Tunable text-to-text

Overview

Tunable text-to-text estimators are estimators that can be tuned to perform a variety of tasks, including but not limited to text summarization, question answering, and text translation. These estimators use the provided data as-is, without any additional preprocessing, or constructing prompts. While this approach allows for more flexibility, it is the user's responsibility to ensure that the data is formatted correctly.

from skllm.models.gpt.text2text.tunable import TunableGPTText2Text

model = TunableGPTText2Text(
        base_model = "gpt-3.5-turbo-0613",
        n_epochs = None, # int or None. When None, will be determined automatically by OpenAI
        system_msg = "You are a text processing model."
)

model.fit(X_train, y_train) # y_train is any desired output text
out = model.transform(X_test)

API Reference

The following API reference only lists the parameters needed for the initialization of the estimator. The remaining methods follow the syntax of a scikit-learn transformer.

TunableGPTText2Text

from skllm.models.gpt.text2text.tunable import TunableGPTText2Text

Parameter	Type	Description
`base_model`	`str`	Model to use, by default "gpt-3.5-turbo-0613".
`key`	`Optional[str]`	Estimator-specific API key; if None, retrieved from the global config, by default None.
`org`	`Optional[str]`	Estimator-specific ORG key; if None, retrieved from the global config, by default None.
`n_epochs`	`Optional[int]`	Number of epochs; if None, determined automatically; by default None.
`custom_suffix`	`Optional[str]`	Custom suffix of the tuned model, used for naming purposes only, by default "skllm".

TunableVertexText2Text

from skllm.models.vertex.text2text.tunable import TunableVertexText2Text

Parameter	Type	Description
`base_model`	`str`	Model to use, by default "text-bison@002".
`n_update_steps`	`int`	Number of epochs, by default 1.