Classifier models based on YandexGPT
Yandex Foundation Models offers you YandexGPT Lite and YandexGPT Pro-based classifier models to classify text requests provided in prompts. The classification is implemented based on the Foundation ModelsText Classification API.
If an out-of-the-box model is not enough, you can fine-tune a YandexGPT Lite-based classifier for the model to classify your requests more accurately. To access a fine-tuned classifier model, use the classify Text Classification API method.
| Model | URI | Operating modes |
|---|---|---|
| Prompt-based classifier built on YandexGPT Lite | cls://<folder_ID>/yandexgpt-lite/latest |
Synchronous |
| Prompt-based classifier built on YandexGPT Pro | cls://<folder_ID>/yandexgpt/latest |
Synchronous |
| Fine-tuned classifier | cls://<basic_model_URI>/<version>@<tuning_suffix> |
Synchronous |
| Classifier model fine-tuned in DataSphere | cls://<folder_ID>/<fine-tuned_model_ID> |
Synchronous |
Accessing models
You can access classifier models in a number of ways.
When operating classifier models via Yandex Cloud ML SDK, use one of the following formats:
-
Model name, provided as a string.
model = ( sdk.models.text_classifiers("yandexgpt-lite") ) -
Model name and version, provided as strings in the
model_nameandmodel_versionfields, respectively.model = ( sdk.models.text_classifiers(model_name="yandexgpt", model_version="latest") ) -
Model URI, provided as a string containing the full URI of the model. You can also use this method to access fine-tuned models.
model = ( sdk.models.text_classifiers("cls://b1gt6g8ht345********/yandexgpt/latest") )
To access a model via the REST API or gRPC API, specify the model's URI containing the folder ID in the modelUri field of the request body. The /latest segment indicates the model version and is optional. To access a YandexGPT-based classifier model, use the fewShotClassify Text Classification API method/call.
Example:
{
"modelUri":
cls://b1gt6g8ht345********/yandexgpt-lite/latest
...
}
To access the Latest version, you do not need to specify the model version explicitly because Latest is used by default.