Datasets:

Modalities:
Text
Formats:
parquet
Languages:
English
ArXiv:
Libraries:
Datasets
pandas
License:
system HF staff commited on
Commit
2f7c340
1 Parent(s): 2717994

Update files from the datasets library (from 1.8.0)

Browse files

Release notes: https://github.com/huggingface/datasets/releases/tag/1.8.0

Files changed (2) hide show
  1. README.md +4 -2
  2. adversarial_qa.py +5 -2
README.md CHANGED
@@ -139,7 +139,9 @@ Data is provided in the same format as SQuAD 1.1. An example is shown below:
139
  - title: the title of the Wikipedia page from which the context is sourced
140
  - context: the context/passage
141
  - id: a string identifier for each question
142
- - answers: a list of all provided answers (one per question in our case, but multiple may exist in SQuAD) with an `answer_start` field which is the character index of the start of the answer span, and a `text` field which is the answer text
 
 
143
 
144
  ### Data Splits
145
 
@@ -223,4 +225,4 @@ This dataset is distributed under [CC BY-SA 3.0](https://creativecommons.org/lic
223
  ```
224
  ### Contributions
225
 
226
- Thanks to [@maxbartolo](https://github.com/maxbartolo) for adding this dataset.
 
139
  - title: the title of the Wikipedia page from which the context is sourced
140
  - context: the context/passage
141
  - id: a string identifier for each question
142
+ - answers: a list of all provided answers (one per question in our case, but multiple may exist in SQuAD) with an `answer_start` field which is the character index of the start of the answer span, and a `text` field which is the answer text.
143
+
144
+ Note that no answers are provided in the test set. Indeed, this dataset is part of the DynaBench benchmark, for which you can submit your predictions on the [website](https://dynabench.org/tasks/2#1).
145
 
146
  ### Data Splits
147
 
 
225
  ```
226
  ### Contributions
227
 
228
+ Thanks to [@maxbartolo](https://github.com/maxbartolo) for adding this dataset.
adversarial_qa.py CHANGED
@@ -169,13 +169,14 @@ class AdversarialQA(datasets.GeneratorBasedBuilder):
169
  logger.info("generating examples from = %s", filepath)
170
  with open(filepath, encoding="utf-8") as f:
171
  squad = json.load(f)
 
172
  for article in squad["data"]:
173
  title = article.get("title", "").strip()
174
  for paragraph in article["paragraphs"]:
175
  context = paragraph["context"].strip()
176
  for qa in paragraph["qas"]:
177
  question = qa["question"].strip()
178
- id_ = qa["id"]
179
 
180
  answer_starts = [answer["answer_start"] for answer in qa["answers"]]
181
  answers = [answer["text"].strip() for answer in qa["answers"]]
@@ -188,10 +189,12 @@ class AdversarialQA(datasets.GeneratorBasedBuilder):
188
  "title": title,
189
  "context": context,
190
  "question": question,
191
- "id": id_,
192
  "answers": {
193
  "answer_start": answer_starts,
194
  "text": answers,
195
  },
196
  "metadata": {"split": split, "model_in_the_loop": model_in_the_loop},
197
  }
 
 
 
169
  logger.info("generating examples from = %s", filepath)
170
  with open(filepath, encoding="utf-8") as f:
171
  squad = json.load(f)
172
+ id_ = 0
173
  for article in squad["data"]:
174
  title = article.get("title", "").strip()
175
  for paragraph in article["paragraphs"]:
176
  context = paragraph["context"].strip()
177
  for qa in paragraph["qas"]:
178
  question = qa["question"].strip()
179
+ qid = qa["id"]
180
 
181
  answer_starts = [answer["answer_start"] for answer in qa["answers"]]
182
  answers = [answer["text"].strip() for answer in qa["answers"]]
 
189
  "title": title,
190
  "context": context,
191
  "question": question,
192
+ "id": qid,
193
  "answers": {
194
  "answer_start": answer_starts,
195
  "text": answers,
196
  },
197
  "metadata": {"split": split, "model_in_the_loop": model_in_the_loop},
198
  }
199
+
200
+ id_ += 1