
}),
..Default::default()
}]),
..Default::default()
};
string GENERAL_MODEL_ID = "aaa03c23b3724a16a56b629203edc62c";

PostModelOutputsRequest request;
request.set_model_id(GENERAL_MODEL_ID);

Input* input = request.add_inputs();
Data* data = input->mutable_data();
Image* image = data->mutable_image();
image->set_url("https://samples.clarifai.com/dog2.jpeg");

MultiOutputResponse response;
grpc::Status status = stub->PostModelOutputs(context.get(), request, &response);

if (!status.ok()) {
cout << "Failure: " << status.error_code() << " " << status.error_message() << endl;
exit(1);
}

if (response.status().code() != status::StatusCode::SUCCESS) {
cout << "Error response: " << response.status().code() << " " << response.status().description() << " " << response.status().details() << endl;
exit(1);
}

Data response_data = response.outputs(0).data();
cout << "Predicted concepts:" << endl;
for (int i = 0; i < response_data.concepts_size(); i++) {
const Concept& c = response_data.concepts(i);
cout << "\t" << c.name() << ": " << c.value() << endl;
}
var response = client.PostModelOutputs(
new PostModelOutputsRequest()
{
ModelId = "aaa03c23b3724a16a56b629203edc62c",
Inputs =
{
new List<Input>()
{
new Input()
{
Data = new Data()
{
Image = new Image()
{
Url = "https://samples.clarifai.com/dog2.jpeg"
}
}
}
}
}
},
metadata
);
let response = try client.postModelOutputs(
Clarifai_Api_PostModelOutputsRequest.with {
$0.modelID = "aaa03c23b3724a16a56b629203edc62c";
$0.inputs = [
Clarifai_Api_Input.with {
$0.data = Clarifai_Api_Data.with {
$0.image = Clarifai_Api_Image.with {
$0.url = "https://samples.clarifai.com/dog2.jpeg"
}
}
}
]
}
).response.wait()

Contact us to get started.

Talk to an AI expert to find the right solution for your use case.

If you are a customer and need support, we'll be happy to help. Contact support

×