Update files from the datasets library (from 1.4.0)

Release notes: https://github.com/huggingface/datasets/releases/tag/1.4.0

README.md

@@ -1,4 +1,26 @@
 ---
+annotations_creators:
+- crowdsourced
+language_creators:
+- crowdsourced
+- found
+languages:
+- en
+licenses:
+- cc-by-3-0
+- cc-by-sa-3-0-at
+- mit
+- other-Open Portion of the American National Corpus
+multilinguality:
+- monolingual
+size_categories:
+- 100K<n<1M
+source_datasets:
+- original
+task_categories:
+- text-scoring
+task_ids:
+- semantic-similarity-scoring
 ---
 
 # Dataset Card for "multi_nli"
@@ -27,7 +49,7 @@
   - [Citation Information](#citation-information)
   - [Contributions](#contributions)
 
-## [Dataset Description](#dataset-description)
+## Dataset Description
 
 - **Homepage:** [https://www.nyu.edu/projects/bowman/multinli/](https://www.nyu.edu/projects/bowman/multinli/)
 - **Repository:** [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
@@ -37,7 +59,7 @@
 - **Size of the generated dataset:** 73.39 MB
 - **Total amount of disk used:** 289.74 MB
 
-### [Dataset Summary](#dataset-summary)
+### Dataset Summary
 
 The Multi-Genre Natural Language Inference (MultiNLI) corpus is a
 crowd-sourced collection of 433k sentence pairs annotated with textual
@@ -46,93 +68,100 @@ that covers a range of genres of spoken and written text, and supports a
 distinctive cross-genre generalization evaluation. The corpus served as the
 basis for the shared task of the RepEval 2017 Workshop at EMNLP in Copenhagen.
 
-### [Supported Tasks](#supported-tasks)
+### Supported Tasks
 
 [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
 
-### [Languages](#languages)
+### Languages
 
-[More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
-
-## [Dataset Structure](#dataset-structure)
+The dataset contains samples in English only.
 
-We show detailed information for up to 5 configurations of the dataset.
+## Dataset Structure
 
-### [Data Instances](#data-instances)
-
-#### plain_text
+### Data Instances
 
 - **Size of downloaded dataset files:** 216.34 MB
 - **Size of the generated dataset:** 73.39 MB
 - **Total amount of disk used:** 289.74 MB
 
-An example of 'validation_matched' looks as follows.
+Example of a data instance:
+
 ```
 {
-    "hypothesis": "flammable",
-    "label": 0,
-    "premise": "inflammable"
+    "promptID": 31193,
+    "pairID": "31193n",
+    "premise": "Conceptually cream skimming has two basic dimensions - product and geography.",
+    "premise_binary_parse": "( ( Conceptually ( cream skimming ) ) ( ( has ( ( ( two ( basic dimensions ) ) - ) ( ( product and ) geography ) ) ) . ) )",
+    "premise_parse": "(ROOT (S (NP (JJ Conceptually) (NN cream) (NN skimming)) (VP (VBZ has) (NP (NP (CD two) (JJ basic) (NNS dimensions)) (: -) (NP (NN product) (CC and) (NN geography)))) (. .)))",
+    "hypothesis": "Product and geography are what make cream skimming work. ",
+    "hypothesis_binary_parse": "( ( ( Product and ) geography ) ( ( are ( what ( make ( cream ( skimming work ) ) ) ) ) . ) )",
+    "hypothesis_parse": "(ROOT (S (NP (NN Product) (CC and) (NN geography)) (VP (VBP are) (SBAR (WHNP (WP what)) (S (VP (VBP make) (NP (NP (NN cream)) (VP (VBG skimming) (NP (NN work)))))))) (. .)))",
+    "genre": "government",
+    "label": 1
 }
 ```
 
-### [Data Fields](#data-fields)
+### Data Fields
 
 The data fields are the same among all splits.
 
-#### plain_text
-- `premise`: a `string` feature.
-- `hypothesis`: a `string` feature.
-- `label`: a classification label, with possible values including `entailment` (0), `neutral` (1), `contradiction` (2).
+- `promptID`: Unique identifier for prompt
+- `pairID`: Unique identifier for pair
+- `{premise,hypothesis}`: combination of `premise` and `hypothesis`
+- `{premise,hypothesis} parse`: Each sentence as parsed by the Stanford PCFG Parser 3.5.2
+- `{premise,hypothesis} binary parse`: parses in unlabeled binary-branching format
+- `genre`: a `string` feature.
+- `label`: a classification label, with possible values including `entailment` (0), `neutral` (1), `contradiction` (2)
 
-### [Data Splits Sample Size](#data-splits-sample-size)
+### Data Splits Sample Size
 
-|   name   |train |validation_matched|validation_mismatched|
-|----------|-----:|-----------------:|--------------------:|
-|plain_text|392702|              9815|                 9832|
+|train |validation_matched|validation_mismatched|
+|-----:|-----------------:|--------------------:|
+|392702|              9815|                 9832|
 
-## [Dataset Creation](#dataset-creation)
+## Dataset Creation
 
-### [Curation Rationale](#curation-rationale)
+### Curation Rationale
 
-[More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
+They constructed MultiNLI so as to make it possible to explicitly evaluate models both on the quality of their sentence representations within the training domain and on their ability to derive reasonable representations in unfamiliar domains.
 
-### [Source Data](#source-data)
+### Source Data
 
-[More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
+They created each sentence pair by selecting a premise sentence from a preexisting text source and asked a human annotator to compose a novel sentence to pair with it as a hypothesis.
 
-### [Annotations](#annotations)
+### Annotations
 
 [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
 
-### [Personal and Sensitive Information](#personal-and-sensitive-information)
+### Personal and Sensitive Information
 
-[More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
+[More Information Needed]
 
-## [Considerations for Using the Data](#considerations-for-using-the-data)
+## Considerations for Using the Data
 
-### [Social Impact of Dataset](#social-impact-of-dataset)
+### Social Impact of Dataset
 
-[More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
+[More Information Needed]
 
-### [Discussion of Biases](#discussion-of-biases)
+### Discussion of Biases
 
-[More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
+[More Information Needed]
 
-### [Other Known Limitations](#other-known-limitations)
+### Other Known Limitations
 
-[More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
+[More Information Needed]
 
-## [Additional Information](#additional-information)
+## Additional Information
 
-### [Dataset Curators](#dataset-curators)
+### Dataset Curators
 
 [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
 
-### [Licensing Information](#licensing-information)
+### Licensing Information
 
-[More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
+The majority of the corpus is released under the OANC’s license, which allows all content to be freely used, modi- fied, and shared under permissive terms. The data in the FICTION section falls under several per- missive licenses; Seven Swords is available under a Creative Commons Share-Alike 3.0 Unported License, and with the explicit permission of the author, Living History and Password Incorrect are available under Creative Commons Attribution 3.0 Unported Licenses; the remaining works of fiction are in the public domain in the United States (but may be licensed differently elsewhere).
 
-### [Citation Information](#citation-information)
+### Citation Information
 
 ```
 @InProceedings{N18-1101,
@@ -152,10 +181,8 @@ The data fields are the same among all splits.
   location = "New Orleans, Louisiana",
   url = "http://aclweb.org/anthology/N18-1101"
 }
-
 ```
 
-
 ### Contributions
 
-Thanks to [@patrickvonplaten](https://github.com/patrickvonplaten), [@thomwolf](https://github.com/thomwolf), [@mariamabarham](https://github.com/mariamabarham) for adding this dataset.
+Thanks to [@bhavitvyamalik](https://github.com/bhavitvyamalik), [@patrickvonplaten](https://github.com/patrickvonplaten), [@thomwolf](https://github.com/thomwolf), [@mariamabarham](https://github.com/mariamabarham) for adding this dataset.

dataset_infos.json

@@ -1 +1 @@
-{"plain_text": {"description": "The Multi-Genre Natural Language Inference (MultiNLI) corpus is a\ncrowd-sourced collection of 433k sentence pairs annotated with textual\nentailment information. The corpus is modeled on the SNLI corpus, but differs in\nthat covers a range of genres of spoken and written text, and supports a\ndistinctive cross-genre generalization evaluation. The corpus served as the\nbasis for the shared task of the RepEval 2017 Workshop at EMNLP in Copenhagen.\n", "citation": "@InProceedings{N18-1101,\n  author = \"Williams, Adina\n            and Nangia, Nikita\n            and Bowman, Samuel\",\n  title = \"A Broad-Coverage Challenge Corpus for\n           Sentence Understanding through Inference\",\n  booktitle = \"Proceedings of the 2018 Conference of\n               the North American Chapter of the\n               Association for Computational Linguistics:\n               Human Language Technologies, Volume 1 (Long\n               Papers)\",\n  year = \"2018\",\n  publisher = \"Association for Computational Linguistics\",\n  pages = \"1112--1122\",\n  location = \"New Orleans, Louisiana\",\n  url = \"http://aclweb.org/anthology/N18-1101\"\n}\n", "homepage": "https://www.nyu.edu/projects/bowman/multinli/", "license": "", "features": {"premise": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "supervised_keys": null, "builder_name": "multi_nli", "config_name": "plain_text", "version": {"version_str": "1.0.0", "description": "", "datasets_version_to_prepare": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 73245222, "num_examples": 392702, "dataset_name": "multi_nli"}, "validation_matched": {"name": "validation_matched", "num_bytes": 1799439, "num_examples": 9815, "dataset_name": "multi_nli"}, "validation_mismatched": {"name": "validation_mismatched", "num_bytes": 1914827, "num_examples": 9832, "dataset_name": "multi_nli"}}, "download_checksums": {"http://storage.googleapis.com/tfds-data/downloads/multi_nli/multinli_1.0.zip": {"num_bytes": 226850426, "checksum": "049f507b9e36b1fcb756cfd5aeb3b7a0cfcb84bf023793652987f7e7e0957822"}}, "download_size": 226850426, "dataset_size": 76959488, "size_in_bytes": 303809914}}
+{"default": {"description": "The Multi-Genre Natural Language Inference (MultiNLI) corpus is a\ncrowd-sourced collection of 433k sentence pairs annotated with textual\nentailment information. The corpus is modeled on the SNLI corpus, but differs in\nthat covers a range of genres of spoken and written text, and supports a\ndistinctive cross-genre generalization evaluation. The corpus served as the\nbasis for the shared task of the RepEval 2017 Workshop at EMNLP in Copenhagen.\n", "citation": "@InProceedings{N18-1101,\n  author = {Williams, Adina\n            and Nangia, Nikita\n            and Bowman, Samuel},\n  title = {A Broad-Coverage Challenge Corpus for\n           Sentence Understanding through Inference},\n  booktitle = {Proceedings of the 2018 Conference of\n               the North American Chapter of the\n               Association for Computational Linguistics:\n               Human Language Technologies, Volume 1 (Long\n               Papers)},\n  year = {2018},\n  publisher = {Association for Computational Linguistics},\n  pages = {1112--1122},\n  location = {New Orleans, Louisiana},\n  url = {http://aclweb.org/anthology/N18-1101}\n}\n", "homepage": "https://www.nyu.edu/projects/bowman/multinli/", "license": "", "features": {"promptID": {"dtype": "int32", "id": null, "_type": "Value"}, "pairID": {"dtype": "string", "id": null, "_type": "Value"}, "premise": {"dtype": "string", "id": null, "_type": "Value"}, "premise_binary_parse": {"dtype": "string", "id": null, "_type": "Value"}, "premise_parse": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis_binary_parse": {"dtype": "string", "id": null, "_type": "Value"}, "hypothesis_parse": {"dtype": "string", "id": null, "_type": "Value"}, "genre": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"num_classes": 3, "names": ["entailment", "neutral", "contradiction"], "names_file": null, "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "builder_name": "multi_nli", "config_name": "default", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 410211586, "num_examples": 392702, "dataset_name": "multi_nli"}, "validation_matched": {"name": "validation_matched", "num_bytes": 10063939, "num_examples": 9815, "dataset_name": "multi_nli"}, "validation_mismatched": {"name": "validation_mismatched", "num_bytes": 10610221, "num_examples": 9832, "dataset_name": "multi_nli"}}, "download_checksums": {"https://cims.nyu.edu/~sbowman/multinli/multinli_1.0.zip": {"num_bytes": 226850426, "checksum": "049f507b9e36b1fcb756cfd5aeb3b7a0cfcb84bf023793652987f7e7e0957822"}}, "download_size": 226850426, "post_processing_size": null, "dataset_size": 430885746, "size_in_bytes": 657736172}}

dummy/{plain_text/1.0.0 → 0.0.0}/dummy_data.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1fa7aca81ea7db5408b84d967c291a89f721e82e0b4eef3563e48ec8edf347e8
-size 1276
+oid sha256:befcc541017d4438bd7105595291e924e6933f3e330c2a41e0053caea457436d
+size 13205

multi_nli.py

@@ -18,6 +18,7 @@
 
 from __future__ import absolute_import, division, print_function
 
+import json
 import os
 
 import datasets
@@ -53,36 +54,25 @@ basis for the shared task of the RepEval 2017 Workshop at EMNLP in Copenhagen.
 """
 
 
-class MultiNLIConfig(datasets.BuilderConfig):
-    """BuilderConfig for MultiNLI."""
-
-    def __init__(self, **kwargs):
-        """BuilderConfig for MultiNLI.
-
-            Args:
-        .
-              **kwargs: keyword arguments forwarded to super.
-        """
-        super(MultiNLIConfig, self).__init__(version=datasets.Version("1.0.0", ""), **kwargs)
-
-
 class MultiNli(datasets.GeneratorBasedBuilder):
     """MultiNLI: The Stanford Question Answering Dataset. Version 1.1."""
 
-    BUILDER_CONFIGS = [
-        MultiNLIConfig(
-            name="plain_text",
-            description="Plain text",
-        ),
-    ]
-
     def _info(self):
         return datasets.DatasetInfo(
             description=_DESCRIPTION,
             features=datasets.Features(
                 {
+                    "promptID": datasets.Value("int32"),
+                    "pairID": datasets.Value("string"),
                     "premise": datasets.Value("string"),
+                    "premise_binary_parse": datasets.Value("string"),  # parses in unlabeled binary-branching format
+                    "premise_parse": datasets.Value("string"),  # sentence as parsed by the Stanford PCFG Parser 3.5.2
                     "hypothesis": datasets.Value("string"),
+                    "hypothesis_binary_parse": datasets.Value("string"),  # parses in unlabeled binary-branching format
+                    "hypothesis_parse": datasets.Value(
+                        "string"
+                    ),  # sentence as parsed by the Stanford PCFG Parser 3.5.2
+                    "genre": datasets.Value("string"),
                     "label": datasets.features.ClassLabel(names=["entailment", "neutral", "contradiction"]),
                 }
             ),
@@ -93,19 +83,13 @@ class MultiNli(datasets.GeneratorBasedBuilder):
             citation=_CITATION,
         )
 
-    def _vocab_text_gen(self, filepath):
-        for _, ex in self._generate_examples(filepath):
-            yield " ".join([ex["premise"], ex["hypothesis"]])
-
     def _split_generators(self, dl_manager):
 
-        downloaded_dir = dl_manager.download_and_extract(
-            "http://storage.googleapis.com/tfds-data/downloads/multi_nli/multinli_1.0.zip"
-        )
+        downloaded_dir = dl_manager.download_and_extract("https://cims.nyu.edu/~sbowman/multinli/multinli_1.0.zip")
         mnli_path = os.path.join(downloaded_dir, "multinli_1.0")
-        train_path = os.path.join(mnli_path, "multinli_1.0_train.txt")
-        matched_validation_path = os.path.join(mnli_path, "multinli_1.0_dev_matched.txt")
-        mismatched_validation_path = os.path.join(mnli_path, "multinli_1.0_dev_mismatched.txt")
+        train_path = os.path.join(mnli_path, "multinli_1.0_train.jsonl")
+        matched_validation_path = os.path.join(mnli_path, "multinli_1.0_dev_matched.jsonl")
+        mismatched_validation_path = os.path.join(mnli_path, "multinli_1.0_dev_mismatched.jsonl")
 
         return [
             datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepath": train_path}),
@@ -114,22 +98,22 @@ class MultiNli(datasets.GeneratorBasedBuilder):
         ]
 
     def _generate_examples(self, filepath):
-        """Generate mnli examples.
-
-        Args:
-          filepath: a string
-
-        Yields:
-          dictionaries containing "premise", "hypothesis" and "label" strings
-        """
-        for idx, line in enumerate(open(filepath, "rb")):
-            if idx == 0:
-                continue  # skip header
-            line = line.strip().decode("utf-8")
-            split_line = line.split("\t")
-            # Examples not marked with a three out of five consensus are marked with
-            # "-" and should not be used in standard evaluations.
-            if split_line[0] == "-":
-                continue
-            # Works for both splits even though dev has some extra human labels.
-            yield idx, {"premise": split_line[5], "hypothesis": split_line[6], "label": split_line[0]}
+        """Generate mnli examples"""
+
+        with open(filepath, encoding="utf-8") as f:
+            for id_, row in enumerate(f):
+                data = json.loads(row)
+                if data["gold_label"] == "-":
+                    continue
+                yield id_, {
+                    "promptID": data["promptID"],
+                    "pairID": data["pairID"],
+                    "premise": data["sentence1"],
+                    "premise_binary_parse": data["sentence1_binary_parse"],
+                    "premise_parse": data["sentence1_parse"],
+                    "hypothesis": data["sentence2"],
+                    "hypothesis_binary_parse": data["sentence2_binary_parse"],
+                    "hypothesis_parse": data["sentence2_parse"],
+                    "genre": data["genre"],
+                    "label": data["gold_label"],
+                }