banking77

This is a text classification dataset. It is intended for machine learning research and experimentation.

This dataset is obtained via formatting another publicly available data to be compatible with our AutoIntent Library.

Usage

It is intended to be used with our AutoIntent Library:

from autointent import Dataset

banking77 = Dataset.from_datasets("AutoIntent/banking77")

Source

This dataset is taken from PolyAI/banking77 and formatted with our AutoIntent Library:

# define utils
import requests
import json
from autointent import Dataset
from datasets import load_dataset

def load_json_from_url(github_file: str):
    raw_text = requests.get(github_file).text
    return json.loads(raw_text)

def convert_banking77(banking77_train, shots_per_intent, intent_names):
    all_labels = sorted(banking77_train.unique("label"))
    n_classes = len(intent_names)
    assert all_labels == list(range(n_classes))

    classwise_utterance_records = [[] for _ in range(n_classes)]
    intents = [
        {
            "id": i,
            "name": name,
        }
        for i, name in enumerate(intent_names)
    ]

    for b77_batch in banking77_train.iter(batch_size=16, drop_last_batch=False):
        for txt, intent_id in zip(b77_batch["text"], b77_batch["label"], strict=False):
            target_list = classwise_utterance_records[intent_id]
            if shots_per_intent is not None and len(target_list) >= shots_per_intent:
                continue
            target_list.append({"utterance": txt, "label": intent_id})

    utterances = [rec for lst in classwise_utterance_records for rec in lst]
    return Dataset.from_dict({"intents": intents, "train": utterances})

# load 
file_url = "https://huggingface.co/datasets/PolyAI/banking77/resolve/main/dataset_infos.json"
dataset_description = load_json_from_url(file_url)
intent_names = dataset_description["default"]["features"]["label"]["names"]
banking77 = load_dataset("PolyAI/banking77", trust_remote_code=True)

# convert
banking77_converted = convert_banking77(
    banking77["train"],
    shots_per_intent=None,
    intent_names=intent_names
)