rllm.llm.LLMWithRetriever¶

class rllm.llm.LLMWithRetriever(file_path: str, metadata_path: str, task_info_path: str, label_info_path: str, llm: BaseChatModel | BaseLLM | None = None, *, shots: int = 4, test_size: float | int = 0.2, target_column: str | None = None, seed: int = 0)[source]¶

Bases: object

A class that integrates a language model with a retriever for classification tasks.

df¶

The dataframe loaded from the CSV file.

Type:: pd.DataFrame

dir¶

The directory of the input file.

Type:: str

data_name¶

The name of the dataset (without extension).

Type:: str

metadata¶

Metadata loaded from the JSON file.

Type:: dict

task_info¶

Task description loaded from the TXT file.

Type:: str

label_info¶

Label information loaded from the TXT file.

Type:: str

llm¶

The language model instance.

Type:: Union[LC.BaseChatModel, LC.BaseLLM]

shots¶

Number of examples to retrieve for few-shot learning.

Type:: int

seed¶

Random seed for reproducibility.

Type:: int

train_file_path¶

Path to the training data CSV file.

Type:: str

test_file_path¶

Path to the testing data CSV file.

Type:: str

label_file_path¶

Path to the label data CSV file.

Type:: str

retriever¶

Instance of the retriever for fetching similar examples.

Type:: SingleTableRetriever

invoke()[source]¶

Executes the inference process by retrieving neighbors for test instances and querying the language model for predictions.

Saves the results to a JSON file in the same directory as the input file.