Upload data files directly to the HF repository (#3)

- Upload data files directly to the HF repository (0a46891173b3eacf17ac7b1e3a5c479c5549cbb7)
- Delete legacy dataset_infos.json (069c6fed2da1f047c3f1e4faa7bb439631b096f7)
- Update metadata in dataset card (d9f69d76cd901db660a03b8099072d5b3de89d51)


Co-authored-by: Pavel Efimov <pefimov@users.noreply.huggingface.co>

README.md

@@ -1,5 +1,4 @@
 ---
-pretty_name: SberQuAD
 annotations_creators:
 - crowdsourced
 language_creators:
@@ -20,7 +19,9 @@ task_categories:
 task_ids:
 - extractive-qa
 paperswithcode_id: sberquad
+pretty_name: SberQuAD
 dataset_info:
+  config_name: sberquad
   features:
   - name: id
     dtype: int32
@@ -36,19 +37,18 @@ dataset_info:
       dtype: string
     - name: answer_start
       dtype: int32
-  config_name: sberquad
   splits:
   - name: train
-    num_bytes: 71631661
+    num_bytes: 71631541
     num_examples: 45328
   - name: validation
-    num_bytes: 7972977
+    num_bytes: 7972953
     num_examples: 5036
   - name: test
-    num_bytes: 36397848
+    num_bytes: 36397776
     num_examples: 23936
-  download_size: 66047276
-  dataset_size: 116002486
+  download_size: 10491714
+  dataset_size: 116002270
 ---
 
 

data/dev_v1.0.json.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9145c18e890e84fd070272fcaa607a688efd61debf36b89bc2b96654ac77d18
+size 1927670

data/origin_test.json.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d24ee67e71ea95163342b72d8fb36914fa7d0b261b3552ed5824f355fe76963
+size 2725921

data/train_v1.0.json.gz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a05221fe35eb1b4fcd17b8acca1d957e12d098783d8e8a2ebb2348f6c437b0a
+size 5838123

dataset_infos.json

@@ -1 +0,0 @@
-{"sberquad": {"description": "Sber Question Answering Dataset (SberQuAD) is a reading comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles, where the answer to every question is a segment of text, or span, from the corresponding reading passage, or the question might be unanswerable. Russian original analogue presented in Sberbank Data Science Journey 2017.\n", "citation": "@article{Efimov_2020,\n   title={SberQuAD \u2013 Russian Reading Comprehension Dataset: Description and Analysis},\n   ISBN={9783030582197},\n   ISSN={1611-3349},\n   url={http://dx.doi.org/10.1007/978-3-030-58219-7_1},\n   DOI={10.1007/978-3-030-58219-7_1},\n   journal={Experimental IR Meets Multilinguality, Multimodality, and Interaction},\n   publisher={Springer International Publishing},\n   author={Efimov, Pavel and Chertok, Andrey and Boytsov, Leonid and Braslavski, Pavel},\n   year={2020},\n   pages={3\u201315}\n}\n ", "homepage": "", "license": "", "features": {"id": {"dtype": "int32", "id": null, "_type": "Value"}, "title": {"dtype": "string", "id": null, "_type": "Value"}, "context": {"dtype": "string", "id": null, "_type": "Value"}, "question": {"dtype": "string", "id": null, "_type": "Value"}, "answers": {"feature": {"text": {"dtype": "string", "id": null, "_type": "Value"}, "answer_start": {"dtype": "int32", "id": null, "_type": "Value"}}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": [{"task": "question-answering-extractive", "question_column": "question", "context_column": "context", "answers_column": "answers"}], "builder_name": "sberquad", "config_name": "sberquad", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 71631661, "num_examples": 45328, "dataset_name": "sberquad"}, "validation": {"name": "validation", "num_bytes": 7972977, "num_examples": 5036, "dataset_name": "sberquad"}, "test": {"name": "test", "num_bytes": 36397848, "num_examples": 23936, "dataset_name": "sberquad"}}, "download_checksums": {"https://sc.link/PNWl": {"num_bytes": 38616884, "checksum": "861b55219f1549139e64b2eed325b54ce9c9c63b792a2c2b3cfbec997aa3d88e"}, "https://sc.link/W6oX": {"num_bytes": 8807953, "checksum": "247bede36a27f076f607117632f39eedb9bb1d80c34d93bbfaeda71fd30fd382"}, "https://sc.link/VOn9": {"num_bytes": 18622439, "checksum": "7793d389208271a76ab38a5dba5cebc98e72f45e99196a99e14b5a37c401c66f"}}, "download_size": 66047276, "post_processing_size": null, "dataset_size": 116002486, "size_in_bytes": 182049762}}

sberquad.py

@@ -1,6 +1,7 @@
 # coding=utf-8
 """SberQUAD: Sber Question Answering Dataset."""
 
+import os
 import json
 
 import datasets
@@ -33,7 +34,7 @@ from the corresponding reading passage, or the question might be unanswerable. \
 Russian original analogue presented in Sberbank Data Science Journey 2017.
 """
 
-_URLS = {"train": "https://sc.link/PNWl", "dev": "https://sc.link/W6oX", "test": "https://sc.link/VOn9"}
+_URLS = {"train": os.path.join("data", "train_v1.0.json.gz"), "dev": os.path.join("data", "dev_v1.0.json.gz"), "test": os.path.join("data", "origin_test.json.gz")}
 
 
 class Sberquad(datasets.GeneratorBasedBuilder):