Helpers¶

class ibm_watsonx_ai.foundation_models_manager.FoundationModelsManager(client)[source]¶

get_model_specs(model_id=None, limit=None, asynchronous=False, get_all=False, filters='function_text_generation,!lifecycle_withdrawn:and', **kwargs)[source]¶

Retrieves a list of specifications for a deployed foundation model.

Parameters:

model_id (str or ModelTypes, optional) – id of the model, defaults to None (all models specifications are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, will work as a generator
get_all (bool, optional) – if True, will get all entries in ‘limited’ chunks

filters (str, list[str], optional) –

a set of filters to specify the list of models, default to function_text_generation Filters are described by the following pattern:

pattern: tfilter[,tfilter][:(or|and)] tfilter: filter | !filter

filter – requires existence of the filter.
!filter – requires absence of the filter.

Available filters:

Filter	Description	Example
`modelid_*`	Filters by model id. Selects a model with a specific model id.	`modelid_ibm/granite-3-3-8b-instruct`
`provider_*`	Filters by provider. Selects all models with a specific provider.	`provider_IBM`
`source_*`	Filters by source. Selects all models with a specific source.	`source_IBM`
`input_tier_*`	Filters by input tier. Selects all models with a specific input tier.	`input_tier_class_1`
`output_tier_*`	Filters by output tier. Selects all models with a specific output tier.	`output_tier_class_c1`
`tier_*`	Filters by tier. Selects all models with a specific input or output tier.	`tier_class_12`
`task_*`	Filters by task id. Selects all models supporting a specific task id.	`task_generation`
`lifecycle_*`	Filters by lifecycle state. Selects all models in the specified lifecycle state.	`lifecycle_available`
`function_*`	Filters by function. Selects all models supporting a specific function.	`function_text_chat`

Returns:

list of specifications for the deployed foundation model

Return type:

dict or generator

Example:

# GET ALL MODEL SPECS
client.foundation_models.get_model_specs()

# GET MODEL SPECS BY MODEL_ID
client.foundation_models.get_model_specs(model_id="google/flan-ul2")

# GET MODEL SPECS WITH ONE FILTER
client.foundation_models.get_model_specs(filters="function_text_generation")

# GET MODEL SPECS WITH MULTIPLE FILTERS
client.foundation_models.get_model_specs(filters="function_text_generation,function_text_chat,!lifecycle_withdrawn:and")

get_chat_model_specs(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Operations to retrieve the list of chat foundation models specifications.

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specs are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, it will work as a generator
get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

list of deployed foundation model specs

Return type:

dict or generator

Example

# GET CHAT MODEL SPECS
client.foundation_models.get_chat_model_specs()

# GET CHAT MODEL SPECS BY MODEL_ID
client.foundation_models.get_chat_model_specs(model_id="ibm/granite-13b-chat-v2")

get_chat_function_calling_model_specs(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Operations to retrieve the list of chat foundation models specifications with function calling support .

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specs are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, it will work as a generator
get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

list of deployed foundation model specs

Return type:

dict or generator

Example

# GET CHAT FUNCTION CALLING MODEL SPECS
client.foundation_models.get_chat_function_calling_model_specs()

# GET CHAT FUNCTION CALLING MODEL SPECS BY MODEL_ID
client.foundation_models.get_chat_function_calling_model_specs(model_id="meta-llama/llama-3-1-70b-instruct")

get_audio_chat_model_specs(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Operations to retrieve the list of audio chat foundation models specifications.

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specs are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, it will work as a generator
get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

list of deployed foundation model specs

Return type:

dict or generator

Example

# GET AUDIO CHAT MODEL SPECS
client.foundation_models.get_audio_chat_model_specs()

# GET AUDIO CHAT MODEL SPECS BY MODEL_ID
client.foundation_models.get_audio_chat_model_specs(model_id="ibm/granite-speech-3-3-8b")

get_text_generation_model_specs(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Operations to retrieve the list of text generation foundation models specifications.

Parameters:

model_id (str or ModelTypes, optional) – id of the model, defaults to None (all models specifications are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, will work as a generator
get_all (bool, optional) – if True, will get all entries in ‘limited’ chunks

Returns:

list of specifications for the deployed foundation model

Return type:

dict or generator

Example:

# GET TEXT GENERATION MODEL SPECS
client.foundation_models.get_text_generation_model_specs()

# GET TEXT GENERATION MODEL SPECS BY MODEL_ID
client.foundation_models.get_text_generation_model_specs(model_id="google/flan-ul2")

get_custom_model_specs(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Get details on available custom model(s) as a dictionary or as a generator (asynchronous). If asynchronous or get_all is set, then model_id is ignored.

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specifications are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, will work as a generator
get_all (bool, optional) – if True, will get all entries in ‘limited’ chunks

Returns:

details of supported custom models, None if no supported custom models are found for the given model_id

Return type:

dict or generator

Example:

client.foundation_models.get_custom_models_spec()
client.foundation_models.get_custom_models_spec()
client.foundation_models.get_custom_models_spec(model_id='mistralai/Mistral-7B-Instruct-v0.2')
client.foundation_models.get_custom_models_spec(limit=20)
client.foundation_models.get_custom_models_spec(limit=20, get_all=True)
for spec in client.foundation_models.get_custom_model_specs(limit=20, asynchronous=True, get_all=True):
    print(spec, end="")

get_embeddings_model_specs(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Retrieves the specifications of an embeddings model.

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specifications are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, will work as a generator
get_all (bool, optional) – if True, will get all entries in ‘limited’ chunks

Returns:

specifications of the embeddings model

Return type:

dict or generator

Example:

client.foundation_models.get_embeddings_model_specs()
client.foundation_models.get_embeddings_model_specs('ibm/slate-125m-english-rtrvr')

get_time_series_model_specs(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Retrieves the specifications of an time series model.

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specifications are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, will work as a generator
get_all (bool, optional) – if True, will get all entries in ‘limited’ chunks

Returns:

specifications of the time series model

Return type:

dict or generator

Example:

client.foundation_models.get_time_series_model_specs()
client.foundation_models.get_time_series_model_specs('ibm/granite-ttm-1536-96-r2')

get_rerank_model_specs(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Retrieves the specifications of a rerank model.

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specifications are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, will work as a generator
get_all (bool, optional) – if True, will get all entries in ‘limited’ chunks

Returns:

specifications of the rerank model

Return type:

dict or generator

Example:

client.foundation_models.get_rerank_model_specs()
client.foundation_models.get_rerank_model_specs('ibm/slate-125m-english-rtrvr-v2')

get_audio_transcriptions_model_specs(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Retrieves the specifications of an audio transcriptions model.

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specifications are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, will work as a generator
get_all (bool, optional) – if True, will get all entries in ‘limited’ chunks

Returns:

specifications of the audio transcriptions model

Return type:

dict or generator

Example:

client.foundation_models.get_audio_transcriptions_model_specs()
client.foundation_models.get_audio_transcriptions_model_specs('openai/whisper-tiny')

get_base_foundation_model_deployable_specs(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Retrieve the specifications of a base deployable foundation models

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specifications are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, will work as a generator
get_all (bool, optional) – if True, will get all entries in ‘limited’ chunks

Returns:

specifications of the base foundation model deployable

Return type:

dict or generator

Example:

client.foundation_models.get_base_foundation_model_deployable_specs()
client.foundation_models.get_base_foundation_model_deployable_specs('meta-llama/llama-3-1-8b')

get_model_specs_with_prompt_tuning_support(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Queries the details of deployed foundation models with prompt tuning support.

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specifications are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, will work as a generator
get_all (bool, optional) – if True, will get all entries in ‘limited’ chunks

Returns:

list of specifications of a deployed foundation model with prompt tuning support

Return type:

dict or generator

Example:

client.foundation_models.get_model_specs_with_prompt_tuning_support()
client.foundation_models.get_model_specs_with_prompt_tuning_support('google/flan-t5-xl')

get_model_specs_with_fine_tuning_support(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Operations to query the details of the deployed foundation models with fine-tuning support.

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specs are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, it will work as a generator
get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

list of deployed foundation model specs with fine-tuning support

Return type:

dict or generator

Example

client.foundation_models.get_model_specs_with_fine_tuning_support()
client.foundation_models.get_model_specs_with_fine_tuning_support('bigscience/bloom')

get_model_specs_with_lora_fine_tuning_support(model_id=None, limit=None, asynchronous=False, get_all=False)[source]¶

Operations to query the details of the deployed foundation models with lora fine-tuning support.

Parameters:

model_id (str, optional) – id of the model, defaults to None (all models specs are returned)
limit (int, optional) – limit number of fetched records
asynchronous (bool, optional) – if True, it will work as a generator
get_all (bool, optional) – if True, it will get all entries in ‘limited’ chunks

Returns:

list of deployed foundation model specs with lora fine-tuning support

Return type:

dict or generator

Example

client.foundation_models.get_model_specs_with_lora_fine_tuning_support()
client.foundation_models.get_model_specs_with_lora_fine_tuning_support('bigscience/bloom')

get_model_lifecycle(model_id, **kwargs)[source]¶

Retrieves a list of lifecycle data of a foundation model.

Parameters:: model_id (str) – id of the model
Returns:: list of lifecycle data of a foundation model
Return type:: list

Example:

client.foundation_models.get_model_lifecycle(
    model_id="ibm/granite-13b-instruct-v2"
    )

ibm_watsonx_ai.foundation_models.get_model_specs(url, model_id=None)[source]¶

Retrieve the list of deployed foundation models specifications.

Deprecated since version v1.0: The get_model_specs() function is deprecated, use the client.foundation_models.get_model_specs() function instead.

Parameters:

url (str) – URL of the environment
model_id (Optional[str, ModelTypes], optional) – ID of the model, defaults to None (all models specs are returned).

Returns:

list of deployed foundation model specs

Return type:

dict

Example:

from ibm_watsonx_ai.foundation_models import get_model_specs

# GET ALL MODEL SPECS
get_model_specs(
    url="https://us-south.ml.cloud.ibm.com"
    )

# GET MODEL SPECS BY MODEL_ID
get_model_specs(
    url="https://us-south.ml.cloud.ibm.com",
    model_id="google/flan-ul2"
    )

ibm_watsonx_ai.foundation_models.get_model_lifecycle(url, model_id)[source]¶

Retrieve the list of model lifecycle data.

Deprecated since version v1.0: The get_model_lifecycle() function is deprecated, use the client.foundation_models.get_model_lifecycle() function instead.

Parameters:

url (str) – URL of environment
model_id (str) – the type of model to use

Returns:

list of deployed foundation model lifecycle data

Return type:

list

Example:

from ibm_watsonx_ai.foundation_models import get_model_lifecycle

get_model_lifecycle(
    url="https://us-south.ml.cloud.ibm.com",
    model_id="ibm/granite-13b-instruct-v2"
    )

ibm_watsonx_ai.foundation_models.get_model_specs_with_prompt_tuning_support(url)[source]¶

Query the details of the deployed foundation models with prompt tuning support.

Deprecated since version v1.0: The get_model_specs_with_prompt_tuning_support() function is deprecated, use the client.foundation_models.get_model_specs_with_prompt_tuning_support() function instead.

Parameters:: url (str) – URL of environment
Returns:: list of deployed foundation model specs with prompt tuning support
Return type:: dict

Example:

from ibm_watsonx_ai.foundation_models import get_model_specs_with_prompt_tuning_support

get_model_specs_with_prompt_tuning_support(
    url="https://us-south.ml.cloud.ibm.com"
    )

ibm_watsonx_ai.foundation_models.get_supported_tasks(url)[source]¶

Retrieves a list of tasks that are supported by the foundation models.

Parameters:: url (str) – URL of the environment
Returns:: list of tasks that are supported by the foundation models
Return type:: dict

Example:

from ibm_watsonx_ai.foundation_models import get_supported_tasks

get_supported_tasks(
    url="https://us-south.ml.cloud.ibm.com"
    )