Datasets:

Modalities:
Text
Languages:
English
ArXiv:
Libraries:
Datasets
License:
babi_qa / babi_qa.py
system's picture
system HF staff
Update files from the datasets library (from 1.16.0)
07549ca
raw
history blame
41.3 kB
# coding=utf-8
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""The bAbI tasks dataset."""
import datasets
_CITATION = """\
@misc{weston2015aicomplete,
title={Towards AI-Complete Question Answering: A Set of Prerequisite Toy Tasks},
author={Jason Weston and Antoine Bordes and Sumit Chopra and Alexander M. Rush and Bart van Merriënboer and Armand Joulin and Tomas Mikolov},
year={2015},
eprint={1502.05698},
archivePrefix={arXiv},
primaryClass={cs.AI}
}
"""
_DESCRIPTION = """\
The (20) QA bAbI tasks are a set of proxy tasks that evaluate reading
comprehension via question answering. Our tasks measure understanding
in several ways: whether a system is able to answer questions via chaining facts,
simple induction, deduction and many more. The tasks are designed to be prerequisites
for any system that aims to be capable of conversing with a human.
The aim is to classify these tasks into skill sets,so that researchers
can identify (and then rectify)the failings of their systems.
"""
_HOMEPAGE = "https://research.fb.com/downloads/babi/"
_LICENSE = """Creative Commons Attribution 3.0 License"""
ZIP_URL = "http://www.thespermwhale.com/jaseweston/babi/tasks_1-20_v1-2.tar.gz"
paths = {
"en": {
"qa9": {
"test": "tasks_1-20_v1-2/en/qa9_simple-negation_test.txt",
"train": "tasks_1-20_v1-2/en/qa9_simple-negation_train.txt",
},
"qa4": {
"train": "tasks_1-20_v1-2/en/qa4_two-arg-relations_train.txt",
"test": "tasks_1-20_v1-2/en/qa4_two-arg-relations_test.txt",
},
"qa6": {
"train": "tasks_1-20_v1-2/en/qa6_yes-no-questions_train.txt",
"test": "tasks_1-20_v1-2/en/qa6_yes-no-questions_test.txt",
},
"qa11": {
"test": "tasks_1-20_v1-2/en/qa11_basic-coreference_test.txt",
"train": "tasks_1-20_v1-2/en/qa11_basic-coreference_train.txt",
},
"qa3": {
"test": "tasks_1-20_v1-2/en/qa3_three-supporting-facts_test.txt",
"train": "tasks_1-20_v1-2/en/qa3_three-supporting-facts_train.txt",
},
"qa15": {
"test": "tasks_1-20_v1-2/en/qa15_basic-deduction_test.txt",
"train": "tasks_1-20_v1-2/en/qa15_basic-deduction_train.txt",
},
"out.txt": {"out": "tasks_1-20_v1-2/en/out.txt"},
"qa17": {
"test": "tasks_1-20_v1-2/en/qa17_positional-reasoning_test.txt",
"train": "tasks_1-20_v1-2/en/qa17_positional-reasoning_train.txt",
},
"qa13": {
"test": "tasks_1-20_v1-2/en/qa13_compound-coreference_test.txt",
"train": "tasks_1-20_v1-2/en/qa13_compound-coreference_train.txt",
},
"qa1": {
"train": "tasks_1-20_v1-2/en/qa1_single-supporting-fact_train.txt",
"test": "tasks_1-20_v1-2/en/qa1_single-supporting-fact_test.txt",
},
"qa14": {
"train": "tasks_1-20_v1-2/en/qa14_time-reasoning_train.txt",
"test": "tasks_1-20_v1-2/en/qa14_time-reasoning_test.txt",
},
"qa16": {
"test": "tasks_1-20_v1-2/en/qa16_basic-induction_test.txt",
"train": "tasks_1-20_v1-2/en/qa16_basic-induction_train.txt",
},
"qa19": {
"test": "tasks_1-20_v1-2/en/qa19_path-finding_test.txt",
"train": "tasks_1-20_v1-2/en/qa19_path-finding_train.txt",
},
"qa18": {
"test": "tasks_1-20_v1-2/en/qa18_size-reasoning_test.txt",
"train": "tasks_1-20_v1-2/en/qa18_size-reasoning_train.txt",
},
"qa10": {
"train": "tasks_1-20_v1-2/en/qa10_indefinite-knowledge_train.txt",
"test": "tasks_1-20_v1-2/en/qa10_indefinite-knowledge_test.txt",
},
"qa7": {
"train": "tasks_1-20_v1-2/en/qa7_counting_train.txt",
"test": "tasks_1-20_v1-2/en/qa7_counting_test.txt",
},
"qa5": {
"test": "tasks_1-20_v1-2/en/qa5_three-arg-relations_test.txt",
"train": "tasks_1-20_v1-2/en/qa5_three-arg-relations_train.txt",
},
"qa12": {
"test": "tasks_1-20_v1-2/en/qa12_conjunction_test.txt",
"train": "tasks_1-20_v1-2/en/qa12_conjunction_train.txt",
},
"qa2": {
"train": "tasks_1-20_v1-2/en/qa2_two-supporting-facts_train.txt",
"test": "tasks_1-20_v1-2/en/qa2_two-supporting-facts_test.txt",
},
"qa20": {
"train": "tasks_1-20_v1-2/en/qa20_agents-motivations_train.txt",
"test": "tasks_1-20_v1-2/en/qa20_agents-motivations_test.txt",
},
"qa8": {
"train": "tasks_1-20_v1-2/en/qa8_lists-sets_train.txt",
"test": "tasks_1-20_v1-2/en/qa8_lists-sets_test.txt",
},
},
"en-10k": {
"qa9": {
"test": "tasks_1-20_v1-2/en-10k/qa9_simple-negation_test.txt",
"train": "tasks_1-20_v1-2/en-10k/qa9_simple-negation_train.txt",
},
"qa4": {
"train": "tasks_1-20_v1-2/en-10k/qa4_two-arg-relations_train.txt",
"test": "tasks_1-20_v1-2/en-10k/qa4_two-arg-relations_test.txt",
},
"qa6": {
"train": "tasks_1-20_v1-2/en-10k/qa6_yes-no-questions_train.txt",
"test": "tasks_1-20_v1-2/en-10k/qa6_yes-no-questions_test.txt",
},
"qa11": {
"test": "tasks_1-20_v1-2/en-10k/qa11_basic-coreference_test.txt",
"train": "tasks_1-20_v1-2/en-10k/qa11_basic-coreference_train.txt",
},
"qa3": {
"test": "tasks_1-20_v1-2/en-10k/qa3_three-supporting-facts_test.txt",
"train": "tasks_1-20_v1-2/en-10k/qa3_three-supporting-facts_train.txt",
},
"qa15": {
"test": "tasks_1-20_v1-2/en-10k/qa15_basic-deduction_test.txt",
"train": "tasks_1-20_v1-2/en-10k/qa15_basic-deduction_train.txt",
},
"qa17": {
"test": "tasks_1-20_v1-2/en-10k/qa17_positional-reasoning_test.txt",
"train": "tasks_1-20_v1-2/en-10k/qa17_positional-reasoning_train.txt",
},
"qa13": {
"test": "tasks_1-20_v1-2/en-10k/qa13_compound-coreference_test.txt",
"train": "tasks_1-20_v1-2/en-10k/qa13_compound-coreference_train.txt",
},
"qa1": {
"train": "tasks_1-20_v1-2/en-10k/qa1_single-supporting-fact_train.txt",
"test": "tasks_1-20_v1-2/en-10k/qa1_single-supporting-fact_test.txt",
},
"qa14": {
"train": "tasks_1-20_v1-2/en-10k/qa14_time-reasoning_train.txt",
"test": "tasks_1-20_v1-2/en-10k/qa14_time-reasoning_test.txt",
},
"qa16": {
"test": "tasks_1-20_v1-2/en-10k/qa16_basic-induction_test.txt",
"train": "tasks_1-20_v1-2/en-10k/qa16_basic-induction_train.txt",
},
"qa19": {
"test": "tasks_1-20_v1-2/en-10k/qa19_path-finding_test.txt",
"train": "tasks_1-20_v1-2/en-10k/qa19_path-finding_train.txt",
},
"qa18": {
"test": "tasks_1-20_v1-2/en-10k/qa18_size-reasoning_test.txt",
"train": "tasks_1-20_v1-2/en-10k/qa18_size-reasoning_train.txt",
},
"qa10": {
"train": "tasks_1-20_v1-2/en-10k/qa10_indefinite-knowledge_train.txt",
"test": "tasks_1-20_v1-2/en-10k/qa10_indefinite-knowledge_test.txt",
},
"qa7": {
"train": "tasks_1-20_v1-2/en-10k/qa7_counting_train.txt",
"test": "tasks_1-20_v1-2/en-10k/qa7_counting_test.txt",
},
"qa5": {
"test": "tasks_1-20_v1-2/en-10k/qa5_three-arg-relations_test.txt",
"train": "tasks_1-20_v1-2/en-10k/qa5_three-arg-relations_train.txt",
},
"qa12": {
"test": "tasks_1-20_v1-2/en-10k/qa12_conjunction_test.txt",
"train": "tasks_1-20_v1-2/en-10k/qa12_conjunction_train.txt",
},
"qa2": {
"train": "tasks_1-20_v1-2/en-10k/qa2_two-supporting-facts_train.txt",
"test": "tasks_1-20_v1-2/en-10k/qa2_two-supporting-facts_test.txt",
},
"qa20": {
"train": "tasks_1-20_v1-2/en-10k/qa20_agents-motivations_train.txt",
"test": "tasks_1-20_v1-2/en-10k/qa20_agents-motivations_test.txt",
},
"qa8": {
"train": "tasks_1-20_v1-2/en-10k/qa8_lists-sets_train.txt",
"test": "tasks_1-20_v1-2/en-10k/qa8_lists-sets_test.txt",
},
},
"en-valid": {
"qa5": {
"train": "tasks_1-20_v1-2/en-valid/qa5_train.txt",
"test": "tasks_1-20_v1-2/en-valid/qa5_test.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa5_valid.txt",
},
"qa16": {
"valid": "tasks_1-20_v1-2/en-valid/qa16_valid.txt",
"test": "tasks_1-20_v1-2/en-valid/qa16_test.txt",
"train": "tasks_1-20_v1-2/en-valid/qa16_train.txt",
},
"qa2": {
"valid": "tasks_1-20_v1-2/en-valid/qa2_valid.txt",
"test": "tasks_1-20_v1-2/en-valid/qa2_test.txt",
"train": "tasks_1-20_v1-2/en-valid/qa2_train.txt",
},
"qa15": {
"train": "tasks_1-20_v1-2/en-valid/qa15_train.txt",
"test": "tasks_1-20_v1-2/en-valid/qa15_test.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa15_valid.txt",
},
"qa9": {
"test": "tasks_1-20_v1-2/en-valid/qa9_test.txt",
"train": "tasks_1-20_v1-2/en-valid/qa9_train.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa9_valid.txt",
},
"qa1": {
"valid": "tasks_1-20_v1-2/en-valid/qa1_valid.txt",
"test": "tasks_1-20_v1-2/en-valid/qa1_test.txt",
"train": "tasks_1-20_v1-2/en-valid/qa1_train.txt",
},
"qa4": {
"test": "tasks_1-20_v1-2/en-valid/qa4_test.txt",
"train": "tasks_1-20_v1-2/en-valid/qa4_train.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa4_valid.txt",
},
"qa14": {
"valid": "tasks_1-20_v1-2/en-valid/qa14_valid.txt",
"train": "tasks_1-20_v1-2/en-valid/qa14_train.txt",
"test": "tasks_1-20_v1-2/en-valid/qa14_test.txt",
},
"qa3": {
"test": "tasks_1-20_v1-2/en-valid/qa3_test.txt",
"train": "tasks_1-20_v1-2/en-valid/qa3_train.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa3_valid.txt",
},
"qa6": {
"valid": "tasks_1-20_v1-2/en-valid/qa6_valid.txt",
"test": "tasks_1-20_v1-2/en-valid/qa6_test.txt",
"train": "tasks_1-20_v1-2/en-valid/qa6_train.txt",
},
"qa8": {
"test": "tasks_1-20_v1-2/en-valid/qa8_test.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa8_valid.txt",
"train": "tasks_1-20_v1-2/en-valid/qa8_train.txt",
},
"qa20": {
"train": "tasks_1-20_v1-2/en-valid/qa20_train.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa20_valid.txt",
"test": "tasks_1-20_v1-2/en-valid/qa20_test.txt",
},
"qa11": {
"test": "tasks_1-20_v1-2/en-valid/qa11_test.txt",
"train": "tasks_1-20_v1-2/en-valid/qa11_train.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa11_valid.txt",
},
"qa12": {
"test": "tasks_1-20_v1-2/en-valid/qa12_test.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa12_valid.txt",
"train": "tasks_1-20_v1-2/en-valid/qa12_train.txt",
},
"qa13": {
"test": "tasks_1-20_v1-2/en-valid/qa13_test.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa13_valid.txt",
"train": "tasks_1-20_v1-2/en-valid/qa13_train.txt",
},
"qa7": {
"train": "tasks_1-20_v1-2/en-valid/qa7_train.txt",
"test": "tasks_1-20_v1-2/en-valid/qa7_test.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa7_valid.txt",
},
"qa19": {
"valid": "tasks_1-20_v1-2/en-valid/qa19_valid.txt",
"test": "tasks_1-20_v1-2/en-valid/qa19_test.txt",
"train": "tasks_1-20_v1-2/en-valid/qa19_train.txt",
},
"qa17": {
"train": "tasks_1-20_v1-2/en-valid/qa17_train.txt",
"test": "tasks_1-20_v1-2/en-valid/qa17_test.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa17_valid.txt",
},
"qa10": {
"test": "tasks_1-20_v1-2/en-valid/qa10_test.txt",
"valid": "tasks_1-20_v1-2/en-valid/qa10_valid.txt",
"train": "tasks_1-20_v1-2/en-valid/qa10_train.txt",
},
"qa18": {
"valid": "tasks_1-20_v1-2/en-valid/qa18_valid.txt",
"train": "tasks_1-20_v1-2/en-valid/qa18_train.txt",
"test": "tasks_1-20_v1-2/en-valid/qa18_test.txt",
},
},
"en-valid-10k": {
"qa5": {
"train": "tasks_1-20_v1-2/en-valid-10k/qa5_train.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa5_test.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa5_valid.txt",
},
"qa16": {
"valid": "tasks_1-20_v1-2/en-valid-10k/qa16_valid.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa16_test.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa16_train.txt",
},
"qa2": {
"valid": "tasks_1-20_v1-2/en-valid-10k/qa2_valid.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa2_test.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa2_train.txt",
},
"qa15": {
"train": "tasks_1-20_v1-2/en-valid-10k/qa15_train.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa15_test.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa15_valid.txt",
},
"qa9": {
"test": "tasks_1-20_v1-2/en-valid-10k/qa9_test.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa9_train.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa9_valid.txt",
},
"qa1": {
"valid": "tasks_1-20_v1-2/en-valid-10k/qa1_valid.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa1_test.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa1_train.txt",
},
"qa4": {
"test": "tasks_1-20_v1-2/en-valid-10k/qa4_test.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa4_train.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa4_valid.txt",
},
"qa14": {
"valid": "tasks_1-20_v1-2/en-valid-10k/qa14_valid.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa14_train.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa14_test.txt",
},
"qa3": {
"test": "tasks_1-20_v1-2/en-valid-10k/qa3_test.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa3_train.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa3_valid.txt",
},
"qa6": {
"valid": "tasks_1-20_v1-2/en-valid-10k/qa6_valid.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa6_test.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa6_train.txt",
},
"qa8": {
"test": "tasks_1-20_v1-2/en-valid-10k/qa8_test.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa8_valid.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa8_train.txt",
},
"qa20": {
"train": "tasks_1-20_v1-2/en-valid-10k/qa20_train.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa20_valid.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa20_test.txt",
},
"qa11": {
"test": "tasks_1-20_v1-2/en-valid-10k/qa11_test.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa11_train.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa11_valid.txt",
},
"qa12": {
"test": "tasks_1-20_v1-2/en-valid-10k/qa12_test.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa12_valid.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa12_train.txt",
},
"qa13": {
"test": "tasks_1-20_v1-2/en-valid-10k/qa13_test.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa13_valid.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa13_train.txt",
},
"qa7": {
"train": "tasks_1-20_v1-2/en-valid-10k/qa7_train.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa7_test.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa7_valid.txt",
},
"qa19": {
"valid": "tasks_1-20_v1-2/en-valid-10k/qa19_valid.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa19_test.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa19_train.txt",
},
"qa17": {
"train": "tasks_1-20_v1-2/en-valid-10k/qa17_train.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa17_test.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa17_valid.txt",
},
"qa10": {
"test": "tasks_1-20_v1-2/en-valid-10k/qa10_test.txt",
"valid": "tasks_1-20_v1-2/en-valid-10k/qa10_valid.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa10_train.txt",
},
"qa18": {
"valid": "tasks_1-20_v1-2/en-valid-10k/qa18_valid.txt",
"train": "tasks_1-20_v1-2/en-valid-10k/qa18_train.txt",
"test": "tasks_1-20_v1-2/en-valid-10k/qa18_test.txt",
},
},
"hn": {
"qa9": {
"test": "tasks_1-20_v1-2/hn/qa9_simple-negation_test.txt",
"train": "tasks_1-20_v1-2/hn/qa9_simple-negation_train.txt",
},
"qa4": {
"train": "tasks_1-20_v1-2/hn/qa4_two-arg-relations_train.txt",
"test": "tasks_1-20_v1-2/hn/qa4_two-arg-relations_test.txt",
},
"qa6": {
"train": "tasks_1-20_v1-2/hn/qa6_yes-no-questions_train.txt",
"test": "tasks_1-20_v1-2/hn/qa6_yes-no-questions_test.txt",
},
"qa11": {
"test": "tasks_1-20_v1-2/hn/qa11_basic-coreference_test.txt",
"train": "tasks_1-20_v1-2/hn/qa11_basic-coreference_train.txt",
},
"qa3": {
"test": "tasks_1-20_v1-2/hn/qa3_three-supporting-facts_test.txt",
"train": "tasks_1-20_v1-2/hn/qa3_three-supporting-facts_train.txt",
},
"qa15": {
"test": "tasks_1-20_v1-2/hn/qa15_basic-deduction_test.txt",
"train": "tasks_1-20_v1-2/hn/qa15_basic-deduction_train.txt",
},
"qa17": {
"test": "tasks_1-20_v1-2/hn/qa17_positional-reasoning_test.txt",
"train": "tasks_1-20_v1-2/hn/qa17_positional-reasoning_train.txt",
},
"qa13": {
"test": "tasks_1-20_v1-2/hn/qa13_compound-coreference_test.txt",
"train": "tasks_1-20_v1-2/hn/qa13_compound-coreference_train.txt",
},
"qa1": {
"train": "tasks_1-20_v1-2/hn/qa1_single-supporting-fact_train.txt",
"test": "tasks_1-20_v1-2/hn/qa1_single-supporting-fact_test.txt",
},
"qa14": {
"train": "tasks_1-20_v1-2/hn/qa14_time-reasoning_train.txt",
"test": "tasks_1-20_v1-2/hn/qa14_time-reasoning_test.txt",
},
"qa16": {
"test": "tasks_1-20_v1-2/hn/qa16_basic-induction_test.txt",
"train": "tasks_1-20_v1-2/hn/qa16_basic-induction_train.txt",
},
"qa19": {
"test": "tasks_1-20_v1-2/hn/qa19_path-finding_test.txt",
"train": "tasks_1-20_v1-2/hn/qa19_path-finding_train.txt",
},
"qa18": {
"test": "tasks_1-20_v1-2/hn/qa18_size-reasoning_test.txt",
"train": "tasks_1-20_v1-2/hn/qa18_size-reasoning_train.txt",
},
"qa10": {
"train": "tasks_1-20_v1-2/hn/qa10_indefinite-knowledge_train.txt",
"test": "tasks_1-20_v1-2/hn/qa10_indefinite-knowledge_test.txt",
},
"qa7": {
"train": "tasks_1-20_v1-2/hn/qa7_counting_train.txt",
"test": "tasks_1-20_v1-2/hn/qa7_counting_test.txt",
},
"qa5": {
"test": "tasks_1-20_v1-2/hn/qa5_three-arg-relations_test.txt",
"train": "tasks_1-20_v1-2/hn/qa5_three-arg-relations_train.txt",
},
"qa12": {
"test": "tasks_1-20_v1-2/hn/qa12_conjunction_test.txt",
"train": "tasks_1-20_v1-2/hn/qa12_conjunction_train.txt",
},
"qa2": {
"train": "tasks_1-20_v1-2/hn/qa2_two-supporting-facts_train.txt",
"test": "tasks_1-20_v1-2/hn/qa2_two-supporting-facts_test.txt",
},
"qa20": {
"train": "tasks_1-20_v1-2/hn/qa20_agents-motivations_train.txt",
"test": "tasks_1-20_v1-2/hn/qa20_agents-motivations_test.txt",
},
"qa8": {
"train": "tasks_1-20_v1-2/hn/qa8_lists-sets_train.txt",
"test": "tasks_1-20_v1-2/hn/qa8_lists-sets_test.txt",
},
},
"hn-10k": {
"qa9": {
"test": "tasks_1-20_v1-2/hn-10k/qa9_simple-negation_test.txt",
"train": "tasks_1-20_v1-2/hn-10k/qa9_simple-negation_train.txt",
},
"qa4": {
"train": "tasks_1-20_v1-2/hn-10k/qa4_two-arg-relations_train.txt",
"test": "tasks_1-20_v1-2/hn-10k/qa4_two-arg-relations_test.txt",
},
"qa6": {
"train": "tasks_1-20_v1-2/hn-10k/qa6_yes-no-questions_train.txt",
"test": "tasks_1-20_v1-2/hn-10k/qa6_yes-no-questions_test.txt",
},
"qa11": {
"test": "tasks_1-20_v1-2/hn-10k/qa11_basic-coreference_test.txt",
"train": "tasks_1-20_v1-2/hn-10k/qa11_basic-coreference_train.txt",
},
"qa3": {
"test": "tasks_1-20_v1-2/hn-10k/qa3_three-supporting-facts_test.txt",
"train": "tasks_1-20_v1-2/hn-10k/qa3_three-supporting-facts_train.txt",
},
"qa15": {
"test": "tasks_1-20_v1-2/hn-10k/qa15_basic-deduction_test.txt",
"train": "tasks_1-20_v1-2/hn-10k/qa15_basic-deduction_train.txt",
},
"qa17": {
"test": "tasks_1-20_v1-2/hn-10k/qa17_positional-reasoning_test.txt",
"train": "tasks_1-20_v1-2/hn-10k/qa17_positional-reasoning_train.txt",
},
"qa13": {
"test": "tasks_1-20_v1-2/hn-10k/qa13_compound-coreference_test.txt",
"train": "tasks_1-20_v1-2/hn-10k/qa13_compound-coreference_train.txt",
},
"qa1": {
"train": "tasks_1-20_v1-2/hn-10k/qa1_single-supporting-fact_train.txt",
"test": "tasks_1-20_v1-2/hn-10k/qa1_single-supporting-fact_test.txt",
},
"qa14": {
"train": "tasks_1-20_v1-2/hn-10k/qa14_time-reasoning_train.txt",
"test": "tasks_1-20_v1-2/hn-10k/qa14_time-reasoning_test.txt",
},
"qa16": {
"test": "tasks_1-20_v1-2/hn-10k/qa16_basic-induction_test.txt",
"train": "tasks_1-20_v1-2/hn-10k/qa16_basic-induction_train.txt",
},
"qa19": {
"test": "tasks_1-20_v1-2/hn-10k/qa19_path-finding_test.txt",
"train": "tasks_1-20_v1-2/hn-10k/qa19_path-finding_train.txt",
},
"qa18": {
"test": "tasks_1-20_v1-2/hn-10k/qa18_size-reasoning_test.txt",
"train": "tasks_1-20_v1-2/hn-10k/qa18_size-reasoning_train.txt",
},
"qa10": {
"train": "tasks_1-20_v1-2/hn-10k/qa10_indefinite-knowledge_train.txt",
"test": "tasks_1-20_v1-2/hn-10k/qa10_indefinite-knowledge_test.txt",
},
"qa7": {
"train": "tasks_1-20_v1-2/hn-10k/qa7_counting_train.txt",
"test": "tasks_1-20_v1-2/hn-10k/qa7_counting_test.txt",
},
"qa5": {
"test": "tasks_1-20_v1-2/hn-10k/qa5_three-arg-relations_test.txt",
"train": "tasks_1-20_v1-2/hn-10k/qa5_three-arg-relations_train.txt",
},
"qa12": {
"test": "tasks_1-20_v1-2/hn-10k/qa12_conjunction_test.txt",
"train": "tasks_1-20_v1-2/hn-10k/qa12_conjunction_train.txt",
},
"qa2": {
"train": "tasks_1-20_v1-2/hn-10k/qa2_two-supporting-facts_train.txt",
"test": "tasks_1-20_v1-2/hn-10k/qa2_two-supporting-facts_test.txt",
},
"qa20": {
"train": "tasks_1-20_v1-2/hn-10k/qa20_agents-motivations_train.txt",
"test": "tasks_1-20_v1-2/hn-10k/qa20_agents-motivations_test.txt",
},
"qa8": {
"train": "tasks_1-20_v1-2/hn-10k/qa8_lists-sets_train.txt",
"test": "tasks_1-20_v1-2/hn-10k/qa8_lists-sets_test.txt",
},
},
"shuffled": {
"qa9": {
"test": "tasks_1-20_v1-2/shuffled/qa9_simple-negation_test.txt",
"train": "tasks_1-20_v1-2/shuffled/qa9_simple-negation_train.txt",
},
"qa4": {
"train": "tasks_1-20_v1-2/shuffled/qa4_two-arg-relations_train.txt",
"test": "tasks_1-20_v1-2/shuffled/qa4_two-arg-relations_test.txt",
},
"qa6": {
"train": "tasks_1-20_v1-2/shuffled/qa6_yes-no-questions_train.txt",
"test": "tasks_1-20_v1-2/shuffled/qa6_yes-no-questions_test.txt",
},
"qa11": {
"test": "tasks_1-20_v1-2/shuffled/qa11_basic-coreference_test.txt",
"train": "tasks_1-20_v1-2/shuffled/qa11_basic-coreference_train.txt",
},
"qa3": {
"test": "tasks_1-20_v1-2/shuffled/qa3_three-supporting-facts_test.txt",
"train": "tasks_1-20_v1-2/shuffled/qa3_three-supporting-facts_train.txt",
},
"qa15": {
"test": "tasks_1-20_v1-2/shuffled/qa15_basic-deduction_test.txt",
"train": "tasks_1-20_v1-2/shuffled/qa15_basic-deduction_train.txt",
},
"qa17": {
"test": "tasks_1-20_v1-2/shuffled/qa17_positional-reasoning_test.txt",
"train": "tasks_1-20_v1-2/shuffled/qa17_positional-reasoning_train.txt",
},
"qa13": {
"test": "tasks_1-20_v1-2/shuffled/qa13_compound-coreference_test.txt",
"train": "tasks_1-20_v1-2/shuffled/qa13_compound-coreference_train.txt",
},
"qa1": {
"train": "tasks_1-20_v1-2/shuffled/qa1_single-supporting-fact_train.txt",
"test": "tasks_1-20_v1-2/shuffled/qa1_single-supporting-fact_test.txt",
},
"qa14": {
"train": "tasks_1-20_v1-2/shuffled/qa14_time-reasoning_train.txt",
"test": "tasks_1-20_v1-2/shuffled/qa14_time-reasoning_test.txt",
},
"qa16": {
"test": "tasks_1-20_v1-2/shuffled/qa16_basic-induction_test.txt",
"train": "tasks_1-20_v1-2/shuffled/qa16_basic-induction_train.txt",
},
"qa19": {
"test": "tasks_1-20_v1-2/shuffled/qa19_path-finding_test.txt",
"train": "tasks_1-20_v1-2/shuffled/qa19_path-finding_train.txt",
},
"qa18": {
"test": "tasks_1-20_v1-2/shuffled/qa18_size-reasoning_test.txt",
"train": "tasks_1-20_v1-2/shuffled/qa18_size-reasoning_train.txt",
},
"qa10": {
"train": "tasks_1-20_v1-2/shuffled/qa10_indefinite-knowledge_train.txt",
"test": "tasks_1-20_v1-2/shuffled/qa10_indefinite-knowledge_test.txt",
},
"qa7": {
"train": "tasks_1-20_v1-2/shuffled/qa7_counting_train.txt",
"test": "tasks_1-20_v1-2/shuffled/qa7_counting_test.txt",
},
"qa5": {
"test": "tasks_1-20_v1-2/shuffled/qa5_three-arg-relations_test.txt",
"train": "tasks_1-20_v1-2/shuffled/qa5_three-arg-relations_train.txt",
},
"qa12": {
"test": "tasks_1-20_v1-2/shuffled/qa12_conjunction_test.txt",
"train": "tasks_1-20_v1-2/shuffled/qa12_conjunction_train.txt",
},
"qa2": {
"train": "tasks_1-20_v1-2/shuffled/qa2_two-supporting-facts_train.txt",
"test": "tasks_1-20_v1-2/shuffled/qa2_two-supporting-facts_test.txt",
},
"qa20": {
"train": "tasks_1-20_v1-2/shuffled/qa20_agents-motivations_train.txt",
"test": "tasks_1-20_v1-2/shuffled/qa20_agents-motivations_test.txt",
},
"qa8": {
"train": "tasks_1-20_v1-2/shuffled/qa8_lists-sets_train.txt",
"test": "tasks_1-20_v1-2/shuffled/qa8_lists-sets_test.txt",
},
},
"shuffled-10k": {
"qa9": {
"test": "tasks_1-20_v1-2/shuffled-10k/qa9_simple-negation_test.txt",
"train": "tasks_1-20_v1-2/shuffled-10k/qa9_simple-negation_train.txt",
},
"qa4": {
"train": "tasks_1-20_v1-2/shuffled-10k/qa4_two-arg-relations_train.txt",
"test": "tasks_1-20_v1-2/shuffled-10k/qa4_two-arg-relations_test.txt",
},
"qa6": {
"train": "tasks_1-20_v1-2/shuffled-10k/qa6_yes-no-questions_train.txt",
"test": "tasks_1-20_v1-2/shuffled-10k/qa6_yes-no-questions_test.txt",
},
"qa11": {
"test": "tasks_1-20_v1-2/shuffled-10k/qa11_basic-coreference_test.txt",
"train": "tasks_1-20_v1-2/shuffled-10k/qa11_basic-coreference_train.txt",
},
"qa3": {
"test": "tasks_1-20_v1-2/shuffled-10k/qa3_three-supporting-facts_test.txt",
"train": "tasks_1-20_v1-2/shuffled-10k/qa3_three-supporting-facts_train.txt",
},
"qa15": {
"test": "tasks_1-20_v1-2/shuffled-10k/qa15_basic-deduction_test.txt",
"train": "tasks_1-20_v1-2/shuffled-10k/qa15_basic-deduction_train.txt",
},
"qa17": {
"test": "tasks_1-20_v1-2/shuffled-10k/qa17_positional-reasoning_test.txt",
"train": "tasks_1-20_v1-2/shuffled-10k/qa17_positional-reasoning_train.txt",
},
"qa13": {
"test": "tasks_1-20_v1-2/shuffled-10k/qa13_compound-coreference_test.txt",
"train": "tasks_1-20_v1-2/shuffled-10k/qa13_compound-coreference_train.txt",
},
"qa1": {
"train": "tasks_1-20_v1-2/shuffled-10k/qa1_single-supporting-fact_train.txt",
"test": "tasks_1-20_v1-2/shuffled-10k/qa1_single-supporting-fact_test.txt",
},
"qa14": {
"train": "tasks_1-20_v1-2/shuffled-10k/qa14_time-reasoning_train.txt",
"test": "tasks_1-20_v1-2/shuffled-10k/qa14_time-reasoning_test.txt",
},
"qa16": {
"test": "tasks_1-20_v1-2/shuffled-10k/qa16_basic-induction_test.txt",
"train": "tasks_1-20_v1-2/shuffled-10k/qa16_basic-induction_train.txt",
},
"qa19": {
"test": "tasks_1-20_v1-2/shuffled-10k/qa19_path-finding_test.txt",
"train": "tasks_1-20_v1-2/shuffled-10k/qa19_path-finding_train.txt",
},
"qa18": {
"test": "tasks_1-20_v1-2/shuffled-10k/qa18_size-reasoning_test.txt",
"train": "tasks_1-20_v1-2/shuffled-10k/qa18_size-reasoning_train.txt",
},
"qa10": {
"train": "tasks_1-20_v1-2/shuffled-10k/qa10_indefinite-knowledge_train.txt",
"test": "tasks_1-20_v1-2/shuffled-10k/qa10_indefinite-knowledge_test.txt",
},
"qa7": {
"train": "tasks_1-20_v1-2/shuffled-10k/qa7_counting_train.txt",
"test": "tasks_1-20_v1-2/shuffled-10k/qa7_counting_test.txt",
},
"qa5": {
"test": "tasks_1-20_v1-2/shuffled-10k/qa5_three-arg-relations_test.txt",
"train": "tasks_1-20_v1-2/shuffled-10k/qa5_three-arg-relations_train.txt",
},
"qa12": {
"test": "tasks_1-20_v1-2/shuffled-10k/qa12_conjunction_test.txt",
"train": "tasks_1-20_v1-2/shuffled-10k/qa12_conjunction_train.txt",
},
"qa2": {
"train": "tasks_1-20_v1-2/shuffled-10k/qa2_two-supporting-facts_train.txt",
"test": "tasks_1-20_v1-2/shuffled-10k/qa2_two-supporting-facts_test.txt",
},
"qa20": {
"train": "tasks_1-20_v1-2/shuffled-10k/qa20_agents-motivations_train.txt",
"test": "tasks_1-20_v1-2/shuffled-10k/qa20_agents-motivations_test.txt",
},
"qa8": {
"train": "tasks_1-20_v1-2/shuffled-10k/qa8_lists-sets_train.txt",
"test": "tasks_1-20_v1-2/shuffled-10k/qa8_lists-sets_test.txt",
},
},
}
class BabiQaConfig(datasets.BuilderConfig):
def __init__(self, *args, type=None, task_no=None, **kwargs):
super().__init__(
*args,
name=f"{type}-{task_no}",
**kwargs,
)
self.type = type
self.task_no = task_no
class BabiQa(datasets.GeneratorBasedBuilder):
"""The bAbI QA (20) tasks Dataset"""
VERSION = datasets.Version("1.2.0")
BUILDER_CONFIG_CLASS = BabiQaConfig
BUILDER_CONFIGS = [
BabiQaConfig(
type="en",
task_no="qa1",
version=VERSION,
description="This part of the config handles the `qa1` task of the bAbI `en` dataset",
),
BabiQaConfig(
type="hn",
task_no="qa1",
version=VERSION,
description="This part of the config handles the `qa1` task of the bAbI `hn` dataset",
),
BabiQaConfig(
type="en-10k",
task_no="qa1",
version=VERSION,
description="This part of the config handles the `qa1` task of the bAbI `en-10k` dataset",
),
BabiQaConfig(
type="en-valid",
task_no="qa1",
version=VERSION,
description="This part of the config handles the `qa1` task of the bAbI `en-valid` dataset",
),
BabiQaConfig(
type="en-valid-10k",
task_no="qa1",
version=VERSION,
description="This part of the config handles the `qa1` task of the bAbI `en-valid-10k` dataset",
),
BabiQaConfig(
type="hn-10k",
task_no="qa1",
version=VERSION,
description="This part of the config handles the `qa1` task of the bAbI `hn-10k` dataset",
),
BabiQaConfig(
type="shuffled",
task_no="qa1",
version=VERSION,
description="This part of the config handles the `qa1` task of the bAbI `shuffled` dataset",
),
BabiQaConfig(
type="shuffled-10k",
task_no="qa1",
version=VERSION,
description="This part of the config handles the `qa1` task of the bAbI `shuffled-10k` dataset",
),
]
def _info(self):
features = datasets.Features(
{
"story": datasets.Sequence(
{
"id": datasets.Value("string"),
"type": datasets.ClassLabel(names=["context", "question"]),
"text": datasets.Value("string"),
"supporting_ids": datasets.Sequence(datasets.Value("string")),
"answer": datasets.Value("string"),
}
),
}
)
return datasets.DatasetInfo(
# This is the description that will appear on the datasets page.
description=_DESCRIPTION,
# This defines the different columns of the dataset and their types
features=features, # Here we define them above because they are different between the two configurations
# If there's a common (input, target) tuple from the features,
# specify them here. They'll be used if as_supervised=True in
# builder.as_dataset.
supervised_keys=None,
# Homepage of the dataset for documentation
homepage=_HOMEPAGE,
# License for the dataset if available
license=_LICENSE,
# Citation for the dataset
citation=_CITATION,
)
def _split_generators(self, dl_manager):
"""Returns SplitGenerators."""
my_urls = ZIP_URL
archive = dl_manager.download(my_urls)
splits = [
datasets.SplitGenerator(
name=datasets.Split.TRAIN,
# These kwargs will be passed to _generate_examples
gen_kwargs={
"filepath": paths[self.config.type][self.config.task_no]["train"],
"files": dl_manager.iter_archive(archive),
},
),
datasets.SplitGenerator(
name=datasets.Split.TEST,
# These kwargs will be passed to _generate_examples
gen_kwargs={
"filepath": paths[self.config.type][self.config.task_no]["test"],
"files": dl_manager.iter_archive(archive),
},
),
]
if "valid" in self.config.type:
splits += [
datasets.SplitGenerator(
name=datasets.Split.VALIDATION,
# These kwargs will be passed to _generate_examples
gen_kwargs={
"filepath": paths[self.config.type][self.config.task_no]["valid"],
"files": dl_manager.iter_archive(archive),
},
),
]
return splits
def _generate_examples(self, filepath, files):
for path, f in files:
if path == filepath:
story = []
example_idx = 0
for idx, line in enumerate(f):
line = line.decode("utf-8")
if line.strip() == "":
if story != []:
yield example_idx, {"story": story}
example_idx += 1
story = []
elif line.strip().split()[0] == "1": # New story
if story != []: # Already some story, flush it out
yield example_idx, {"story": story}
example_idx += 1
story = []
line_no = line.split()[0]
line_split = line[len(line_no) :].strip().split("\t")
if len(line_split) > 1:
story.append(
{
"id": line_no,
"type": 1, # question
"supporting_ids": line_split[-1].split(" "),
"text": line_split[0].strip(),
"answer": line_split[1].strip(),
}
)
else:
story.append(
{
"id": line_no,
"type": 0, # context
"supporting_ids": [],
"text": line_split[0].strip(),
"answer": "",
}
)
else:
line_no = line.split()[0]
line_split = line[len(line_no) :].strip().split("\t")
if len(line_split) > 1:
story.append(
{
"id": line_no,
"type": 1, # question
"supporting_ids": line_split[-1].split(" "),
"text": line_split[0].strip(),
"answer": line_split[1].strip(),
}
)
else:
story.append(
{
"id": line_no,
"type": 0, # context
"supporting_ids": [],
"text": line_split[0].strip(),
"answer": "",
}
)
else: # After last line
if story != []:
yield example_idx, {"story": story}
break