Overview

Models

aiXplain has an ever-expanding catalog of ready-to-use AI models by different suppliers (e.g. AWS, Microsoft, Google, Meta, etc.) for various functions (e.g. Machine Translation, Speech Recognition, Large Language Modeling, Sentiment Analysis, etc.). These models are available on-demand, can be connected together into Pipelines and Agents.

Curation

aiXplain's models are categorized using several filters, such as Function, Supplier and Modalities (e.g. source language, target language) to make searching easier.

Standardization

All aiXplain models can be run using the same syntax via the SDK, and our standardizations allow for model swapping in your pipelines.

Onboarding

You can onboard your own models onto aiXplain, making them accessible for deployment and utilization within your applications.

How to search the marketplace

This will walk you through the process of searching the marketplace for various aiXplain assets including models, metrics, and data.

Models

Models

aiXplain's collection of models are searchable using queries, filters and are directly accessible using thier IDs (unique identifiers).

Models are searchable on the SDK using the following search parameters:

Python

Parameter	Type	Default	Description
query	Optional[Text]	""	Search query to filter models based on their name or description.
function	Optional[Function]	None	AI function filter (e.g., Translation, Text Generation).
suppliers	Optional[Union[Supplier, List[Supplier]]]	None	Filter models by suppliers (e.g., AWS, Google, Microsoft).
source_languages	Optional[Union[Language, List[Language]]]	None	Filter models based on their input language(s).
target_languages	Optional[Union[Language, List[Language]]]	None	Filter models based on their output language(s).
is_finetunable	Optional[bool]	None	Specify if models should support fine-tuning.
ownership	Optional[Tuple[OwnershipType, List[OwnershipType]]]	None	Filter models by ownership type (e.g., SUBSCRIBED, OWNER).
sort_by	Optional[SortBy]	None	Attribute to sort the retrieved models (e.g., name, creation date).
sort_order	SortOrder	SortOrder.ASCENDING	Specify the sorting order (ascending or descending).
page_number	int	0	Page number for paginated results.
page_size	int	20	Number of results to retrieve per page.

Swift

Parameter	Type	Default	Description
query	Optional[Text]	""	Search query to filter models based on their name or description.
function	Optional[Function]	None	AI function filter (e.g., Translation, Text Generation).

Python

Let's use query, function, source_languages, target_languages and suppliers to search for translation models from English to Canadian French.

from aixplain.factories import ModelFactory
from aixplain.enums import Function, Language, Supplier

model_list = ModelFactory.list(
    "Canada",
    function=Function.TRANSLATION,
    source_languages=Language.English,
    target_languages=Language.French,
    suppliers=[Supplier.AWS, Supplier.GOOGLE, Supplier.MICROSOFT],
    )["results"]

for model in model_list:
    print(model.__dict__)

Show output

tip

Use the _member_names_ attribute to see the list of available function types, languages and suppliers.

Function._member_names_

Show output

Language._member_names_

Show output

Supplier._member_names_

Show output

Swift

Let's use query and function to search for translation models from English to Canadian French.

let provider = ModelProvider()
Task {
    let query = ModelQuery(query: "English to French Canada", functions: ["translation"])
    let result = try? await provider.list(query)

    result?.forEach { model in
        dump(model)
    }
}

Show output

tip

The best way to view a list of all available functions is to check Discover. We will add a programmatic way to get this information to Swift in future. We will also soon display all models (and functions) in the Docs. 👍

Direct Access

Once you know a model's ID, you can access the model directly (without searching for it).

EXAMPLE OpenAI's GPT-4 model has ID 6414bd3cd09663e9225130e8.

Instantiate a model object

Python

from aixplain.factories import ModelFactory
model = ModelFactory.get('6414bd3cd09663e9225130e8')
model.__dict__

Show output

Swift

import AiXplainKit
let model = ModelProvider().get("6414bd3cd09663e9225130e8")
dump(model)

Show output

Once you have identified a suitable model, you can use it for inference, benchmarking, finetuning, or integrating it into custom AI pipelines.

Metrics

Metrics

Metrics are searchable on the SDK using the following search parameters:

• hypothesis (Hypothesis): The metric's hypothesis modality. (Text, Audio, Image, Video, Label, Number.)
• is_source_required (Source required): Should the metric use source. Defaults to None.
• is_reference_required (Reference required):Should the metric use reference. Defaults to None.
• model_id: ID of model for which metric is to be used. Defaults to None.

All parameters except query can be passed as single items or lists.

info

View all available metrics by specifying no filters.

Python

from aixplain.factories import MetricFactory

metric_list = MetricFactory.list(page_size=50)["results"]

for metric in metric_list:
    print(metric.__dict__)

Show output

Let's search for all metrics compatible with GPT-4 that require a reference input.

Python

metric_list = MetricFactory.list(
model_id=model.id,
is_reference_required=True,
)["results"]

for metric in metric_list:
    print(metric.__dict__)

Show output

Direct Access

Once you know a metric's ID, you can access the metric directly (without searching for it).

EXAMPLE
BLEU has ID 639874ab506c987b1ae1acc6.

Instantiate a metric object

Python

from aixplain.factories import MetricFactory
bleu = MetricFactory.get('639874ab506c987b1ae1acc6')
bleu.__dict__

Show output

Once you have found your metric, you can use it for inference, finetuning and benchmarking.

Data

Data

Data assets are searchable using queries and filters and directly accessible using their IDs (unique identifiers).

note

You cannot download the contents of a Dataset or Corpus.

Corpora

Python

from aixplain.factories import CorpusFactory

corpus_list = CorpusFactory.list(
  # query="",
  # function=Function.TRANSLATION,
  # language=[Language.English],
  # data_type=DataType.TEXT,
  # page_size=5
)["results"]

for corpus in corpus_list:
  print(corpus.__dict__)

corpus = CorpusFactory.get("<CORPUS_ID>")

Datasets

Python

from aixplain.factories import DatasetFactory

dataset_list = DatasetFactory.list(
  # query="",
  # function=Function.TRANSLATION,
  # source_languages=[Language.English],
  # target_languages=[Language.Spanish],
  # data_type=DataType.TEXT,
  # page_size=5
)["results"]

for dataset in dataset_list:
  print(dataset.__dict__)

dataset = DatasetFactory.get("<DATASET_ID>")

Next, create or integrate these assets into AI agents to enhance your applications.

How to call an asset

The aiXplain SDK allows you to run models synchronously (Python) or asynchronously (Python and Swift). You can also process Data assets (Python).

Let's use Groq's Llama 70B as an example.

Python

from aixplain.factories import ModelFactory
from aixplain.enums import Supplier

model_list = ModelFactory.list(suppliers=Supplier.GROQ)["results"]
for model in model_list:
    print(model.__dict__)

Show output

model = ModelFactory.get("6626a3a8c8f1d089790cf5a2")

Swift

let provider = ModelProvider()
Task {
    let query = ModelQuery(query: "groq")
    let result = try? await provider.list(query)

    result?.forEach {
        print($0.id, $0.name)
    }
}

Show output

let model = ModelProvider().get("6626a3a8c8f1d089790cf5a2")

note

Model (and pipeline) inputs can be URLs, file paths, or direct text/labels (if applicable).
The examples below use only direct text.

Synchronous

Python

model.run("Tell me a joke about dogs.")

Show output

Use a dictionary to specify additional parameters or if the model takes multiple inputs.

model.run(
    {
        "text": "Tell me a joke about dogs.",
        "max_tokens": 10,
        "temperature": 0.5,
    }
)

Show output

Asynchronous

Python

start_response = model.run_async("Tell me a joke about dogs.")
start_response

Show output

Use the poll method to monitor inference progress.

while True:
  status = model.poll(start_response['url'])
  print(status)
  if status['status'] != 'IN_PROGRESS':
      break
  time.sleep(1)  # wait for 1 second before checking again

Show output

Swift

let result = model.run("Tell me a joke about dogs.")
print(result)

Show output

Use a dictionary to specify additional parameters or if the model takes multiple inputs.

let result = try await model.run({
    "text": "Tell me a joke about dogs.",
    "max_tokens": 10,
    "temperature": 0.5
})
print(result)

Show output

Process Data Assets

You can also perform inference on Data assets (Corpora or Datasets). You will need to onboard a data asset to use it.

note

Inference on Data assets is only available in Python.

Each data asset has an ID, and each column in that data asset has an ID, too. Specify both to perform inference:

Python

Run

result = model.run(
  data="64acbad666608858f693a3a0",
  data_asset="64acbad666608858f693a39f"
)

Run Async

start_response = model.run_async(
  data="64acbad666608858f693a3a0",
  data_asset="64acbad666608858f693a39f"
)

With these steps, you can easily call models, pipelines, and integrate them into your agents to perform AI tasks.