FlyPig23 commited on
Commit
26616d3
·
verified ·
1 Parent(s): c2a17cf

Upload batch 216 (20 files, last=huggingface_dataset/Dataset_Card/MicPie_unpredictable_cluster20.md)

Browse files
Files changed (20) hide show
  1. huggingface_dataset/Dataset_Card/JeremyAlain_123_test.md +196 -0
  2. huggingface_dataset/Dataset_Card/Lo_clip-bert-data.md +16 -0
  3. huggingface_dataset/Dataset_Card/MicPie_unpredictable_cluster13.md +250 -0
  4. huggingface_dataset/Dataset_Card/MicPie_unpredictable_cluster20.md +250 -0
  5. huggingface_dataset/Dataset_Card/Nicky0007_cointelegraph_news_English.md +20 -0
  6. huggingface_dataset/Dataset_Card/autoevaluate_autoeval-eval-anli-plain_text-f2dca1-2066067125.md +34 -0
  7. huggingface_dataset/Dataset_Card/autoevaluate_autoeval-eval-futin__feed-sen_en_-1de085-2240171542.md +34 -0
  8. huggingface_dataset/Dataset_Card/autoevaluate_autoeval-eval-futin__guess-en_3-8ea950-2087767174.md +34 -0
  9. huggingface_dataset/Dataset_Card/autoevaluate_autoeval-staging-eval-project-ab647f27-7704970.md +31 -0
  10. huggingface_dataset/Dataset_Card/bethecloud_golf-courses.md +93 -0
  11. huggingface_dataset/Dataset_Card/casino.md +342 -0
  12. huggingface_dataset/Dataset_Card/huggingartists_billy-talent.md +204 -0
  13. huggingface_dataset/Dataset_Card/huggingnft_cryptoadz-by-gremplin.md +175 -0
  14. huggingface_dataset/Dataset_Card/macavaney_d2q-msmarco-passage-scores-monot5.md +88 -0
  15. huggingface_dataset/Dataset_Card/nlpso_m0_qualitative_analysis_ref_cmbert_io.md +46 -0
  16. huggingface_dataset/Dataset_Card/nlpso_m0_qualitative_analysis_ref_ptrn_cmbert_io.md +46 -0
  17. huggingface_dataset/Dataset_Card/nntadotzip_iuQAchatbot.md +21 -0
  18. huggingface_dataset/Dataset_Card/pietrolesci_robust_nli.md +178 -0
  19. huggingface_dataset/Dataset_Card/polsum.md +762 -0
  20. huggingface_dataset/Dataset_Card/readerbench_ro-fb-offense.md +178 -0
huggingface_dataset/Dataset_Card/JeremyAlain_123_test.md ADDED
@@ -0,0 +1,196 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - no-annotation
4
+ language_creators:
5
+ - found
6
+ language:
7
+ - en
8
+ license:
9
+ - apache-2.0
10
+ multilinguality:
11
+ - monolingual
12
+ pretty_name: Fewshot Table Dataset
13
+ size_categories:
14
+ - 100K<n<1M
15
+ source_datasets: []
16
+ task_categories:
17
+ - multiple-choice
18
+ - question-answering
19
+ - zero-shot-classification
20
+ - text2text-generation
21
+ - table-question-answering
22
+ - text-generation
23
+ - text-classification
24
+ - tabular-classification
25
+ task_ids:
26
+ - multiple-choice-qa
27
+ - extractive-qa
28
+ - open-domain-qa
29
+ - closed-domain-qa
30
+ - closed-book-qa
31
+ - open-book-qa
32
+ - language-modeling
33
+ - multi-class-classification
34
+ - natural-language-inference
35
+ - topic-classification
36
+ - multi-label-classification
37
+ - tabular-multi-class-classification
38
+ - tabular-multi-label-classification
39
+ ---
40
+
41
+
42
+ # Dataset Card for Fewshot Table Dataset
43
+
44
+ ## Table of Contents
45
+ - [Dataset Description](#dataset-description)
46
+ - [Dataset Summary](#dataset-summary)
47
+ - [Supported Tasks](#supported-tasks-and-leaderboards)
48
+ - [Languages](#languages)
49
+ - [Dataset Structure](#dataset-structure)
50
+ - [Data Instances](#data-instances)
51
+ - [Data Fields](#data-instances)
52
+ - [Data Splits](#data-instances)
53
+ - [Dataset Creation](#dataset-creation)
54
+ - [Curation Rationale](#curation-rationale)
55
+ - [Source Data](#source-data)
56
+ - [Annotations](#annotations)
57
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
58
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
59
+ - [Social Impact of Dataset](#social-impact-of-dataset)
60
+ - [Discussion of Biases](#discussion-of-biases)
61
+ - [Other Known Limitations](#other-known-limitations)
62
+ - [Additional Information](#additional-information)
63
+ - [Dataset Curators](#dataset-curators)
64
+ - [Licensing Information](#licensing-information)
65
+ - [Citation Information](#citation-information)
66
+
67
+ ## Dataset Description
68
+
69
+ - **Homepage:** [Needs More Information]
70
+ - **Repository:** https://github.com/JunShern/few-shot-pretraining
71
+ - **Paper:** Paper-Title
72
+ - **Leaderboard:** [Needs More Information]
73
+ - **Point of Contact:** junshern@nyu.edu, perez@nyu.edu
74
+
75
+ ### Dataset Summary
76
+
77
+ The Fewshot Table dataset consists of tables that naturally occur on the web, that are formatted as few-shot tasks for fine-tuning language models to improve their few-shot performance. The dataset consists of approximately 413K tables that are extracted from the [WDC Web Table Corpora](http://webdatacommons.org/webtables/) 2015, which is released under the Apache-2.0 license. The WDC Web Table Corpora "contains vast amounts of HTML tables. [...] The Web Data Commons project extracts relational Web tables from the [Common Crawl](https://commoncrawl.org/), the largest and most up-to-date Web corpus that is currently available to the public."
78
+
79
+ ### Supported Tasks and Leaderboards
80
+
81
+ Since the tables come from the web, the distribution of tasks and topics is very broad. The shape of our dataset is very wide i.e. we have 1000's tasks, while each task has only a few examples, compared to most current NLP datasets which are very deep, i.e. 10s of tasks with many examples. This implies that our dataset covers a broad range of potential tasks, e.g. multiple-choice, question-answering, table-question-answering, text-classification, etc.
82
+
83
+ The intended use of this dataset is to improve few-shot performance by finetuning/pretraining onour dataset.
84
+
85
+ ### Languages
86
+
87
+ English
88
+
89
+ ## Dataset Structure
90
+
91
+ ### Data Instances
92
+
93
+ Each table, i.e. task is represented as a json-lines file and consists of several few-shot examples. Each example is a dictionary containing a field 'task', which identifies the task, followed by an 'input', 'options', and 'output' field. The 'input' field contains several column elements of the same row in the table, while the 'output' field is a target which represents an individual column of the same row. Each task contains several such examples which can be concatenated as a few-shot task. In the case of multiple choice classification, the 'options' field contains the possible classes that a model needs to choose from.
94
+
95
+ There are also additional meta-data fields such as 'pageTitle', 'title', 'outputColName', 'url', 'wdcFile'.
96
+
97
+ ### Data Fields
98
+
99
+ 'task': task identifier
100
+
101
+ 'input': column elements of a specific row in table.
102
+
103
+ 'options': for multiple choice classification, it provides the options to choose from.
104
+
105
+ 'output': target column element of same row as input.
106
+
107
+ 'pageTitle': the title of the page containing the table.
108
+
109
+ 'outputColName': ?? (potentially remove this from data)
110
+
111
+ 'url': url to the website containing the table
112
+
113
+ 'wdcFile': ? (potentially remove this from data)
114
+
115
+ ### Data Splits
116
+
117
+ [Needs More Information]
118
+
119
+ ## Dataset Creation
120
+
121
+ ### Curation Rationale
122
+
123
+ How do we convert tables to few-shot tasks?
124
+ Unlike unstructured text, structured data in the form of tables lends itself easily to the few-shot task format. Given a table where each row is an instance of a similar class and the columns describe the attributes of each instance, we can turn each row into a task example to predict one attribute given the others. When the table has more than one row, we instantly have multiple examples of this task by using each row as a single example, and thus each table becomes a few-shot dataset for a particular task.
125
+
126
+ The few-shot setting in this setting is significant: Tables often do not come with clear instructions for each field, so tasks may be underspecified if prompted in a zero-shot manner, but the intended task becomes clearer when examples are provided. This makes a good two-way match: The few-shot format is a perfect setup for table learning, and tables provide a natural dataset for few-shot training.
127
+
128
+ ### Source Data
129
+
130
+ #### Initial Data Collection and Normalization
131
+
132
+ We downloaded the [WDC Web Table Corpora](http://webdatacommons.org/webtables/) 2015 dataset and focus on relational tables. In the following, we describe the steps we executed to filter the WDC Web Table Corpora and create our task dataset. Given a set of relation tables, we apply defined preprocessing steps to ensure all the tables can be handled consistently. Each table can then spawn one or more tasks using a simple predict-one-column approach. Finally, all tasks produced in this manner undergo simple rule-based checks, i.e. any candidates that do not meet some defined minimum requirements for a well-formed task are rejected. Following this approach, we start with 50 million tables in the initial corpus and produce a longlist of 400K tasks.
133
+
134
+ 1. We select only relational tables.
135
+ 2. We make sure all tables are vertical (horizontal tables are simply transposed) and remove duplicate rows.
136
+ 3. To create task we use what in the literature is referred to as verbalizers. For example, a table with 3 columns may be cast as three different tasks: predict column A given B and C, predict column B given A and C, and predict column C given A and B.
137
+ 4. Rule-based-checks to reject tables:
138
+ a) We reject 25M tables that have fewer than 6 rows (so we can do at least k=5-shot learning)
139
+ b) We reject tables with > 20% non-English text as measured by [SpaCy](https://spacy.io/)
140
+ c) Given 2 Million passing tables we consider each table column as a potential output column, and concatenate all other columns to form the input (which produces 5.6 M candidate tasks)
141
+ 5. Rule-based-checks to reject tasks
142
+ a) We reject a task if it has less than 6 rows. Note that tasks may have fewer rows than their origin tables since we remove rows where the output column is empty.
143
+ b) We reject tasks if any input maps to multiple outputs.
144
+ c) We reject tasks if it has fewer than 2 output classes.
145
+ d) We reject a task if the output column alone has >20% non-English text.
146
+ e) We reject a task if the classes are heavily imbalanced.
147
+
148
+ 6. Lastly we apply domain-level filtering. Initial iterations of our dataset found a significant imbalance in terms of the website of origin for our generated tasks. In particular, we found that the mos-frequent domain in the WDC corpus, Cappex.com, was emphasized by our export criteria such that this website alone represented 41% of our total tasks. Since we want our dataset to represent the diversity of all the tables available on the web, we apply a hard fix for this imbalance by limiting the number of tasks per domain. Starting from the initial corpus of 50M tables from 323160 web domains, our resulting longlist of tasks comprises more than X for a total of 413350 tasks.
149
+
150
+ #### Who are the source language producers?
151
+
152
+ The dataset is extracted from [WDC Web Table Corpora](http://webdatacommons.org/webtables/).
153
+
154
+ ### Annotations
155
+
156
+ #### Annotation process
157
+
158
+ No annotation Process
159
+
160
+ #### Who are the annotators?
161
+
162
+ -
163
+
164
+ ### Personal and Sensitive Information
165
+
166
+ The data was extracted from [WDC Web Table Corpora](http://webdatacommons.org/webtables/), which in turn extracted tables from the [Common Crawl](https://commoncrawl.org/). We did not filter the data in any way. Thus any user identities or otherwise sensitive information (e.g. data that reveals racial or ethnic origins, sexual orientations, religious beliefs, political opinions or union memberships, or locations; financial or health data; biometric or genetic data; forms of government identification, such as social security numbers; criminal history, etc.) might be contained in our dataset.
167
+
168
+ ## Considerations for Using the Data
169
+
170
+ ### Social Impact of Dataset
171
+
172
+ The purpose of this dataset is to help develop models that are better at few-shot learning and have higher few-shot performance by fine-tuning few-shot tasks extracted from tables.
173
+
174
+ While tables have a similar structure to few-shot tasks and we do see an improved performance on few-shot tasks in our paper, we want to make clear that finetuning on tables also has its risks. First of all, since the tables are extracted from the web, they may contain user identities or otherwise sensitive information which a model might reveal at inference, or which could influence the learning process of a model in a negative way. Second, since tables are very diverse in nature, the model also trains on low-quality data or data with an unusual structure. While it is interesting that training on such data improves few-shot performance on downstream tasks, this could also imply that the model learns concepts that are very dissimilar to human concepts that would be useful for a certain downstream task. In other words, it is possible that the model learns weird things that are helpful on the evaluated downstream tasks, but might lead to bad out-of-distribution behavior.
175
+
176
+ ### Discussion of Biases
177
+
178
+ Since our dataset contains tables that are scraped from the web, it will also contain many toxic, racist, sexist, and otherwise harmful biases and texts. We have not run any analysis on the biases prevalent in our datasets. Neither have we explicitly filtered the content for toxic content.
179
+ This implies that a model trained on our dataset will reinforce harmful biases and toxic text that exist in our dataset.
180
+
181
+
182
+ ### Other Known Limitations
183
+
184
+ [Needs More Information]
185
+
186
+ ## Additional Information
187
+
188
+ ### Dataset Curators
189
+ Mention all authors
190
+
191
+ ### Licensing Information
192
+ Apache 2.0
193
+
194
+ ### Citation Information
195
+
196
+ [Needs More Information]
huggingface_dataset/Dataset_Card/Lo_clip-bert-data.md ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ license:
5
+ - cc-by-4.0
6
+ multilinguality:
7
+ - monolingual
8
+ ---
9
+
10
+ # CLIP-BERT training data
11
+
12
+ This data was used to train the CLIP-BERT model first described in [this paper](https://arxiv.org/abs/2109.11321).
13
+
14
+ The dataset is based on text and images from MS COCO, SBU Captions, Visual Genome QA and Conceptual Captions.
15
+
16
+ The image features have been extracted using the CLIP model [openai/clip-vit-base-patch32](https://huggingface.co/openai/clip-vit-base-patch32) available on Huggingface.
huggingface_dataset/Dataset_Card/MicPie_unpredictable_cluster13.md ADDED
@@ -0,0 +1,250 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - no-annotation
4
+ language_creators:
5
+ - found
6
+ language:
7
+ - en
8
+ license:
9
+ - apache-2.0
10
+ multilinguality:
11
+ - monolingual
12
+ pretty_name: UnpredicTable-cluster13
13
+ size_categories:
14
+ - 100K<n<1M
15
+ source_datasets: []
16
+ task_categories:
17
+ - multiple-choice
18
+ - question-answering
19
+ - zero-shot-classification
20
+ - text2text-generation
21
+ - table-question-answering
22
+ - text-generation
23
+ - text-classification
24
+ - tabular-classification
25
+ task_ids:
26
+ - multiple-choice-qa
27
+ - extractive-qa
28
+ - open-domain-qa
29
+ - closed-domain-qa
30
+ - closed-book-qa
31
+ - open-book-qa
32
+ - language-modeling
33
+ - multi-class-classification
34
+ - natural-language-inference
35
+ - topic-classification
36
+ - multi-label-classification
37
+ - tabular-multi-class-classification
38
+ - tabular-multi-label-classification
39
+ ---
40
+
41
+
42
+ # Dataset Card for "UnpredicTable-cluster13" - Dataset of Few-shot Tasks from Tables
43
+
44
+ ## Table of Contents
45
+ - [Dataset Description](#dataset-description)
46
+ - [Dataset Summary](#dataset-summary)
47
+ - [Supported Tasks](#supported-tasks-and-leaderboards)
48
+ - [Languages](#languages)
49
+ - [Dataset Structure](#dataset-structure)
50
+ - [Data Instances](#data-instances)
51
+ - [Data Fields](#data-instances)
52
+ - [Data Splits](#data-instances)
53
+ - [Dataset Creation](#dataset-creation)
54
+ - [Curation Rationale](#curation-rationale)
55
+ - [Source Data](#source-data)
56
+ - [Annotations](#annotations)
57
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
58
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
59
+ - [Social Impact of Dataset](#social-impact-of-dataset)
60
+ - [Discussion of Biases](#discussion-of-biases)
61
+ - [Other Known Limitations](#other-known-limitations)
62
+ - [Additional Information](#additional-information)
63
+ - [Dataset Curators](#dataset-curators)
64
+ - [Licensing Information](#licensing-information)
65
+ - [Citation Information](#citation-information)
66
+
67
+ ## Dataset Description
68
+
69
+ - **Homepage:** https://ethanperez.net/unpredictable
70
+ - **Repository:** https://github.com/JunShern/few-shot-adaptation
71
+ - **Paper:** Few-shot Adaptation Works with UnpredicTable Data
72
+ - **Point of Contact:** junshern@nyu.edu, perez@nyu.edu
73
+
74
+ ### Dataset Summary
75
+
76
+ The UnpredicTable dataset consists of web tables formatted as few-shot tasks for fine-tuning language models to improve their few-shot performance.
77
+
78
+ There are several dataset versions available:
79
+
80
+ * [UnpredicTable-full](https://huggingface.co/datasets/MicPie/unpredictable_full): Starting from the initial WTC corpus of 50M tables, we apply our tables-to-tasks procedure to produce our resulting dataset, [UnpredicTable-full](https://huggingface.co/datasets/MicPie/unpredictable_full), which comprises 413,299 tasks from 23,744 unique websites.
81
+
82
+ * [UnpredicTable-unique](https://huggingface.co/datasets/MicPie/unpredictable_unique): This is the same as [UnpredicTable-full](https://huggingface.co/datasets/MicPie/unpredictable_full) but filtered to have a maximum of one task per website. [UnpredicTable-unique](https://huggingface.co/datasets/MicPie/unpredictable_unique) contains exactly 23,744 tasks from 23,744 websites.
83
+
84
+ * [UnpredicTable-5k](https://huggingface.co/datasets/MicPie/unpredictable_5k): This dataset contains 5k random tables from the full dataset.
85
+
86
+ * UnpredicTable data subsets based on a manual human quality rating (please see our publication for details of the ratings):
87
+ * [UnpredicTable-rated-low](https://huggingface.co/datasets/MicPie/unpredictable_rated-low)
88
+ * [UnpredicTable-rated-medium](https://huggingface.co/datasets/MicPie/unpredictable_rated-medium)
89
+ * [UnpredicTable-rated-high](https://huggingface.co/datasets/MicPie/unpredictable_rated-high)
90
+
91
+ * UnpredicTable data subsets based on the website of origin:
92
+ * [UnpredicTable-baseball-fantasysports-yahoo-com](https://huggingface.co/datasets/MicPie/unpredictable_baseball-fantasysports-yahoo-com)
93
+ * [UnpredicTable-bulbapedia-bulbagarden-net](https://huggingface.co/datasets/MicPie/unpredictable_bulbapedia-bulbagarden-net)
94
+ * [UnpredicTable-cappex-com](https://huggingface.co/datasets/MicPie/unpredictable_cappex-com)
95
+ * [UnpredicTable-cram-com](https://huggingface.co/datasets/MicPie/unpredictable_cram-com)
96
+ * [UnpredicTable-dividend-com](https://huggingface.co/datasets/MicPie/unpredictable_dividend-com)
97
+ * [UnpredicTable-dummies-com](https://huggingface.co/datasets/MicPie/unpredictable_dummies-com)
98
+ * [UnpredicTable-en-wikipedia-org](https://huggingface.co/datasets/MicPie/unpredictable_en-wikipedia-org)
99
+ * [UnpredicTable-ensembl-org](https://huggingface.co/datasets/MicPie/unpredictable_ensembl-org)
100
+ * [UnpredicTable-gamefaqs-com](https://huggingface.co/datasets/MicPie/unpredictable_gamefaqs-com)
101
+ * [UnpredicTable-mgoblog-com](https://huggingface.co/datasets/MicPie/unpredictable_mgoblog-com)
102
+ * [UnpredicTable-mmo-champion-com](https://huggingface.co/datasets/MicPie/unpredictable_mmo-champion-com)
103
+ * [UnpredicTable-msdn-microsoft-com](https://huggingface.co/datasets/MicPie/unpredictable_msdn-microsoft-com)
104
+ * [UnpredicTable-phonearena-com](https://huggingface.co/datasets/MicPie/unpredictable_phonearena-com)
105
+ * [UnpredicTable-sittercity-com](https://huggingface.co/datasets/MicPie/unpredictable_sittercity-com)
106
+ * [UnpredicTable-sporcle-com](https://huggingface.co/datasets/MicPie/unpredictable_sporcle-com)
107
+ * [UnpredicTable-studystack-com](https://huggingface.co/datasets/MicPie/unpredictable_studystack-com)
108
+ * [UnpredicTable-support-google-com](https://huggingface.co/datasets/MicPie/unpredictable_support-google-com)
109
+ * [UnpredicTable-w3-org](https://huggingface.co/datasets/MicPie/unpredictable_w3-org)
110
+ * [UnpredicTable-wiki-openmoko-org](https://huggingface.co/datasets/MicPie/unpredictable_wiki-openmoko-org)
111
+ * [UnpredicTable-wkdu-org](https://huggingface.co/datasets/MicPie/unpredictable_wkdu-org)
112
+
113
+
114
+ * UnpredicTable data subsets based on clustering (for the clustering details please see our publication):
115
+ * [UnpredicTable-cluster00](https://huggingface.co/datasets/MicPie/unpredictable_cluster00)
116
+ * [UnpredicTable-cluster01](https://huggingface.co/datasets/MicPie/unpredictable_cluster01)
117
+ * [UnpredicTable-cluster02](https://huggingface.co/datasets/MicPie/unpredictable_cluster02)
118
+ * [UnpredicTable-cluster03](https://huggingface.co/datasets/MicPie/unpredictable_cluster03)
119
+ * [UnpredicTable-cluster04](https://huggingface.co/datasets/MicPie/unpredictable_cluster04)
120
+ * [UnpredicTable-cluster05](https://huggingface.co/datasets/MicPie/unpredictable_cluster05)
121
+ * [UnpredicTable-cluster06](https://huggingface.co/datasets/MicPie/unpredictable_cluster06)
122
+ * [UnpredicTable-cluster07](https://huggingface.co/datasets/MicPie/unpredictable_cluster07)
123
+ * [UnpredicTable-cluster08](https://huggingface.co/datasets/MicPie/unpredictable_cluster08)
124
+ * [UnpredicTable-cluster09](https://huggingface.co/datasets/MicPie/unpredictable_cluster09)
125
+ * [UnpredicTable-cluster10](https://huggingface.co/datasets/MicPie/unpredictable_cluster10)
126
+ * [UnpredicTable-cluster11](https://huggingface.co/datasets/MicPie/unpredictable_cluster11)
127
+ * [UnpredicTable-cluster12](https://huggingface.co/datasets/MicPie/unpredictable_cluster12)
128
+ * [UnpredicTable-cluster13](https://huggingface.co/datasets/MicPie/unpredictable_cluster13)
129
+ * [UnpredicTable-cluster14](https://huggingface.co/datasets/MicPie/unpredictable_cluster14)
130
+ * [UnpredicTable-cluster15](https://huggingface.co/datasets/MicPie/unpredictable_cluster15)
131
+ * [UnpredicTable-cluster16](https://huggingface.co/datasets/MicPie/unpredictable_cluster16)
132
+ * [UnpredicTable-cluster17](https://huggingface.co/datasets/MicPie/unpredictable_cluster17)
133
+ * [UnpredicTable-cluster18](https://huggingface.co/datasets/MicPie/unpredictable_cluster18)
134
+ * [UnpredicTable-cluster19](https://huggingface.co/datasets/MicPie/unpredictable_cluster19)
135
+ * [UnpredicTable-cluster20](https://huggingface.co/datasets/MicPie/unpredictable_cluster20)
136
+ * [UnpredicTable-cluster21](https://huggingface.co/datasets/MicPie/unpredictable_cluster21)
137
+ * [UnpredicTable-cluster22](https://huggingface.co/datasets/MicPie/unpredictable_cluster22)
138
+ * [UnpredicTable-cluster23](https://huggingface.co/datasets/MicPie/unpredictable_cluster23)
139
+ * [UnpredicTable-cluster24](https://huggingface.co/datasets/MicPie/unpredictable_cluster24)
140
+ * [UnpredicTable-cluster25](https://huggingface.co/datasets/MicPie/unpredictable_cluster25)
141
+ * [UnpredicTable-cluster26](https://huggingface.co/datasets/MicPie/unpredictable_cluster26)
142
+ * [UnpredicTable-cluster27](https://huggingface.co/datasets/MicPie/unpredictable_cluster27)
143
+ * [UnpredicTable-cluster28](https://huggingface.co/datasets/MicPie/unpredictable_cluster28)
144
+ * [UnpredicTable-cluster29](https://huggingface.co/datasets/MicPie/unpredictable_cluster29)
145
+ * [UnpredicTable-cluster-noise](https://huggingface.co/datasets/MicPie/unpredictable_cluster-noise)
146
+
147
+ ### Supported Tasks and Leaderboards
148
+
149
+ Since the tables come from the web, the distribution of tasks and topics is very broad. The shape of our dataset is very wide, i.e., we have 1000's of tasks, while each task has only a few examples, compared to most current NLP datasets which are very deep, i.e., 10s of tasks with many examples. This implies that our dataset covers a broad range of potential tasks, e.g., multiple-choice, question-answering, table-question-answering, text-classification, etc.
150
+
151
+ The intended use of this dataset is to improve few-shot performance by fine-tuning/pre-training on our dataset.
152
+
153
+ ### Languages
154
+
155
+ English
156
+
157
+ ## Dataset Structure
158
+
159
+ ### Data Instances
160
+
161
+ Each task is represented as a jsonline file and consists of several few-shot examples. Each example is a dictionary containing a field 'task', which identifies the task, followed by an 'input', 'options', and 'output' field. The 'input' field contains several column elements of the same row in the table, while the 'output' field is a target which represents an individual column of the same row. Each task contains several such examples which can be concatenated as a few-shot task. In the case of multiple choice classification, the 'options' field contains the possible classes that a model needs to choose from.
162
+
163
+ There are also additional meta-data fields such as 'pageTitle', 'title', 'outputColName', 'url', 'wdcFile'.
164
+
165
+ ### Data Fields
166
+
167
+ 'task': task identifier
168
+
169
+ 'input': column elements of a specific row in the table.
170
+
171
+ 'options': for multiple choice classification, it provides the options to choose from.
172
+
173
+ 'output': target column element of the same row as input.
174
+
175
+ 'pageTitle': the title of the page containing the table.
176
+
177
+ 'outputColName': output column name
178
+
179
+ 'url': url to the website containing the table
180
+
181
+ 'wdcFile': WDC Web Table Corpus file
182
+
183
+ ### Data Splits
184
+
185
+ The UnpredicTable datasets do not come with additional data splits.
186
+
187
+ ## Dataset Creation
188
+
189
+ ### Curation Rationale
190
+
191
+ Few-shot training on multi-task datasets has been demonstrated to improve language models' few-shot learning (FSL) performance on new tasks, but it is unclear which training tasks lead to effective downstream task adaptation. Few-shot learning datasets are typically produced with expensive human curation, limiting the scale and diversity of the training tasks available to study. As an alternative source of few-shot data, we automatically extract 413,299 tasks from diverse internet tables. We provide this as a research resource to investigate the relationship between training data and few-shot learning.
192
+
193
+ ### Source Data
194
+
195
+ #### Initial Data Collection and Normalization
196
+
197
+ We use internet tables from the English-language Relational Subset of the WDC Web Table Corpus 2015 (WTC). The WTC dataset tables were extracted from the July 2015 Common Crawl web corpus (http://webdatacommons.org/webtables/2015/EnglishStatistics.html). The dataset contains 50,820,165 tables from 323,160 web domains. We then convert the tables into few-shot learning tasks. Please see our publication for more details on the data collection and conversion pipeline.
198
+
199
+ #### Who are the source language producers?
200
+
201
+ The dataset is extracted from [WDC Web Table Corpora](http://webdatacommons.org/webtables/).
202
+
203
+ ### Annotations
204
+
205
+ #### Annotation process
206
+
207
+ Manual annotation was only carried out for the [UnpredicTable-rated-low](https://huggingface.co/datasets/MicPie/unpredictable_rated-low),
208
+ [UnpredicTable-rated-medium](https://huggingface.co/datasets/MicPie/unpredictable_rated-medium), and [UnpredicTable-rated-high](https://huggingface.co/datasets/MicPie/unpredictable_rated-high) data subsets to rate task quality. Detailed instructions of the annotation instructions can be found in our publication.
209
+
210
+ #### Who are the annotators?
211
+
212
+ Annotations were carried out by a lab assistant.
213
+
214
+ ### Personal and Sensitive Information
215
+
216
+ The data was extracted from [WDC Web Table Corpora](http://webdatacommons.org/webtables/), which in turn extracted tables from the [Common Crawl](https://commoncrawl.org/). We did not filter the data in any way. Thus any user identities or otherwise sensitive information (e.g., data that reveals racial or ethnic origins, sexual orientations, religious beliefs, political opinions or union memberships, or locations; financial or health data; biometric or genetic data; forms of government identification, such as social security numbers; criminal history, etc.) might be contained in our dataset.
217
+
218
+ ## Considerations for Using the Data
219
+
220
+ ### Social Impact of Dataset
221
+
222
+ This dataset is intended for use as a research resource to investigate the relationship between training data and few-shot learning. As such, it contains high- and low-quality data, as well as diverse content that may be untruthful or inappropriate. Without careful investigation, it should not be used for training models that will be deployed for use in decision-critical or user-facing situations.
223
+
224
+ ### Discussion of Biases
225
+
226
+ Since our dataset contains tables that are scraped from the web, it will also contain many toxic, racist, sexist, and otherwise harmful biases and texts. We have not run any analysis on the biases prevalent in our datasets. Neither have we explicitly filtered the content. This implies that a model trained on our dataset may potentially reflect harmful biases and toxic text that exist in our dataset.
227
+
228
+ ### Other Known Limitations
229
+
230
+ No additional known limitations.
231
+
232
+ ## Additional Information
233
+
234
+ ### Dataset Curators
235
+ Jun Shern Chan, Michael Pieler, Jonathan Jao, Jérémy Scheurer, Ethan Perez
236
+
237
+ ### Licensing Information
238
+ Apache 2.0
239
+
240
+ ### Citation Information
241
+
242
+ ```
243
+ @misc{chan2022few,
244
+ author = {Chan, Jun Shern and Pieler, Michael and Jao, Jonathan and Scheurer, Jérémy and Perez, Ethan},
245
+ title = {Few-shot Adaptation Works with UnpredicTable Data},
246
+ publisher={arXiv},
247
+ year = {2022},
248
+ url = {https://arxiv.org/abs/2208.01009}
249
+ }
250
+ ```
huggingface_dataset/Dataset_Card/MicPie_unpredictable_cluster20.md ADDED
@@ -0,0 +1,250 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - no-annotation
4
+ language_creators:
5
+ - found
6
+ language:
7
+ - en
8
+ license:
9
+ - apache-2.0
10
+ multilinguality:
11
+ - monolingual
12
+ pretty_name: UnpredicTable-cluster20
13
+ size_categories:
14
+ - 100K<n<1M
15
+ source_datasets: []
16
+ task_categories:
17
+ - multiple-choice
18
+ - question-answering
19
+ - zero-shot-classification
20
+ - text2text-generation
21
+ - table-question-answering
22
+ - text-generation
23
+ - text-classification
24
+ - tabular-classification
25
+ task_ids:
26
+ - multiple-choice-qa
27
+ - extractive-qa
28
+ - open-domain-qa
29
+ - closed-domain-qa
30
+ - closed-book-qa
31
+ - open-book-qa
32
+ - language-modeling
33
+ - multi-class-classification
34
+ - natural-language-inference
35
+ - topic-classification
36
+ - multi-label-classification
37
+ - tabular-multi-class-classification
38
+ - tabular-multi-label-classification
39
+ ---
40
+
41
+
42
+ # Dataset Card for "UnpredicTable-cluster20" - Dataset of Few-shot Tasks from Tables
43
+
44
+ ## Table of Contents
45
+ - [Dataset Description](#dataset-description)
46
+ - [Dataset Summary](#dataset-summary)
47
+ - [Supported Tasks](#supported-tasks-and-leaderboards)
48
+ - [Languages](#languages)
49
+ - [Dataset Structure](#dataset-structure)
50
+ - [Data Instances](#data-instances)
51
+ - [Data Fields](#data-instances)
52
+ - [Data Splits](#data-instances)
53
+ - [Dataset Creation](#dataset-creation)
54
+ - [Curation Rationale](#curation-rationale)
55
+ - [Source Data](#source-data)
56
+ - [Annotations](#annotations)
57
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
58
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
59
+ - [Social Impact of Dataset](#social-impact-of-dataset)
60
+ - [Discussion of Biases](#discussion-of-biases)
61
+ - [Other Known Limitations](#other-known-limitations)
62
+ - [Additional Information](#additional-information)
63
+ - [Dataset Curators](#dataset-curators)
64
+ - [Licensing Information](#licensing-information)
65
+ - [Citation Information](#citation-information)
66
+
67
+ ## Dataset Description
68
+
69
+ - **Homepage:** https://ethanperez.net/unpredictable
70
+ - **Repository:** https://github.com/JunShern/few-shot-adaptation
71
+ - **Paper:** Few-shot Adaptation Works with UnpredicTable Data
72
+ - **Point of Contact:** junshern@nyu.edu, perez@nyu.edu
73
+
74
+ ### Dataset Summary
75
+
76
+ The UnpredicTable dataset consists of web tables formatted as few-shot tasks for fine-tuning language models to improve their few-shot performance.
77
+
78
+ There are several dataset versions available:
79
+
80
+ * [UnpredicTable-full](https://huggingface.co/datasets/MicPie/unpredictable_full): Starting from the initial WTC corpus of 50M tables, we apply our tables-to-tasks procedure to produce our resulting dataset, [UnpredicTable-full](https://huggingface.co/datasets/MicPie/unpredictable_full), which comprises 413,299 tasks from 23,744 unique websites.
81
+
82
+ * [UnpredicTable-unique](https://huggingface.co/datasets/MicPie/unpredictable_unique): This is the same as [UnpredicTable-full](https://huggingface.co/datasets/MicPie/unpredictable_full) but filtered to have a maximum of one task per website. [UnpredicTable-unique](https://huggingface.co/datasets/MicPie/unpredictable_unique) contains exactly 23,744 tasks from 23,744 websites.
83
+
84
+ * [UnpredicTable-5k](https://huggingface.co/datasets/MicPie/unpredictable_5k): This dataset contains 5k random tables from the full dataset.
85
+
86
+ * UnpredicTable data subsets based on a manual human quality rating (please see our publication for details of the ratings):
87
+ * [UnpredicTable-rated-low](https://huggingface.co/datasets/MicPie/unpredictable_rated-low)
88
+ * [UnpredicTable-rated-medium](https://huggingface.co/datasets/MicPie/unpredictable_rated-medium)
89
+ * [UnpredicTable-rated-high](https://huggingface.co/datasets/MicPie/unpredictable_rated-high)
90
+
91
+ * UnpredicTable data subsets based on the website of origin:
92
+ * [UnpredicTable-baseball-fantasysports-yahoo-com](https://huggingface.co/datasets/MicPie/unpredictable_baseball-fantasysports-yahoo-com)
93
+ * [UnpredicTable-bulbapedia-bulbagarden-net](https://huggingface.co/datasets/MicPie/unpredictable_bulbapedia-bulbagarden-net)
94
+ * [UnpredicTable-cappex-com](https://huggingface.co/datasets/MicPie/unpredictable_cappex-com)
95
+ * [UnpredicTable-cram-com](https://huggingface.co/datasets/MicPie/unpredictable_cram-com)
96
+ * [UnpredicTable-dividend-com](https://huggingface.co/datasets/MicPie/unpredictable_dividend-com)
97
+ * [UnpredicTable-dummies-com](https://huggingface.co/datasets/MicPie/unpredictable_dummies-com)
98
+ * [UnpredicTable-en-wikipedia-org](https://huggingface.co/datasets/MicPie/unpredictable_en-wikipedia-org)
99
+ * [UnpredicTable-ensembl-org](https://huggingface.co/datasets/MicPie/unpredictable_ensembl-org)
100
+ * [UnpredicTable-gamefaqs-com](https://huggingface.co/datasets/MicPie/unpredictable_gamefaqs-com)
101
+ * [UnpredicTable-mgoblog-com](https://huggingface.co/datasets/MicPie/unpredictable_mgoblog-com)
102
+ * [UnpredicTable-mmo-champion-com](https://huggingface.co/datasets/MicPie/unpredictable_mmo-champion-com)
103
+ * [UnpredicTable-msdn-microsoft-com](https://huggingface.co/datasets/MicPie/unpredictable_msdn-microsoft-com)
104
+ * [UnpredicTable-phonearena-com](https://huggingface.co/datasets/MicPie/unpredictable_phonearena-com)
105
+ * [UnpredicTable-sittercity-com](https://huggingface.co/datasets/MicPie/unpredictable_sittercity-com)
106
+ * [UnpredicTable-sporcle-com](https://huggingface.co/datasets/MicPie/unpredictable_sporcle-com)
107
+ * [UnpredicTable-studystack-com](https://huggingface.co/datasets/MicPie/unpredictable_studystack-com)
108
+ * [UnpredicTable-support-google-com](https://huggingface.co/datasets/MicPie/unpredictable_support-google-com)
109
+ * [UnpredicTable-w3-org](https://huggingface.co/datasets/MicPie/unpredictable_w3-org)
110
+ * [UnpredicTable-wiki-openmoko-org](https://huggingface.co/datasets/MicPie/unpredictable_wiki-openmoko-org)
111
+ * [UnpredicTable-wkdu-org](https://huggingface.co/datasets/MicPie/unpredictable_wkdu-org)
112
+
113
+
114
+ * UnpredicTable data subsets based on clustering (for the clustering details please see our publication):
115
+ * [UnpredicTable-cluster00](https://huggingface.co/datasets/MicPie/unpredictable_cluster00)
116
+ * [UnpredicTable-cluster01](https://huggingface.co/datasets/MicPie/unpredictable_cluster01)
117
+ * [UnpredicTable-cluster02](https://huggingface.co/datasets/MicPie/unpredictable_cluster02)
118
+ * [UnpredicTable-cluster03](https://huggingface.co/datasets/MicPie/unpredictable_cluster03)
119
+ * [UnpredicTable-cluster04](https://huggingface.co/datasets/MicPie/unpredictable_cluster04)
120
+ * [UnpredicTable-cluster05](https://huggingface.co/datasets/MicPie/unpredictable_cluster05)
121
+ * [UnpredicTable-cluster06](https://huggingface.co/datasets/MicPie/unpredictable_cluster06)
122
+ * [UnpredicTable-cluster07](https://huggingface.co/datasets/MicPie/unpredictable_cluster07)
123
+ * [UnpredicTable-cluster08](https://huggingface.co/datasets/MicPie/unpredictable_cluster08)
124
+ * [UnpredicTable-cluster09](https://huggingface.co/datasets/MicPie/unpredictable_cluster09)
125
+ * [UnpredicTable-cluster10](https://huggingface.co/datasets/MicPie/unpredictable_cluster10)
126
+ * [UnpredicTable-cluster11](https://huggingface.co/datasets/MicPie/unpredictable_cluster11)
127
+ * [UnpredicTable-cluster12](https://huggingface.co/datasets/MicPie/unpredictable_cluster12)
128
+ * [UnpredicTable-cluster13](https://huggingface.co/datasets/MicPie/unpredictable_cluster13)
129
+ * [UnpredicTable-cluster14](https://huggingface.co/datasets/MicPie/unpredictable_cluster14)
130
+ * [UnpredicTable-cluster15](https://huggingface.co/datasets/MicPie/unpredictable_cluster15)
131
+ * [UnpredicTable-cluster16](https://huggingface.co/datasets/MicPie/unpredictable_cluster16)
132
+ * [UnpredicTable-cluster17](https://huggingface.co/datasets/MicPie/unpredictable_cluster17)
133
+ * [UnpredicTable-cluster18](https://huggingface.co/datasets/MicPie/unpredictable_cluster18)
134
+ * [UnpredicTable-cluster19](https://huggingface.co/datasets/MicPie/unpredictable_cluster19)
135
+ * [UnpredicTable-cluster20](https://huggingface.co/datasets/MicPie/unpredictable_cluster20)
136
+ * [UnpredicTable-cluster21](https://huggingface.co/datasets/MicPie/unpredictable_cluster21)
137
+ * [UnpredicTable-cluster22](https://huggingface.co/datasets/MicPie/unpredictable_cluster22)
138
+ * [UnpredicTable-cluster23](https://huggingface.co/datasets/MicPie/unpredictable_cluster23)
139
+ * [UnpredicTable-cluster24](https://huggingface.co/datasets/MicPie/unpredictable_cluster24)
140
+ * [UnpredicTable-cluster25](https://huggingface.co/datasets/MicPie/unpredictable_cluster25)
141
+ * [UnpredicTable-cluster26](https://huggingface.co/datasets/MicPie/unpredictable_cluster26)
142
+ * [UnpredicTable-cluster27](https://huggingface.co/datasets/MicPie/unpredictable_cluster27)
143
+ * [UnpredicTable-cluster28](https://huggingface.co/datasets/MicPie/unpredictable_cluster28)
144
+ * [UnpredicTable-cluster29](https://huggingface.co/datasets/MicPie/unpredictable_cluster29)
145
+ * [UnpredicTable-cluster-noise](https://huggingface.co/datasets/MicPie/unpredictable_cluster-noise)
146
+
147
+ ### Supported Tasks and Leaderboards
148
+
149
+ Since the tables come from the web, the distribution of tasks and topics is very broad. The shape of our dataset is very wide, i.e., we have 1000's of tasks, while each task has only a few examples, compared to most current NLP datasets which are very deep, i.e., 10s of tasks with many examples. This implies that our dataset covers a broad range of potential tasks, e.g., multiple-choice, question-answering, table-question-answering, text-classification, etc.
150
+
151
+ The intended use of this dataset is to improve few-shot performance by fine-tuning/pre-training on our dataset.
152
+
153
+ ### Languages
154
+
155
+ English
156
+
157
+ ## Dataset Structure
158
+
159
+ ### Data Instances
160
+
161
+ Each task is represented as a jsonline file and consists of several few-shot examples. Each example is a dictionary containing a field 'task', which identifies the task, followed by an 'input', 'options', and 'output' field. The 'input' field contains several column elements of the same row in the table, while the 'output' field is a target which represents an individual column of the same row. Each task contains several such examples which can be concatenated as a few-shot task. In the case of multiple choice classification, the 'options' field contains the possible classes that a model needs to choose from.
162
+
163
+ There are also additional meta-data fields such as 'pageTitle', 'title', 'outputColName', 'url', 'wdcFile'.
164
+
165
+ ### Data Fields
166
+
167
+ 'task': task identifier
168
+
169
+ 'input': column elements of a specific row in the table.
170
+
171
+ 'options': for multiple choice classification, it provides the options to choose from.
172
+
173
+ 'output': target column element of the same row as input.
174
+
175
+ 'pageTitle': the title of the page containing the table.
176
+
177
+ 'outputColName': output column name
178
+
179
+ 'url': url to the website containing the table
180
+
181
+ 'wdcFile': WDC Web Table Corpus file
182
+
183
+ ### Data Splits
184
+
185
+ The UnpredicTable datasets do not come with additional data splits.
186
+
187
+ ## Dataset Creation
188
+
189
+ ### Curation Rationale
190
+
191
+ Few-shot training on multi-task datasets has been demonstrated to improve language models' few-shot learning (FSL) performance on new tasks, but it is unclear which training tasks lead to effective downstream task adaptation. Few-shot learning datasets are typically produced with expensive human curation, limiting the scale and diversity of the training tasks available to study. As an alternative source of few-shot data, we automatically extract 413,299 tasks from diverse internet tables. We provide this as a research resource to investigate the relationship between training data and few-shot learning.
192
+
193
+ ### Source Data
194
+
195
+ #### Initial Data Collection and Normalization
196
+
197
+ We use internet tables from the English-language Relational Subset of the WDC Web Table Corpus 2015 (WTC). The WTC dataset tables were extracted from the July 2015 Common Crawl web corpus (http://webdatacommons.org/webtables/2015/EnglishStatistics.html). The dataset contains 50,820,165 tables from 323,160 web domains. We then convert the tables into few-shot learning tasks. Please see our publication for more details on the data collection and conversion pipeline.
198
+
199
+ #### Who are the source language producers?
200
+
201
+ The dataset is extracted from [WDC Web Table Corpora](http://webdatacommons.org/webtables/).
202
+
203
+ ### Annotations
204
+
205
+ #### Annotation process
206
+
207
+ Manual annotation was only carried out for the [UnpredicTable-rated-low](https://huggingface.co/datasets/MicPie/unpredictable_rated-low),
208
+ [UnpredicTable-rated-medium](https://huggingface.co/datasets/MicPie/unpredictable_rated-medium), and [UnpredicTable-rated-high](https://huggingface.co/datasets/MicPie/unpredictable_rated-high) data subsets to rate task quality. Detailed instructions of the annotation instructions can be found in our publication.
209
+
210
+ #### Who are the annotators?
211
+
212
+ Annotations were carried out by a lab assistant.
213
+
214
+ ### Personal and Sensitive Information
215
+
216
+ The data was extracted from [WDC Web Table Corpora](http://webdatacommons.org/webtables/), which in turn extracted tables from the [Common Crawl](https://commoncrawl.org/). We did not filter the data in any way. Thus any user identities or otherwise sensitive information (e.g., data that reveals racial or ethnic origins, sexual orientations, religious beliefs, political opinions or union memberships, or locations; financial or health data; biometric or genetic data; forms of government identification, such as social security numbers; criminal history, etc.) might be contained in our dataset.
217
+
218
+ ## Considerations for Using the Data
219
+
220
+ ### Social Impact of Dataset
221
+
222
+ This dataset is intended for use as a research resource to investigate the relationship between training data and few-shot learning. As such, it contains high- and low-quality data, as well as diverse content that may be untruthful or inappropriate. Without careful investigation, it should not be used for training models that will be deployed for use in decision-critical or user-facing situations.
223
+
224
+ ### Discussion of Biases
225
+
226
+ Since our dataset contains tables that are scraped from the web, it will also contain many toxic, racist, sexist, and otherwise harmful biases and texts. We have not run any analysis on the biases prevalent in our datasets. Neither have we explicitly filtered the content. This implies that a model trained on our dataset may potentially reflect harmful biases and toxic text that exist in our dataset.
227
+
228
+ ### Other Known Limitations
229
+
230
+ No additional known limitations.
231
+
232
+ ## Additional Information
233
+
234
+ ### Dataset Curators
235
+ Jun Shern Chan, Michael Pieler, Jonathan Jao, Jérémy Scheurer, Ethan Perez
236
+
237
+ ### Licensing Information
238
+ Apache 2.0
239
+
240
+ ### Citation Information
241
+
242
+ ```
243
+ @misc{chan2022few,
244
+ author = {Chan, Jun Shern and Pieler, Michael and Jao, Jonathan and Scheurer, Jérémy and Perez, Ethan},
245
+ title = {Few-shot Adaptation Works with UnpredicTable Data},
246
+ publisher={arXiv},
247
+ year = {2022},
248
+ url = {https://arxiv.org/abs/2208.01009}
249
+ }
250
+ ```
huggingface_dataset/Dataset_Card/Nicky0007_cointelegraph_news_English.md ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ task_categories:
3
+ - token-classification
4
+ - question-answering
5
+ language:
6
+ - en
7
+ size_categories:
8
+ - 10K<n<100K
9
+ ---
10
+ # Dataset cointelegraph English
11
+
12
+ ## Dataset Description
13
+
14
+ It is a dataset where information about the title, description, author, etc. is collected.
15
+ approx: 10041 row
16
+ page: https://cointelegraph.com/
17
+
18
+
19
+
20
+ categorie: #cryptocurrency, #Bitcoin, #Ethereum ...
huggingface_dataset/Dataset_Card/autoevaluate_autoeval-eval-anli-plain_text-f2dca1-2066067125.md ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ type: predictions
3
+ tags:
4
+ - autotrain
5
+ - evaluation
6
+ datasets:
7
+ - anli
8
+ eval_info:
9
+ task: natural_language_inference
10
+ model: MoritzLaurer/DeBERTa-v3-base-mnli-fever-anli
11
+ metrics: []
12
+ dataset_name: anli
13
+ dataset_config: plain_text
14
+ dataset_split: dev_r1
15
+ col_mapping:
16
+ text1: premise
17
+ text2: hypothesis
18
+ target: label
19
+ ---
20
+ # Dataset Card for AutoTrain Evaluator
21
+
22
+ This repository contains model predictions generated by [AutoTrain](https://huggingface.co/autotrain) for the following task and dataset:
23
+
24
+ * Task: Natural Language Inference
25
+ * Model: MoritzLaurer/DeBERTa-v3-base-mnli-fever-anli
26
+ * Dataset: anli
27
+ * Config: plain_text
28
+ * Split: dev_r1
29
+
30
+ To run new evaluation jobs, visit Hugging Face's [automatic model evaluator](https://huggingface.co/spaces/autoevaluate/model-evaluator).
31
+
32
+ ## Contributions
33
+
34
+ Thanks to [@ctkang](https://huggingface.co/ctkang) for evaluating this model.
huggingface_dataset/Dataset_Card/autoevaluate_autoeval-eval-futin__feed-sen_en_-1de085-2240171542.md ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ type: predictions
3
+ tags:
4
+ - autotrain
5
+ - evaluation
6
+ datasets:
7
+ - futin/feed
8
+ eval_info:
9
+ task: text_zero_shot_classification
10
+ model: bigscience/bloom-1b7
11
+ metrics: []
12
+ dataset_name: futin/feed
13
+ dataset_config: sen_en_
14
+ dataset_split: test
15
+ col_mapping:
16
+ text: text
17
+ classes: classes
18
+ target: target
19
+ ---
20
+ # Dataset Card for AutoTrain Evaluator
21
+
22
+ This repository contains model predictions generated by [AutoTrain](https://huggingface.co/autotrain) for the following task and dataset:
23
+
24
+ * Task: Zero-Shot Text Classification
25
+ * Model: bigscience/bloom-1b7
26
+ * Dataset: futin/feed
27
+ * Config: sen_en_
28
+ * Split: test
29
+
30
+ To run new evaluation jobs, visit Hugging Face's [automatic model evaluator](https://huggingface.co/spaces/autoevaluate/model-evaluator).
31
+
32
+ ## Contributions
33
+
34
+ Thanks to [@futin](https://huggingface.co/futin) for evaluating this model.
huggingface_dataset/Dataset_Card/autoevaluate_autoeval-eval-futin__guess-en_3-8ea950-2087767174.md ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ type: predictions
3
+ tags:
4
+ - autotrain
5
+ - evaluation
6
+ datasets:
7
+ - futin/guess
8
+ eval_info:
9
+ task: text_zero_shot_classification
10
+ model: bigscience/bloomz-560m
11
+ metrics: []
12
+ dataset_name: futin/guess
13
+ dataset_config: en_3
14
+ dataset_split: test
15
+ col_mapping:
16
+ text: text
17
+ classes: classes
18
+ target: target
19
+ ---
20
+ # Dataset Card for AutoTrain Evaluator
21
+
22
+ This repository contains model predictions generated by [AutoTrain](https://huggingface.co/autotrain) for the following task and dataset:
23
+
24
+ * Task: Zero-Shot Text Classification
25
+ * Model: bigscience/bloomz-560m
26
+ * Dataset: futin/guess
27
+ * Config: en_3
28
+ * Split: test
29
+
30
+ To run new evaluation jobs, visit Hugging Face's [automatic model evaluator](https://huggingface.co/spaces/autoevaluate/model-evaluator).
31
+
32
+ ## Contributions
33
+
34
+ Thanks to [@futin](https://huggingface.co/futin) for evaluating this model.
huggingface_dataset/Dataset_Card/autoevaluate_autoeval-staging-eval-project-ab647f27-7704970.md ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ type: predictions
3
+ tags:
4
+ - autotrain
5
+ - evaluation
6
+ datasets:
7
+ - masakhaner
8
+ eval_info:
9
+ task: entity_extraction
10
+ model: mbeukman/xlm-roberta-base-finetuned-yoruba-finetuned-ner-swahili
11
+ metrics: []
12
+ dataset_name: masakhaner
13
+ dataset_config: yor
14
+ dataset_split: test
15
+ col_mapping:
16
+ tokens: tokens
17
+ tags: ner_tags
18
+ ---
19
+ # Dataset Card for AutoTrain Evaluator
20
+
21
+ This repository contains model predictions generated by [AutoTrain](https://huggingface.co/autotrain) for the following task and dataset:
22
+
23
+ * Task: Token Classification
24
+ * Model: mbeukman/xlm-roberta-base-finetuned-yoruba-finetuned-ner-swahili
25
+ * Dataset: masakhaner
26
+
27
+ To run new evaluation jobs, visit Hugging Face's [automatic model evaluator](https://huggingface.co/spaces/autoevaluate/model-evaluator).
28
+
29
+ ## Contributions
30
+
31
+ Thanks to [@lewtun](https://huggingface.co/lewtun) for evaluating this model.
huggingface_dataset/Dataset_Card/bethecloud_golf-courses.md ADDED
@@ -0,0 +1,93 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - machine-generated
4
+ language:
5
+ - en
6
+ language_creators:
7
+ - found
8
+ license:
9
+ - mit
10
+ multilinguality:
11
+ - monolingual
12
+ pretty_name: bethecloud/golf-courses
13
+ size_categories:
14
+ - n<1K
15
+ source_datasets: []
16
+ tags:
17
+ - golf-courses
18
+ task_categories:
19
+ - image-classification
20
+ task_ids:
21
+ - multi-label-image-classification
22
+ ---
23
+
24
+ ## Dataset Description
25
+
26
+ - **Homepage: https://mirror.xyz/bitkevin.eth**
27
+ - **Repository: https://colab.research.google.com/drive/1EnqpDiKOVYhR0c6f4CgmDg2zqcbYZJpB#scrollTo=c1ef3d21-6e0e-46c9-a459-8a2ab856a5ca**
28
+ - **Point of Contact: Kevin Leffew – kleffew94@gmail.com**
29
+
30
+ ### Dataset Summary: golf-course
31
+
32
+ This dataset (bethecloud/golf-courses) includes 21 unique images of golf courses pulled from Unsplash.
33
+
34
+ The dataset is a collection of photographs taken at various golf courses around the world. The images depict a variety of scenes, including fairways, greens, bunkers, water hazards, and clubhouse facilities. The images are high resolution and have been carefully selected to provide a diverse range of visual content for fine-tuning a machine learning model.
35
+
36
+ The dataset is intended to be used in the context of the Hugging Face Dream Booth hackathon, a competition that challenges participants to build innovative applications using the Hugging Face transformers library. The submission is for the category of landscape.
37
+
38
+ Overall, this dataset provides a rich source of visual data for machine learning models looking to understand and classify elements of golf courses. Its diverse range of images and high-quality resolution make it well-suited for use in fine-tuning models for tasks such as image classification, object detection, and image segmentation.
39
+
40
+ By using the golf course images as part of their training data, participants can fine-tune their models to recognize and classify specific features and elements commonly found on golf courses. The ultimate goal after the hackathon is to pull this dataset from decentralized cloud storage (like Storj DCS), increasing its accessibility, performance, and resilience by distributing across an edge of over 17,000 uncorrelated participants.
41
+
42
+
43
+ ## Example Output
44
+ ![golf-acropolis.jpg]https://link.storjshare.io/juid5vc27dbajh6zyzplf4fah5xq/golf-course-output%2Fgolf-acropolis.png
45
+
46
+ # Usage
47
+ The golf-courses dataset can be used by modifying the instance_prompt: a photo of golf course
48
+
49
+ ### Languages
50
+
51
+ The language data in golf-courses is in English (BCP-47 en)
52
+
53
+ ## Dataset Structure
54
+
55
+ The complete dataset is GBs and consists of 21 objects.
56
+
57
+ ### Parallelized download using Decentralized Object Storage (Storj DCS)
58
+
59
+ A direct download for the dataset is located at https://link.storjshare.io/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses.
60
+
61
+ In the future, Storj DCS will be used to download large datasets (exceeding 1TB) in a highly parallel, highly performant, and highly economical manner (by utilizing a network of over 17,000 diverse and economically incentivized datacenter node endpoints.
62
+
63
+ ### Curation Rationale
64
+
65
+ This model was created as a sample by Kevin Leffew as part of the DreamBooth Hackathon.
66
+
67
+ ### Source Data
68
+
69
+ The source data for the dataset is simply pulled from Unsplash
70
+
71
+ ### Licensing Information
72
+
73
+ MIT License
74
+
75
+ ## Thanks to John Whitaker and Lewis Tunstall
76
+
77
+ Thanks to [John Whitaker](https://github.com/johnowhitaker) and [Lewis Tunstall](https://github.com/lewtun)for writing out and describing the initial hackathon parameters at https://huggingface.co/dreambooth-hackathon.
78
+
79
+ ## Example Training Data
80
+ ![golf-course1.jpg](https://link.storjshare.io/raw/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/andrew-anderson-CtyC2JjLhVg-unsplash.jpg)
81
+ ![golf-course2.jpg](https://link.storjshare.io/raw/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/dean-SuGEzQkeJno-unsplash.jpg)
82
+ ![golf-course3.jpg](https://link.storjshare.io/raw/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/amauri-cruz-filho-kBNV9WpCs5k-unsplash.jpg)
83
+ ![golf-course4.jpg](https://link.storjshare.io/raw/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/minho-yoon-_ZVEio7AkGc-unsplash.jpg)
84
+ ![golf-course5.jpg](https://link.storjshare.io/raw/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/minho-yoon-_ZVEio7AkGc-unsplash.jpg)
85
+ ![golf-course6.jpg](https://link.storjshare.io/s/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/jura-FegOaqn_4GQ-unsplash%20%281%29.jpg?wrap=1)
86
+ ![golf-course7.jpg](https://link.storjshare.io/s/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/sly-dizzle-cE6SpYTfqqg-unsplash.jpg?wrap=1)
87
+ ![golf-course8.jpg](https://link.storjshare.io/s/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/jura-FegOaqn_4GQ-unsplash.jpg)
88
+ ![golf-course9.jpg](https://link.storjshare.io/s/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/sly-dizzle-cE6SpYTfqqg-unsplash.jpg?wrap=1)
89
+ ![golf-course10.jpg](https://link.storjshare.io/s/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/dean-ricciardi-08Ipbe8GpWw-unsplash.jpg)
90
+ ![golf-course11.jpg](https://link.storjshare.io/s/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/jonas-from-berlin-UgwkaRUt2d0-unsplash.jpg)
91
+ ![golf-course12.jpg](https://link.storjshare.io/s/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/rob-tol-Ner8kdSXh0M-unsplash.jpg)
92
+ ![golf-course13.jpg](https://link.storjshare.io/s/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/richard-brutyo-HQXFhq8FNJ8-unsplash.jpg?wrap=1)
93
+ ![golf-course14.jpg](https://link.storjshare.io/s/juo7ynuvpe5svxj3hh454v6fnhba/golf-courses/edwin-compton-Z8XlmAj65iM-unsplash.jpg?wrap=1)
huggingface_dataset/Dataset_Card/casino.md ADDED
@@ -0,0 +1,342 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - expert-generated
4
+ language_creators:
5
+ - crowdsourced
6
+ language:
7
+ - en
8
+ license:
9
+ - cc-by-4.0
10
+ multilinguality:
11
+ - monolingual
12
+ size_categories:
13
+ - 1K<n<10K
14
+ source_datasets:
15
+ - original
16
+ task_categories:
17
+ - conversational
18
+ - text-generation
19
+ - fill-mask
20
+ task_ids:
21
+ - dialogue-modeling
22
+ pretty_name: Campsite Negotiation Dialogues
23
+ paperswithcode_id: casino
24
+ dataset_info:
25
+ features:
26
+ - name: chat_logs
27
+ list:
28
+ - name: text
29
+ dtype: string
30
+ - name: task_data
31
+ struct:
32
+ - name: data
33
+ dtype: string
34
+ - name: issue2youget
35
+ struct:
36
+ - name: Firewood
37
+ dtype: string
38
+ - name: Water
39
+ dtype: string
40
+ - name: Food
41
+ dtype: string
42
+ - name: issue2theyget
43
+ struct:
44
+ - name: Firewood
45
+ dtype: string
46
+ - name: Water
47
+ dtype: string
48
+ - name: Food
49
+ dtype: string
50
+ - name: id
51
+ dtype: string
52
+ - name: participant_info
53
+ struct:
54
+ - name: mturk_agent_1
55
+ struct:
56
+ - name: value2issue
57
+ struct:
58
+ - name: Low
59
+ dtype: string
60
+ - name: Medium
61
+ dtype: string
62
+ - name: High
63
+ dtype: string
64
+ - name: value2reason
65
+ struct:
66
+ - name: Low
67
+ dtype: string
68
+ - name: Medium
69
+ dtype: string
70
+ - name: High
71
+ dtype: string
72
+ - name: outcomes
73
+ struct:
74
+ - name: points_scored
75
+ dtype: int32
76
+ - name: satisfaction
77
+ dtype: string
78
+ - name: opponent_likeness
79
+ dtype: string
80
+ - name: demographics
81
+ struct:
82
+ - name: age
83
+ dtype: int32
84
+ - name: gender
85
+ dtype: string
86
+ - name: ethnicity
87
+ dtype: string
88
+ - name: education
89
+ dtype: string
90
+ - name: personality
91
+ struct:
92
+ - name: svo
93
+ dtype: string
94
+ - name: big-five
95
+ struct:
96
+ - name: extraversion
97
+ dtype: float32
98
+ - name: agreeableness
99
+ dtype: float32
100
+ - name: conscientiousness
101
+ dtype: float32
102
+ - name: emotional-stability
103
+ dtype: float32
104
+ - name: openness-to-experiences
105
+ dtype: float32
106
+ - name: mturk_agent_2
107
+ struct:
108
+ - name: value2issue
109
+ struct:
110
+ - name: Low
111
+ dtype: string
112
+ - name: Medium
113
+ dtype: string
114
+ - name: High
115
+ dtype: string
116
+ - name: value2reason
117
+ struct:
118
+ - name: Low
119
+ dtype: string
120
+ - name: Medium
121
+ dtype: string
122
+ - name: High
123
+ dtype: string
124
+ - name: outcomes
125
+ struct:
126
+ - name: points_scored
127
+ dtype: int32
128
+ - name: satisfaction
129
+ dtype: string
130
+ - name: opponent_likeness
131
+ dtype: string
132
+ - name: demographics
133
+ struct:
134
+ - name: age
135
+ dtype: int32
136
+ - name: gender
137
+ dtype: string
138
+ - name: ethnicity
139
+ dtype: string
140
+ - name: education
141
+ dtype: string
142
+ - name: personality
143
+ struct:
144
+ - name: svo
145
+ dtype: string
146
+ - name: big-five
147
+ struct:
148
+ - name: extraversion
149
+ dtype: float32
150
+ - name: agreeableness
151
+ dtype: float32
152
+ - name: conscientiousness
153
+ dtype: float32
154
+ - name: emotional-stability
155
+ dtype: float32
156
+ - name: openness-to-experiences
157
+ dtype: float32
158
+ - name: annotations
159
+ list:
160
+ list: string
161
+ splits:
162
+ - name: train
163
+ num_bytes: 3211555
164
+ num_examples: 1030
165
+ download_size: 4300019
166
+ dataset_size: 3211555
167
+ ---
168
+
169
+
170
+ # Dataset Card for Casino
171
+
172
+ ## Table of Contents
173
+ - [Dataset Description](#dataset-description)
174
+ - [Dataset Summary](#dataset-summary)
175
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
176
+ - [Languages](#languages)
177
+ - [Dataset Structure](#dataset-structure)
178
+ - [Data Instances](#data-instances)
179
+ - [Data Fields](#data-fields)
180
+ - [Data Splits](#data-splits)
181
+ - [Dataset Creation](#dataset-creation)
182
+ - [Curation Rationale](#curation-rationale)
183
+ - [Source Data](#source-data)
184
+ - [Annotations](#annotations)
185
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
186
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
187
+ - [Social Impact of Dataset](#social-impact-of-dataset)
188
+ - [Discussion of Biases](#discussion-of-biases)
189
+ - [Other Known Limitations](#other-known-limitations)
190
+ - [Additional Information](#additional-information)
191
+ - [Dataset Curators](#dataset-curators)
192
+ - [Licensing Information](#licensing-information)
193
+ - [Citation Information](#citation-information)
194
+ - [Contributions](#contributions)
195
+
196
+ ## Dataset Description
197
+
198
+ - **Repository:** [Github: Kushal Chawla CaSiNo](https://github.com/kushalchawla/CaSiNo)
199
+ - **Paper:** [CaSiNo: A Corpus of Campsite Negotiation Dialogues for Automatic Negotiation Systems](https://aclanthology.org/2021.naacl-main.254.pdf)
200
+ - **Point of Contact:** [Kushal Chawla](kchawla@usc.edu)
201
+
202
+ ### Dataset Summary
203
+
204
+ We provide a novel dataset (referred to as CaSiNo) of 1030 negotiation dialogues. Two participants take the role of campsite neighbors and negotiate for Food, Water, and Firewood packages, based on their individual preferences and requirements. This design keeps the task tractable, while still facilitating linguistically rich and personal conversations. This helps to overcome the limitations of prior negotiation datasets such as Deal or No Deal and Craigslist Bargain. Each dialogue consists of rich meta-data including participant demographics, personality, and their subjective evaluation of the negotiation in terms of satisfaction and opponent likeness.
205
+
206
+ ### Supported Tasks and Leaderboards
207
+
208
+ Train end-to-end models for negotiation
209
+
210
+ ### Languages
211
+
212
+ English
213
+
214
+ ## Dataset Structure
215
+
216
+ ### Data Instances
217
+
218
+ ```
219
+ {
220
+ "chat_logs": [
221
+ {
222
+ "text": "Hello! \ud83d\ude42 Let's work together on a deal for these packages, shall we? What are you most interested in?",
223
+ "task_data": {},
224
+ "id": "mturk_agent_1"
225
+ },
226
+ ...
227
+ ],
228
+ "participant_info": {
229
+ "mturk_agent_1":
230
+ {
231
+ "value2issue": ...
232
+ "value2reason": ...
233
+ "outcomes": ...
234
+ "demographics": ...
235
+ "personality": ...
236
+ },
237
+ "mturk_agent_2": ...
238
+ },
239
+ "annotations": [
240
+ ["Hello! \ud83d\ude42 Let's work together on a deal for these packages, shall we? What are you most interested in?", "promote-coordination,elicit-pref"],
241
+ ...
242
+ ]
243
+ }
244
+ ```
245
+
246
+ ### Data Fields
247
+
248
+ - `chat_logs`: The negotiation dialogue between two participants
249
+ - `text`: The dialogue utterance
250
+ - `task_data`: Meta-data associated with the utterance such as the deal submitted by a participant
251
+ - `id`: The ID of the participant who typed this utterance
252
+ - `participant_info`: Meta-data about the two participants in this conversation
253
+ - `mturk_agent_1`: For the first participant (Note that 'first' is just for reference. There is no order between the participants and any participant can start the conversation)
254
+ - `value2issue`: The priority order of this participant among Food, Water, Firewood
255
+ - `value2reason`: The personal arguments given by the participants themselves, consistent with the above preference order. This preference order and these arguments were submitted before the negotiation began.
256
+ - `outcomes`: The negotiation outcomes for this participant including objective and subjective assessment.
257
+ - `demographics`: Demographic attributes of the participant in terms of age, gender, ethnicity, and education.
258
+ - `personality`: Personality attributes for this participant, in terms of Big-5 and Social Value Orientation
259
+ - `mturk_agent_2`: For the second participant; follows the same structure as above
260
+ - `annotations`: Strategy annotations for each utterance in the dialogue, wherever available. The first element represents the utterance and the second represents a comma-separated list of all strategies present in that utterance.
261
+
262
+ ### Data Splits
263
+
264
+ No default data split has been provided. Hence, all 1030 data points are under the 'train' split.
265
+
266
+ | | Train |
267
+ | ----- | ----- |
268
+ | total dialogues | 1030 |
269
+ | annotated dialogues | 396 |
270
+
271
+ ## Dataset Creation
272
+
273
+ ### Curation Rationale
274
+
275
+ The dataset was collected to address the limitations in prior negotiation datasets from the perspective of downstream applications in pedagogy and conversational AI. Please refer to the original paper published at NAACL 2021 for details about the rationale and data curation steps ([source paper](https://aclanthology.org/2021.naacl-main.254.pdf)).
276
+
277
+ ### Source Data
278
+
279
+ #### Initial Data Collection and Normalization
280
+
281
+ The dialogues were crowdsourced on Amazon Mechanical Turk. The strategy annotations were performed by expert annotators (first three authors of the paper). Please refer to the original dataset paper published at NAACL 2021 for more details ([source paper](https://aclanthology.org/2021.naacl-main.254.pdf)).
282
+
283
+ #### Who are the source language producers?
284
+
285
+ The primary producers are Turkers on Amazon Mechanical Turk platform. Two turkers were randomly paired with each other to engage in a negotiation via a chat interface. Please refer to the original dataset paper published at NAACL 2021 for more details ([source paper](https://aclanthology.org/2021.naacl-main.254.pdf)).
286
+
287
+ ### Annotations
288
+
289
+ #### Annotation process
290
+
291
+ From the [source paper](https://aclanthology.org/2021.naacl-main.254.pdf) for this dataset:
292
+
293
+ >Three expert annotators independently annotated 396 dialogues containing 4615 utterances. The annotation guidelines were iterated over a subset of 5 dialogues, while the reliability scores were computed on a different subset of 10 dialogues. We use the nominal form of Krippendorff’s alpha (Krippendorff, 2018) to measure the inter-annotator agreement. We provide the annotation statistics in Table 2. Although we release all the annotations, we skip Coordination and Empathy for our analysis in this work, due to higher subjectivity resulting in relatively lower reliability scores.
294
+
295
+ #### Who are the annotators?
296
+
297
+ Three expert annotators (first three authors of the paper).
298
+
299
+ ### Personal and Sensitive Information
300
+
301
+ All personally identifiable information about the participants such as MTurk Ids or HIT Ids was removed before releasing the data.
302
+
303
+ ## Considerations for Using the Data
304
+
305
+ ### Social Impact of Dataset
306
+
307
+ Please refer to Section 8.2 in the [source paper](https://aclanthology.org/2021.naacl-main.254.pdf).
308
+
309
+ ### Discussion of Biases
310
+
311
+ Please refer to Section 8.2 in the [source paper](https://aclanthology.org/2021.naacl-main.254.pdf).
312
+
313
+ ### Other Known Limitations
314
+
315
+ Please refer to Section 7 in the [source paper](https://aclanthology.org/2021.naacl-main.254.pdf).
316
+
317
+ ## Additional Information
318
+
319
+ ### Dataset Curators
320
+
321
+ Corresponding Author: Kushal Chawla (`kchawla@usc.edu`)\
322
+ Affiliation: University of Southern California\
323
+ Please refer to the [source paper](https://aclanthology.org/2021.naacl-main.254.pdf) for the complete author list.
324
+
325
+ ### Licensing Information
326
+
327
+ The project is licensed under CC-by-4.0
328
+
329
+ ### Citation Information
330
+ ```
331
+ @inproceedings{chawla2021casino,
332
+ title={CaSiNo: A Corpus of Campsite Negotiation Dialogues for Automatic Negotiation Systems},
333
+ author={Chawla, Kushal and Ramirez, Jaysa and Clever, Rene and Lucas, Gale and May, Jonathan and Gratch, Jonathan},
334
+ booktitle={Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies},
335
+ pages={3167--3185},
336
+ year={2021}
337
+ }
338
+ ```
339
+
340
+ ### Contributions
341
+
342
+ Thanks to [Kushal Chawla](https://kushalchawla.github.io/) for adding this dataset.
huggingface_dataset/Dataset_Card/huggingartists_billy-talent.md ADDED
@@ -0,0 +1,204 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ tags:
5
+ - huggingartists
6
+ - lyrics
7
+ ---
8
+
9
+ # Dataset Card for "huggingartists/billy-talent"
10
+
11
+ ## Table of Contents
12
+ - [Dataset Description](#dataset-description)
13
+ - [Dataset Summary](#dataset-summary)
14
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
15
+ - [Languages](#languages)
16
+ - [How to use](#how-to-use)
17
+ - [Dataset Structure](#dataset-structure)
18
+ - [Data Fields](#data-fields)
19
+ - [Data Splits](#data-splits)
20
+ - [Dataset Creation](#dataset-creation)
21
+ - [Curation Rationale](#curation-rationale)
22
+ - [Source Data](#source-data)
23
+ - [Annotations](#annotations)
24
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
25
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
26
+ - [Social Impact of Dataset](#social-impact-of-dataset)
27
+ - [Discussion of Biases](#discussion-of-biases)
28
+ - [Other Known Limitations](#other-known-limitations)
29
+ - [Additional Information](#additional-information)
30
+ - [Dataset Curators](#dataset-curators)
31
+ - [Licensing Information](#licensing-information)
32
+ - [Citation Information](#citation-information)
33
+ - [About](#about)
34
+
35
+ ## Dataset Description
36
+
37
+ - **Homepage:** [https://github.com/AlekseyKorshuk/huggingartists](https://github.com/AlekseyKorshuk/huggingartists)
38
+ - **Repository:** [https://github.com/AlekseyKorshuk/huggingartists](https://github.com/AlekseyKorshuk/huggingartists)
39
+ - **Paper:** [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
40
+ - **Point of Contact:** [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
41
+ - **Size of the generated dataset:** 0.222716 MB
42
+
43
+
44
+ <div class="inline-flex flex-col" style="line-height: 1.5;">
45
+ <div class="flex">
46
+ <div style="display:DISPLAY_1; margin-left: auto; margin-right: auto; width: 92px; height:92px; border-radius: 50%; background-size: cover; background-image: url(&#39;https://images.genius.com/66f0650a5d8acadaed4292d6e3df6b9b.1000x1000x1.jpg&#39;)">
47
+ </div>
48
+ </div>
49
+ <a href="https://huggingface.co/huggingartists/billy-talent">
50
+ <div style="text-align: center; margin-top: 3px; font-size: 16px; font-weight: 800">🤖 HuggingArtists Model 🤖</div>
51
+ </a>
52
+ <div style="text-align: center; font-size: 16px; font-weight: 800">Billy Talent</div>
53
+ <a href="https://genius.com/artists/billy-talent">
54
+ <div style="text-align: center; font-size: 14px;">@billy-talent</div>
55
+ </a>
56
+ </div>
57
+
58
+ ### Dataset Summary
59
+
60
+ The Lyrics dataset parsed from Genius. This dataset is designed to generate lyrics with HuggingArtists.
61
+ Model is available [here](https://huggingface.co/huggingartists/billy-talent).
62
+
63
+ ### Supported Tasks and Leaderboards
64
+
65
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
66
+
67
+ ### Languages
68
+
69
+ en
70
+
71
+ ## How to use
72
+
73
+ How to load this dataset directly with the datasets library:
74
+
75
+ ```python
76
+ from datasets import load_dataset
77
+
78
+ dataset = load_dataset("huggingartists/billy-talent")
79
+ ```
80
+
81
+ ## Dataset Structure
82
+
83
+ An example of 'train' looks as follows.
84
+ ```
85
+ This example was too long and was cropped:
86
+
87
+ {
88
+ "text": "Look, I was gonna go easy on you\nNot to hurt your feelings\nBut I'm only going to get this one chance\nSomething's wrong, I can feel it..."
89
+ }
90
+ ```
91
+
92
+ ### Data Fields
93
+
94
+ The data fields are the same among all splits.
95
+
96
+ - `text`: a `string` feature.
97
+
98
+
99
+ ### Data Splits
100
+
101
+ | train |validation|test|
102
+ |------:|---------:|---:|
103
+ |122| -| -|
104
+
105
+ 'Train' can be easily divided into 'train' & 'validation' & 'test' with few lines of code:
106
+
107
+ ```python
108
+ from datasets import load_dataset, Dataset, DatasetDict
109
+ import numpy as np
110
+
111
+ datasets = load_dataset("huggingartists/billy-talent")
112
+
113
+ train_percentage = 0.9
114
+ validation_percentage = 0.07
115
+ test_percentage = 0.03
116
+
117
+ train, validation, test = np.split(datasets['train']['text'], [int(len(datasets['train']['text'])*train_percentage), int(len(datasets['train']['text'])*(train_percentage + validation_percentage))])
118
+
119
+ datasets = DatasetDict(
120
+ {
121
+ 'train': Dataset.from_dict({'text': list(train)}),
122
+ 'validation': Dataset.from_dict({'text': list(validation)}),
123
+ 'test': Dataset.from_dict({'text': list(test)})
124
+ }
125
+ )
126
+ ```
127
+
128
+ ## Dataset Creation
129
+
130
+ ### Curation Rationale
131
+
132
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
133
+
134
+ ### Source Data
135
+
136
+ #### Initial Data Collection and Normalization
137
+
138
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
139
+
140
+ #### Who are the source language producers?
141
+
142
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
143
+
144
+ ### Annotations
145
+
146
+ #### Annotation process
147
+
148
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
149
+
150
+ #### Who are the annotators?
151
+
152
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
153
+
154
+ ### Personal and Sensitive Information
155
+
156
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
157
+
158
+ ## Considerations for Using the Data
159
+
160
+ ### Social Impact of Dataset
161
+
162
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
163
+
164
+ ### Discussion of Biases
165
+
166
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
167
+
168
+ ### Other Known Limitations
169
+
170
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
171
+
172
+ ## Additional Information
173
+
174
+ ### Dataset Curators
175
+
176
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
177
+
178
+ ### Licensing Information
179
+
180
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
181
+
182
+ ### Citation Information
183
+
184
+ ```
185
+ @InProceedings{huggingartists,
186
+ author={Aleksey Korshuk}
187
+ year=2021
188
+ }
189
+ ```
190
+
191
+
192
+ ## About
193
+
194
+ *Built by Aleksey Korshuk*
195
+
196
+ [![Follow](https://img.shields.io/github/followers/AlekseyKorshuk?style=social)](https://github.com/AlekseyKorshuk)
197
+
198
+ [![Follow](https://img.shields.io/twitter/follow/alekseykorshuk?style=social)](https://twitter.com/intent/follow?screen_name=alekseykorshuk)
199
+
200
+ [![Follow](https://img.shields.io/badge/dynamic/json?color=blue&label=Telegram%20Channel&query=%24.result&url=https%3A%2F%2Fapi.telegram.org%2Fbot1929545866%3AAAFGhV-KKnegEcLiyYJxsc4zV6C-bdPEBtQ%2FgetChatMemberCount%3Fchat_id%3D-1001253621662&style=social&logo=telegram)](https://t.me/joinchat/_CQ04KjcJ-4yZTky)
201
+
202
+ For more details, visit the project repository.
203
+
204
+ [![GitHub stars](https://img.shields.io/github/stars/AlekseyKorshuk/huggingartists?style=social)](https://github.com/AlekseyKorshuk/huggingartists)
huggingface_dataset/Dataset_Card/huggingnft_cryptoadz-by-gremplin.md ADDED
@@ -0,0 +1,175 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - huggingnft
4
+ - nft
5
+ - huggan
6
+ - gan
7
+ - image
8
+ - images
9
+ task:
10
+ - unconditional-image-generation
11
+ datasets:
12
+ - huggingnft/cryptoadz-by-gremplin
13
+ license: mit
14
+ ---
15
+
16
+ # Dataset Card
17
+
18
+ ## Disclaimer
19
+
20
+ All rights belong to their owners.
21
+ Models and datasets can be removed from the site at the request of the copyright holder.
22
+
23
+ ## Table of Contents
24
+ - [Dataset Description](#dataset-description)
25
+ - [Dataset Summary](#dataset-summary)
26
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
27
+ - [Languages](#languages)
28
+ - [How to use](#how-to-use)
29
+ - [Dataset Structure](#dataset-structure)
30
+ - [Data Fields](#data-fields)
31
+ - [Data Splits](#data-splits)
32
+ - [Dataset Creation](#dataset-creation)
33
+ - [Curation Rationale](#curation-rationale)
34
+ - [Source Data](#source-data)
35
+ - [Annotations](#annotations)
36
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
37
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
38
+ - [Social Impact of Dataset](#social-impact-of-dataset)
39
+ - [Discussion of Biases](#discussion-of-biases)
40
+ - [Other Known Limitations](#other-known-limitations)
41
+ - [Additional Information](#additional-information)
42
+ - [Dataset Curators](#dataset-curators)
43
+ - [Licensing Information](#licensing-information)
44
+ - [Citation Information](#citation-information)
45
+ - [About](#about)
46
+
47
+ ## Dataset Description
48
+
49
+ - **Homepage:** [https://github.com/AlekseyKorshuk/huggingnft](https://github.com/AlekseyKorshuk/huggingnft)
50
+ - **Repository:** [https://github.com/AlekseyKorshuk/huggingnft](https://github.com/AlekseyKorshuk/huggingnft)
51
+ - **Paper:** [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
52
+ - **Point of Contact:** [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
53
+
54
+
55
+ ### Dataset Summary
56
+
57
+ NFT images dataset for unconditional generation.
58
+
59
+ NFT collection available [here](https://opensea.io/collection/cryptoadz-by-gremplin).
60
+
61
+ Model is available [here](https://huggingface.co/huggingnft/cryptoadz-by-gremplin).
62
+
63
+ Check Space: [link](https://huggingface.co/spaces/AlekseyKorshuk/huggingnft).
64
+
65
+ ### Supported Tasks and Leaderboards
66
+
67
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
68
+
69
+
70
+ ## How to use
71
+
72
+ How to load this dataset directly with the datasets library:
73
+
74
+ ```python
75
+ from datasets import load_dataset
76
+
77
+ dataset = load_dataset("huggingnft/cryptoadz-by-gremplin")
78
+ ```
79
+
80
+ ## Dataset Structure
81
+
82
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
83
+
84
+
85
+ ### Data Fields
86
+
87
+ The data fields are the same among all splits.
88
+
89
+ - `image`: an `image` feature.
90
+ - `id`: an `int` feature.
91
+ - `token_metadata`: a `str` feature.
92
+ - `image_original_url`: a `str` feature.
93
+
94
+ ### Data Splits
95
+
96
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
97
+
98
+
99
+ ## Dataset Creation
100
+
101
+ ### Curation Rationale
102
+
103
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
104
+
105
+ ### Source Data
106
+
107
+ #### Initial Data Collection and Normalization
108
+
109
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
110
+
111
+ #### Who are the source language producers?
112
+
113
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
114
+
115
+ ### Annotations
116
+
117
+ #### Annotation process
118
+
119
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
120
+
121
+ #### Who are the annotators?
122
+
123
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
124
+
125
+ ### Personal and Sensitive Information
126
+
127
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
128
+
129
+ ## Considerations for Using the Data
130
+
131
+ ### Social Impact of Dataset
132
+
133
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
134
+
135
+ ### Discussion of Biases
136
+
137
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
138
+
139
+ ### Other Known Limitations
140
+
141
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
142
+
143
+ ## Additional Information
144
+
145
+ ### Dataset Curators
146
+
147
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
148
+
149
+ ### Licensing Information
150
+
151
+ [More Information Needed](https://github.com/huggingface/datasets/blob/master/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
152
+
153
+ ### Citation Information
154
+
155
+ ```
156
+ @InProceedings{huggingnft,
157
+ author={Aleksey Korshuk}
158
+ year=2022
159
+ }
160
+ ```
161
+
162
+
163
+ ## About
164
+
165
+ *Built by Aleksey Korshuk*
166
+
167
+ [![Follow](https://img.shields.io/github/followers/AlekseyKorshuk?style=social)](https://github.com/AlekseyKorshuk)
168
+
169
+ [![Follow](https://img.shields.io/twitter/follow/alekseykorshuk?style=social)](https://twitter.com/intent/follow?screen_name=alekseykorshuk)
170
+
171
+ [![Follow](https://img.shields.io/badge/dynamic/json?color=blue&label=Telegram%20Channel&query=%24.result&url=https%3A%2F%2Fapi.telegram.org%2Fbot1929545866%3AAAFGhV-KKnegEcLiyYJxsc4zV6C-bdPEBtQ%2FgetChatMemberCount%3Fchat_id%3D-1001253621662&style=social&logo=telegram)](https://t.me/joinchat/_CQ04KjcJ-4yZTky)
172
+
173
+ For more details, visit the project repository.
174
+
175
+ [![GitHub stars](https://img.shields.io/github/stars/AlekseyKorshuk/huggingnft?style=social)](https://github.com/AlekseyKorshuk/huggingnft)
huggingface_dataset/Dataset_Card/macavaney_d2q-msmarco-passage-scores-monot5.md ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - no-annotation
4
+ language: []
5
+ language_creators:
6
+ - machine-generated
7
+ license: []
8
+ pretty_name: Doc2Query monoT5 Relevance Scores for `msmarco-passage`
9
+ source_datasets: [msmarco-passage]
10
+ tags:
11
+ - document-expansion
12
+ - doc2query--
13
+ task_categories:
14
+ - text-retrieval
15
+ task_ids:
16
+ - document-retrieval
17
+ viewer: false
18
+ ---
19
+
20
+ # Doc2Query monoT5 Relevance Scores for `msmarco-passage`
21
+
22
+ This dataset provides the pre-computed query relevance scores for the [`msmarco-passage`](https://ir-datasets.com/msmarco-passage) dataset,
23
+ for use with Doc2Query--.
24
+
25
+ The generated queries come from [`macavaney/d2q-msmarco-passage`](https://huggingface.co/datasets/macavaney/d2q-msmarco-passage) and
26
+ were scored with [`castorini/monot5-base-msmarco`](https://huggingface.co/castorini/monot5-base-msmarco).
27
+
28
+ ## Getting started
29
+
30
+ This artefact is meant to be used with the [`pyterrier_doc2query`](https://github.com/terrierteam/pyterrier_doc2query) pacakge. It can
31
+ be installed as:
32
+
33
+ ```bash
34
+ pip install git+https://github.com/terrierteam/pyterrier_doc2query
35
+ ```
36
+
37
+ Depending on what you are using this aretefact for, you may also need the following additional packages:
38
+
39
+ ```bash
40
+ pip install git+https://github.com/terrierteam/pyterrier_pisa # for indexing / retrieval
41
+ pip install git+https://github.com/terrierteam/pyterrier_t5 # for reproducing this aretefact
42
+ ```
43
+
44
+ ## Using this artefact
45
+
46
+ The main use case is to use this aretefact in a Doc2Query&minus;&minus; indexing pipeline:
47
+
48
+ ```python
49
+ import pyterrier as pt ; pt.init()
50
+ from pyterrier_pisa import PisaIndex
51
+ from pyterrier_doc2query import QueryScoreStore, QueryFilter
52
+
53
+ store = QueryScoreStore.from_repo('https://huggingface.co/datasets/macavaney/d2q-msmarco-passage-scores-monot5')
54
+ index = PisaIndex('path/to/index')
55
+ pipeline = store.query_scorer(limit_k=40) >> QueryFilter(t=store.percentile(70)) >> index
56
+
57
+ dataset = pt.get_dataset('irds:msmarco-passage')
58
+ pipeline.index(dataset.get_corpus_iter())
59
+ ```
60
+
61
+ You can also use the store directly as a dataset to look up or iterate over the data:
62
+
63
+ ```python
64
+ store.lookup('100')
65
+ # {'querygen': ..., 'querygen_store': ...}
66
+ for record in store:
67
+ pass
68
+ ```
69
+
70
+ ## Reproducing this aretefact
71
+
72
+ This aretefact can be reproduced using the following pipeline:
73
+
74
+ ```python
75
+ import pyterrier as pt ; pt.init()
76
+ from pyterrier_t5 import MonoT5ReRanker
77
+ from pyterrier_doc2query import Doc2QueryStore, QueryScoreStore, QueryScorer
78
+
79
+ doc2query_generator = Doc2QueryStore.from_repo('https://huggingface.co/datasets/macavaney/d2q-msmarco-passage').generator()
80
+ store = QueryScoreStore('path/to/store')
81
+ pipeline = doc2query_generator >> QueryScorer(MonoT5ReRanker()) >> store
82
+
83
+ dataset = pt.get_dataset('irds:msmarco-passage')
84
+ pipeline.index(dataset.get_corpus_iter())
85
+ ```
86
+
87
+ Note that this process will take quite some time; it computes the relevance score for 80 generated queries
88
+ for every document in the dataset.
huggingface_dataset/Dataset_Card/nlpso_m0_qualitative_analysis_ref_cmbert_io.md ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - fr
4
+ multilinguality:
5
+ - monolingual
6
+ task_categories:
7
+ - token-classification
8
+ ---
9
+
10
+ # m0_qualitative_analysis_ref_cmbert_io
11
+
12
+ ## Introduction
13
+
14
+ This dataset was used to perform **qualitative analysis** of [Jean-Baptiste/camembert-ner](https://huggingface.co/Jean-Baptiste/camembert-ner) on **flat NER task** using Flat NER approach [M0].
15
+ It contains 19th-century Paris trade directories' entries.
16
+
17
+ ## Dataset parameters
18
+
19
+ * Approach : M0
20
+ * Dataset type : ground-truth
21
+ * Tokenizer : [Jean-Baptiste/camembert-ner](https://huggingface.co/Jean-Baptiste/camembert-ner)
22
+ * Tagging format : IO
23
+ * Counts :
24
+ * Train : 6084
25
+ * Dev : 676
26
+ * Test : 1685
27
+ * Associated fine-tuned model : [nlpso/m0_flat_ner_ref_cmbert_io](https://huggingface.co/nlpso/m0_flat_ner_ref_cmbert_io)
28
+
29
+ ## Entity types
30
+
31
+ Abbreviation|Description
32
+ -|-
33
+ O |Outside of a named entity
34
+ PER |Person or company name
35
+ ACT |Person or company professional activity
36
+ TITRE |Distinction
37
+ LOC |Street name
38
+ CARDINAL |Street number
39
+ FT |Geographical feature
40
+
41
+ ## How to use this dataset
42
+
43
+ ```python
44
+ from datasets import load_dataset
45
+
46
+ train_dev_test = load_dataset("nlpso/m0_qualitative_analysis_ref_cmbert_io")
huggingface_dataset/Dataset_Card/nlpso_m0_qualitative_analysis_ref_ptrn_cmbert_io.md ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - fr
4
+ multilinguality:
5
+ - monolingual
6
+ task_categories:
7
+ - token-classification
8
+ ---
9
+
10
+ # m0_qualitative_analysis_ref_ptrn_cmbert_io
11
+
12
+ ## Introduction
13
+
14
+ This dataset was used to perform **qualitative analysis** of [HueyNemud/das22-10-camembert_pretrained](https://huggingface.co/HueyNemud/das22-10-camembert_pretrained) on **flat NER task** using Flat NER approach [M0].
15
+ It contains 19th-century Paris trade directories' entries.
16
+
17
+ ## Dataset parameters
18
+
19
+ * Approach : M0
20
+ * Dataset type : ground-truth
21
+ * Tokenizer : [HueyNemud/das22-10-camembert_pretrained](https://huggingface.co/HueyNemud/das22-10-camembert_pretrained)
22
+ * Tagging format : IO
23
+ * Counts :
24
+ * Train : 6084
25
+ * Dev : 676
26
+ * Test : 1685
27
+ * Associated fine-tuned model : [nlpso/m0_flat_ner_ref_ptrn_cmbert_io](https://huggingface.co/nlpso/m0_flat_ner_ref_ptrn_cmbert_io)
28
+
29
+ ## Entity types
30
+
31
+ Abbreviation|Description
32
+ -|-
33
+ O |Outside of a named entity
34
+ PER |Person or company name
35
+ ACT |Person or company professional activity
36
+ TITRE |Distinction
37
+ LOC |Street name
38
+ CARDINAL |Street number
39
+ FT |Geographical feature
40
+
41
+ ## How to use this dataset
42
+
43
+ ```python
44
+ from datasets import load_dataset
45
+
46
+ train_dev_test = load_dataset("nlpso/m0_qualitative_analysis_ref_ptrn_cmbert_io")
huggingface_dataset/Dataset_Card/nntadotzip_iuQAchatbot.md ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ annotations_creators:
2
+ - crowdsourced
3
+ language_creators:
4
+ - crowdsourced
5
+ - found
6
+ languages:
7
+ - en
8
+ licenses:
9
+ - cc-by-4.0
10
+ multilinguality:
11
+ - monolingual
12
+ paperswithcode_id: squad
13
+ pretty_name: SQuAD
14
+ size_categories:
15
+ - 10K<n<100K
16
+ source_datasets:
17
+ - extended|wikipedia
18
+ task_categories:
19
+ - question-answering
20
+ task_ids:
21
+ - extractive-qa
huggingface_dataset/Dataset_Card/pietrolesci_robust_nli.md ADDED
@@ -0,0 +1,178 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ## Overview
2
+ Original dataset is available in the original [Github repo](https://github.com/tyliupku/nli-debiasing-datasets).
3
+
4
+ This dataset is a collection of NLI benchmarks constructed as described in the paper
5
+ [An Empirical Study on Model-agnostic Debiasing Strategies for Robust Natural Language Inference](https://aclanthology.org/2020.conll-1.48/)
6
+ published at CoNLL 2020.
7
+
8
+
9
+ ## Dataset curation
10
+ No specific curation for this dataset. Label encoding follows exactly what is reported in the paper by the authors.
11
+ Also, from the paper:
12
+
13
+ > _all the following datasets are collected based on the public available resources proposed by their authors, thus the experimental results in this paper are comparable to the numbers reported in the original papers and the other papers that use these datasets_
14
+
15
+ Most of the datasets included follow the custom 3-class NLI convention `{"entailment": 0, "neutral": 1, "contradiction": 2}`.
16
+ However, the following datasets have a particular label mapping
17
+
18
+ - `IS-SD`: `{"non-entailment": 0, "entailment": 1}`
19
+
20
+ - `LI_TS`: `{"non-contradiction": 0, "contradiction": 1}`
21
+
22
+
23
+ ## Dataset structure
24
+ This benchmark dataset includes 10 adversarial datasets. To provide more insights on how the adversarial
25
+ datasets attack the models, the authors categorized them according to the bias(es) they test and they renamed
26
+ them accordingly. More details in section 2 of the paper.
27
+ A mapping with the original dataset names is provided below
28
+
29
+ | | Name | Original Name | Original Paper | Original Curation |
30
+ |---:|:-------|:-----------------------|:--------------------------------------------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
31
+ | 0 | PI-CD | SNLI-Hard | [Gururangan et al. (2018)](https://aclanthology.org/N18-2017/) | SNLI test sets instances that cannot be correctly classified by a neural classifier (fastText) trained on only the hypothesis sentences. |
32
+ | 1 | PI-SP | MNLI-Hard | [Liu et al. (2020)](https://aclanthology.org/2020.lrec-1.846/) | MNLI-mismatched dev sets instances that cannot be correctly classified by surface patterns that are highly correlated with the labels. |
33
+ | 2 | IS-SD | HANS | [McCoy et al. (2019)](https://aclanthology.org/P19-1334/) | Dataset that tests lexical overlap, subsequence, and constituent heuristics between the hypothesis and premises sentences. |
34
+ | 3 | IS-CS | SoSwap-AddAMod | [Nie et al. (2019)](https://dl.acm.org/doi/abs/10.1609/aaai.v33i01.33016867) | Pairs of sentences whose logical relations cannot be extracted from lexical information alone. Premise are taken from SNLI dev set and modified. The original paper assigns a Lexically Misleading Scores (LMS) to each instance. Here, only the subset with LMS > 0.7 is reported. |
35
+ | 4 | LI-LI | Stress tests (antonym) | [Naik et al. (2018)](https://aclanthology.org/C18-1198/) and [Glockner et al. (2018)](https://aclanthology.org/P18-2103/) | Merge of the 'antonym' category in Naik et al. (2018) (from MNLI matched and mismatched dev sets) and Glockner et al. (2018) (SNLI training set). |
36
+ | 5 | LI-TS | Created by the authors | Created by the authors | Swap the two sentences in the original MultiNLI mismatched dev sets. If the gold label is 'contradiction', the corresponding label in the swapped instance remains unchanged, otherwise it becomes 'non-contradicted'. |
37
+ | 6 | ST-WO | Word overlap | [Naik et al. (2018)](https://aclanthology.org/C18-1198/) | 'Word overlap' category in Naik et al. (2018). |
38
+ | 7 | ST-NE | Negation | [Naik et al. (2018)](https://aclanthology.org/C18-1198/) | 'Negation' category in Naik et al. (2018). |
39
+ | 8 | ST-LM | Length mismatch | [Naik et al. (2018)](https://aclanthology.org/C18-1198/) | 'Length mismatch' category in Naik et al. (2018). |
40
+ | 9 | ST-SE | Spelling errors | [Naik et al. (2018)](https://aclanthology.org/C18-1198/) | 'Spelling errors' category in Naik et al. (2018). |
41
+
42
+ ## Code to create the dataset
43
+
44
+ ```python
45
+
46
+ import pandas as pd
47
+ from datasets import Dataset, ClassLabel, Value, Features, DatasetDict
48
+
49
+
50
+ Tri_dataset = ["IS_CS", "LI_LI", "PI_CD", "PI_SP", "ST_LM", "ST_NE", "ST_SE", "ST_WO"]
51
+ Ent_bin_dataset = ["IS_SD"]
52
+ Con_bin_dataset = ["LI_TS"]
53
+
54
+
55
+ # read data
56
+ with open("<path to file>/robust_nli.txt", encoding="utf-8", mode="r") as fl:
57
+ f = fl.read().strip().split("\n")
58
+ f = [eval(i) for i in f]
59
+ df = pd.DataFrame.from_dict(f)
60
+
61
+ # rename to map common names
62
+ df = df.rename(columns={"prem": "premise", "hypo": "hypothesis"})
63
+
64
+ # reorder columns
65
+ df = df.loc[:, ["idx", "split", "premise", "hypothesis", "label"]]
66
+
67
+ # create split-specific features
68
+ Tri_features = Features(
69
+ {
70
+ "idx": Value(dtype="int64"),
71
+ "premise": Value(dtype="string"),
72
+ "hypothesis": Value(dtype="string"),
73
+ "label": ClassLabel(num_classes=3, names=["entailment", "neutral", "contradiction"]),
74
+ }
75
+ )
76
+
77
+ Ent_features = Features(
78
+ {
79
+ "idx": Value(dtype="int64"),
80
+ "premise": Value(dtype="string"),
81
+ "hypothesis": Value(dtype="string"),
82
+ "label": ClassLabel(num_classes=2, names=["non-entailment", "entailment"]),
83
+ }
84
+ )
85
+
86
+ Con_features = Features(
87
+ {
88
+ "idx": Value(dtype="int64"),
89
+ "premise": Value(dtype="string"),
90
+ "hypothesis": Value(dtype="string"),
91
+ "label": ClassLabel(num_classes=2, names=["non-contradiction", "contradiction"]),
92
+ }
93
+ )
94
+
95
+ # convert to datasets
96
+ dataset_splits = {}
97
+
98
+ for split in df["split"].unique():
99
+ print(split)
100
+ df_split = df.loc[df["split"] == split].copy()
101
+
102
+ if split in Tri_dataset:
103
+ df_split["label"] = df_split["label"].map({"entailment": 0, "neutral": 1, "contradiction": 2})
104
+ ds = Dataset.from_pandas(df_split, features=Tri_features)
105
+
106
+ elif split in Ent_bin_dataset:
107
+ df_split["label"] = df_split["label"].map({"non-entailment": 0, "entailment": 1})
108
+ ds = Dataset.from_pandas(df_split, features=Ent_features)
109
+
110
+ elif split in Con_bin_dataset:
111
+ df_split["label"] = df_split["label"].map({"non-contradiction": 0, "contradiction": 1})
112
+ ds = Dataset.from_pandas(df_split, features=Con_features)
113
+
114
+ else:
115
+ print("ERROR:", split)
116
+ dataset_splits[split] = ds
117
+ datasets = DatasetDict(dataset_splits)
118
+ datasets.push_to_hub("pietrolesci/robust_nli", token="<your token>")
119
+
120
+
121
+ # check overlap between splits
122
+ from itertools import combinations
123
+ for i, j in combinations(datasets.keys(), 2):
124
+ print(
125
+ f"{i} - {j}: ",
126
+ pd.merge(
127
+ datasets[i].to_pandas(),
128
+ datasets[j].to_pandas(),
129
+ on=["premise", "hypothesis", "label"],
130
+ how="inner",
131
+ ).shape[0],
132
+ )
133
+ #> PI_SP - ST_LM: 0
134
+ #> PI_SP - ST_NE: 0
135
+ #> PI_SP - IS_CS: 0
136
+ #> PI_SP - LI_TS: 1
137
+ #> PI_SP - LI_LI: 0
138
+ #> PI_SP - ST_SE: 0
139
+ #> PI_SP - PI_CD: 0
140
+ #> PI_SP - IS_SD: 0
141
+ #> PI_SP - ST_WO: 0
142
+ #> ST_LM - ST_NE: 0
143
+ #> ST_LM - IS_CS: 0
144
+ #> ST_LM - LI_TS: 0
145
+ #> ST_LM - LI_LI: 0
146
+ #> ST_LM - ST_SE: 0
147
+ #> ST_LM - PI_CD: 0
148
+ #> ST_LM - IS_SD: 0
149
+ #> ST_LM - ST_WO: 0
150
+ #> ST_NE - IS_CS: 0
151
+ #> ST_NE - LI_TS: 0
152
+ #> ST_NE - LI_LI: 0
153
+ #> ST_NE - ST_SE: 0
154
+ #> ST_NE - PI_CD: 0
155
+ #> ST_NE - IS_SD: 0
156
+ #> ST_NE - ST_WO: 0
157
+ #> IS_CS - LI_TS: 0
158
+ #> IS_CS - LI_LI: 0
159
+ #> IS_CS - ST_SE: 0
160
+ #> IS_CS - PI_CD: 0
161
+ #> IS_CS - IS_SD: 0
162
+ #> IS_CS - ST_WO: 0
163
+ #> LI_TS - LI_LI: 0
164
+ #> LI_TS - ST_SE: 0
165
+ #> LI_TS - PI_CD: 0
166
+ #> LI_TS - IS_SD: 0
167
+ #> LI_TS - ST_WO: 0
168
+ #> LI_LI - ST_SE: 0
169
+ #> LI_LI - PI_CD: 0
170
+ #> LI_LI - IS_SD: 0
171
+ #> LI_LI - ST_WO: 0
172
+ #> ST_SE - PI_CD: 0
173
+ #> ST_SE - IS_SD: 0
174
+ #> ST_SE - ST_WO: 0
175
+ #> PI_CD - IS_SD: 0
176
+ #> PI_CD - ST_WO: 0
177
+ #> IS_SD - ST_WO: 0
178
+ ```
huggingface_dataset/Dataset_Card/polsum.md ADDED
@@ -0,0 +1,762 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - expert-generated
4
+ language_creators:
5
+ - expert-generated
6
+ language:
7
+ - pl
8
+ license:
9
+ - cc-by-3.0
10
+ multilinguality:
11
+ - monolingual
12
+ size_categories:
13
+ - n<1K
14
+ source_datasets:
15
+ - original
16
+ task_categories:
17
+ - summarization
18
+ task_ids:
19
+ - news-articles-summarization
20
+ paperswithcode_id: null
21
+ pretty_name: Polish Summaries Corpus
22
+ dataset_info:
23
+ features:
24
+ - name: id
25
+ dtype: string
26
+ - name: date
27
+ dtype: string
28
+ - name: title
29
+ dtype: string
30
+ - name: section
31
+ dtype: string
32
+ - name: authors
33
+ dtype: string
34
+ - name: body
35
+ dtype: string
36
+ - name: summaries
37
+ sequence:
38
+ - name: ratio
39
+ dtype: int32
40
+ - name: type
41
+ dtype: string
42
+ - name: author
43
+ dtype: string
44
+ - name: body
45
+ dtype: string
46
+ - name: spans
47
+ sequence:
48
+ - name: start
49
+ dtype: int32
50
+ - name: end
51
+ dtype: int32
52
+ - name: span_text
53
+ dtype: string
54
+ splits:
55
+ - name: train
56
+ num_bytes: 34787575
57
+ num_examples: 569
58
+ download_size: 6082812
59
+ dataset_size: 34787575
60
+ ---
61
+
62
+ # Dataset Card for Polish Summaries Corpus
63
+
64
+ ## Table of Contents
65
+ - [Dataset Description](#dataset-description)
66
+ - [Dataset Summary](#dataset-summary)
67
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
68
+ - [Languages](#languages)
69
+ - [Dataset Structure](#dataset-structure)
70
+ - [Data Instances](#data-instances)
71
+ - [Data Fields](#data-fields)
72
+ - [Data Splits](#data-splits)
73
+ - [Dataset Creation](#dataset-creation)
74
+ - [Curation Rationale](#curation-rationale)
75
+ - [Source Data](#source-data)
76
+ - [Annotations](#annotations)
77
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
78
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
79
+ - [Social Impact of Dataset](#social-impact-of-dataset)
80
+ - [Discussion of Biases](#discussion-of-biases)
81
+ - [Other Known Limitations](#other-known-limitations)
82
+ - [Additional Information](#additional-information)
83
+ - [Dataset Curators](#dataset-curators)
84
+ - [Licensing Information](#licensing-information)
85
+ - [Citation Information](#citation-information)
86
+ - [Contributions](#contributions)
87
+
88
+ ## Dataset Description
89
+
90
+ - **Homepage:** http://zil.ipipan.waw.pl/PolishSummariesCorpus
91
+ - **Repository:** http://zil.ipipan.waw.pl/PolishSummariesCorpus
92
+ - **Paper:** http://nlp.ipipan.waw.pl/Bib/ogro:kop:14:lrec.pdf
93
+ - **Leaderboard:** [Needs More Information]
94
+ - **Point of Contact:** [Mateusz Kopeć](http://zil.ipipan.waw.pl/MateuszKopec)
95
+
96
+ ### Dataset Summary
97
+
98
+ The Corpus contains a large number of manual summaries of news articles,
99
+ with many independently created summaries for a single text. Such approach is supposed to overcome the annotator bias, which is often described as a problem during the evaluation of the summarization algorithms against a single gold standard.
100
+
101
+ ### Supported Tasks and Leaderboards
102
+
103
+ [Needs More Information]
104
+
105
+ ### Languages
106
+
107
+ Polish
108
+
109
+ ## Dataset Structure
110
+
111
+ ### Data Instances
112
+
113
+ See below an example from the dataset. Detailed descriptions of the fields are provided in the following section.
114
+
115
+ ```
116
+ {'authors': 'Krystyna Forowicz',
117
+ 'body': "ROZMOWA\n\nProf. Krzysztof Ernst, kierownik Zakładu Optyki Instytutu Fizyki Doświadczalnej Uniwersytetu Warszawskiego\n\nLidarowe oczy\n\nRYS. MAREK KONECKI\n\nJutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar? \n\nPROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym.\n\nCzy to kosztowne urządzenie będzie służyło tylko naukowcom?\n\nTego typu lidar jest rzeczywiście drogi, kosztuje około miliona marek niemieckich. Jest to najnowsza generacja tego typu lidarów. DIAL - lidar absorbcji różnicowej jest urządzeniem inteligentnym, to znaczy potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, korzyść naukową - rozwijamy badania nad tym urządzeniem, staramy się m.in. rozszerzyć jego zastosowanie także na inne substancje występujące w atmosferze. I korzyść dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska. Nad lidarem pracują specjaliści od laserów i od komputerów. Współpracujemy z doskonałym laboratorium prof. Ludgera Wöste z Freie Universitat Berlin rozwijającym m.in. problematykę lidarową. Pakiet software'u wzbogacamy o nowe algorytmy, które potrafią lepiej i dokładniej rozszyfrowywać sygnał lidarowy, a w konsekwencji skażenia. Żeby przetworzyć tzw. sygnał lidarowy, czyli to co wraca po rozproszeniu światła do układu, i otrzymać rozsądne dane dotyczące rozkładu koncentracji - trzeba dokonać skomplikowanych operacji. \n\nBadania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej, dzięki której ten lidar u nas zaistniał i dla której w ramach naszych zobowiązań wykonujemy pomiary zanieczyszczeń nad naszą wspólną granicą. Zasadniczy koszt jego budowy pokryła uzyskana od Fundacji dotacja. Część pieniędzy przekazał też Narodowy Fundusz Ochrony Środowiska i Gospodarki Wodnej oraz Komitet Badań Naukowych.\n\nCzy wszystkie zanieczyszczenia będzie można wykryć za pomocą lidaru?\n\nNie ma takiego jednostkowego urządzenia, które by wykrywało i mierzyło wszystkie szkodliwe gazy w atmosferze łącznie z dostarczeniem informacji o ich rozkładzie. Ale np. obecnie prowadzimy badania mające na celu rozszerzenie możliwości lidaru o taką substancję jak fosgen. Tym szkodliwym gazem może być skażone powietrze w miastach, w których zlokalizowane są zakłady chemiczne, np. w Bydgoszczy pewne ilości fosgenu emitują Zakłady Chemiczne Organika- Zachem. \n\nLidar typu DIAL jest oparty na pomiarze absorbcji różnicowej, czyli muszą być zastosowane dwie wiązki laserowe o dwóch różnych długościach fali, z których jedna jest absorbowana, a druga nie jest absorbowana przez substancję, którą chcemy wykryć. Cząsteczki, które wykrywamy mają pasma absorbcji w bliskim nadfiolecie. Możemy np. badać zawartość ozonu w troposferze. Okazuje się bowiem, że o ile brak tego gazu w wysokich warstwach atmosfery powoduje groźny efekt cieplarniany, to jego nadmiar tuż nad Ziemią jest szkodliwy. Groźne są też substancje gazowe, jak np. tlenki azotu, będące następstwem spalin samochodowych. A samochodów przybywa.\n\nCzy stać nas będzie na prowadzenie pomiarów ozonu w miastach? \n\nKoszt jednego dnia kampanii pomiarowej firmy zachodnie szacują na kilka tysięcy DM. Potrzebne są pieniądze na utrzymanie lidaru, na prowadzenie badań. Nasze przedsięwzięcie nie ma charakteru komercyjnego. Koszt pomiarów będzie znacznie niższy. Chcemy np. mierzyć w Warszawie rozkłady koncentracji tlenków azotu, ich ewolucję czasową nad różnymi arteriami miasta. Chcielibyśmy rozwinąć tutaj współpracę z państwowymi i wojewódzkimi służbami ochrony środowiska. Tego typu badania były prowadzone np. w Lyonie. Okazało się, że najwięcej tlenków azotu występuje niekoniecznie tam gdzie są one produkowane, to znaczy nie przy najruchliwszych ulicach, jeśli są one dobrze wentylowane a gromadzą się one w małych uliczkach. Przede wszystkim jednak do końca tego roku zamierzamy zakończyć pomiary skażeń atmosferycznych nad granicą polsko-niemiecką. Koncentrujemy się głównie na Czarnym Trójkącie - obszarze u zbiegu trzech granic: Polski, Niemiec i Czech, do niedawna uważanym za najbardziej zdegradowany region w Europie. Prowadziliśmy pomiary w samym Turowie, gdzie elektrownia Turoszowska jest głównym źródłem emisji. W planie mamy Bogatynię, zagłębie miedziowe. \n\nW Czarnym Trójkącie istnieje wiele stacjonarnych stacji monitoringowych.\n\nNasz lidar ma większe możliwości niż stacje monitoringowe. Mierzy zanieczyszczenia nie tylko lokalnie, ale też ich rozkład w przestrzeni, z wysoką rozdzielczością przestrzenną i na odległość kilku kilometrów. Możemy zatem śledzić ewolucję rozprzestrzeniania się tych zanieczyszczeń, ich kierunek i zmiany spowodowane m.in. warunkami atmosferycznymi. Wyniki naszych pomiarów porównujemy z danymi uzyskanymi ze stacji monitoringowych. \n\nJak wypadł Czarny Trójkąt?\n\nKiedy występowaliśmy o finansowanie tego projektu do Fundacji Współpracy Polsko-Niemieckiej zanieczyszczenie powietrza w Czarnym Trójkącie było dużo większe niż obecnie i wszystko wskazuje na to, że będzie dalej spadać. Obecnie stężenie dwutlenku siarki jest na granicy naszych możliwości pomiarowych. Dla regionu Turoszowskiego to dobra wiadomość i dla stosunków polsko-niemieckich też.\n\nTypów lidarów jest wiele \n\nTen lidar pracuje w obszarze bliskiego nadfioletu i promieniowania widzialnego, które jest wynikiem wykorzystania drugiej lub trzeciej harmonicznej lasera szafirowego, pracującego na granicy czerwieni i podczerwieni. DIAL jest tym typem lidara, który dzisiaj ma zdecydowanie największe wzięcie w ochronie środowiska. Z lidarów korzysta meteorologia. W Stanach Zjednoczonych lidary umieszcza się na satelitach (program NASA). Określają na przestrzeni kilkudziesięciu kilometrów rozkłady temperatury, wilgotności, ciśnienia, a także prędkości wiatru. Wykrywają pojawianie się huraganów, a nawet mogą określać rozmiary oka tajfunu.\n\nIle takich urządzeń jest w Europie?\n\n- W Europie takich lidarów jak nasz jest zaledwie kilka. Większość z nich mierzy ozon, dwutlenek siarki i tlenek azotu. Wykrywanie toluenu i benzenu jest oryginalnym rozwiązaniem. Długość fali dla benzenu jest już na skraju możliwości widmowych. Nasz lidar typu DIAL jest najnowocześniejszym w Polsce. Ponadto jest lidarem ruchomym, zainstalowanym na samochodzie. Ale historia lidarów w naszym kraju jest dłuższa i zaczęła się na początku lat 60. Pierwsze próby prowadzone były w stacji geofizycznej PAN w Belsku, niedługo po skonstruowaniu pierwszego w świecie lasera rubinowego. Potem powstał lidar stacjonarny, również typu DIAL, w Gdańsku, a w Krakowie sodary - urządzenia oparte na falach akustycznych, wygodne np. do pomiarów szybkości wiatru. Lidar umieszczony na samochodzie i zbudowany w latach 80 na Politechnice Poznańskiej w perspektywie miał być lidarem typu DIAL.\n\nFizycy dotychczas nie zajmowali się ochroną środowiska?\n\nTaka specjalizacja powstala na Wydziale Fizyki UW dwa lata temu. Gwarancją sukcesu naszego programu dydaktyczno-badawczego jest udział w nim zakładów należących do Instytutu Fizyki Doświadczalnej UW, Pracowni Przetwarzania Informacji (zdjęć satelitarnych) Instytutu Geofizyki i, co bardzo ważne, współpraca z Freie Universität Berlin. Mamy również na UW Międzywydziałowe Studia Ochrony Środowiska i studentom przekazujemy informacje o lidarze i fizycznych metodach badania środowiska. Nasze działania dydaktyczne bardzo efektywnie wspiera NFOŚ.\n\nRozmawiała Krystyna Forowicz",
118
+ 'date': '1997-04-21',
119
+ 'id': '199704210011',
120
+ 'section': 'Nauka i Technika',
121
+ 'summaries': {'author': ['I',
122
+ 'I',
123
+ 'I',
124
+ 'C',
125
+ 'C',
126
+ 'C',
127
+ 'K',
128
+ 'K',
129
+ 'K',
130
+ 'G',
131
+ 'G',
132
+ 'G',
133
+ 'J',
134
+ 'J',
135
+ 'J'],
136
+ 'body': ['Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar?PROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym.Czy to kosztowne urządzenie będzie służyło tylko naukowcom? Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, naukową - rozwijamy badania nad tym urządzeniem. I korzyść dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska. Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej, dzięki której ten lidar u nas zaistniał i dla której w ramach naszych zobowiązań wykonujemy pomiary zanieczyszczeń nad naszą wspólną granicą.',
137
+ 'Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar?PROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym.Czy to kosztowne urządzenie będzie służyło tylko naukowcom? Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, naukową - rozwijamy badania nad tym urządzeniem. I korzyść dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska. Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej, dzięki której ten lidar u nas zaistniał i dla której w ramach naszych zobowiązań wykonujemy pomiary zanieczyszczeń nad naszą wspólną granicą. Czy wszystkie zanieczyszczenia będzie można wykryć za pomocą lidaru?Nie ma takiego jednostkowego urządzenia, które by wykrywało i mierzyło wszystkie szkodliwe gazy w atmosferze łącznie z dostarczeniem informacji o ich rozkładzie. Możemy np. badać zawartość ozonu w troposferze. W Europie takich lidarów jak nasz jest zaledwie kilka. Większość z nich mierzy ozon, dwutlenek siarki i tlenek azotu. Fizycy dotychczas nie zajmowali się ochroną środowiska?Taka specjalizacja powstala na Wydziale Fizyki UW dwa lata temu. Gwarancją sukcesu naszego programu dydaktyczno-badawczego jest udział w nim zakładów należących do Instytutu Fizyki Doświadczalnej UW, Pracowni Przetwarzania Informacji Instytutu Geofizyki i współpraca z Freie Universität Berlin.',
138
+ 'Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar?PROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym. Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej, dzięki której ten lidar u nas zaistniał.',
139
+ 'Jutro odbędzie sie pokaz nowego polskiego lidara typu DIAL. lidar Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. DIAL - lidar absorbcji różnicowej jest urządzeniem inteligentnym, to znaczy potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. Z lidara korzyść mamy potrójną: użyteczną, naukową I dydaktyczną. Żeby przetworzyć sygnał lidarowy, czyli to co wraca po rozproszeniu światła do układu, i otrzymać dane dotyczące rozkładu koncentracji - trzeba dokonać skomplikowanych operacji. muszą być zastosowane dwie wiązki laserowe o dwóch różnych długościach fali, z których jedna jest absorbowana, a druga nie jest absorbowana przez substancję, którą chcemy wykryć.',
140
+ 'Jutro odbędzie sie pokaz nowego polskiego lidara typu DIAL. lidar Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym. Jest to najnowsza generacja tego typu lidarów. DIAL - lidar absorbcji różnicowej jest urządzeniem inteligentnym, to znaczy potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, korzyść naukową - rozwijamy badania nad tym urządzeniem. I korzyść dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska. Żeby przetworzyć tzw. sygnał lidarowy, czyli to co wraca po rozproszeniu światła do układu, i otrzymać rozsądne dane dotyczące rozkładu koncentracji - trzeba dokonać skomplikowanych operacji. Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej, dzięki której ten lidar u nas zaistniał i dla której w ramach naszych zobowiązań wykonujemy pomiary zanieczyszczeń nad naszą wspólną granicą. Zasadniczy koszt jego budowy pokryła uzyskana od Fundacji dotacja. Część pieniędzy przekazał też Narodowy Fundusz Ochrony Środowiska i Gospodarki Wodnej oraz Komitet Badań Naukowych. Lidar typu DIAL jest oparty na pomiarze absorbcji różnicowej, czyli muszą być zastosowane dwie wiązki laserowe o dwóch różnych długościach fali, z których jedna jest absorbowana, a druga nie jest absorbowana przez substancję, którą chcemy wykryć.',
141
+ 'Jutro odbędzie sie pokaz nowego polskiego lidara typu DIAL. lidar Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. DIAL - lidar absorbcji różnicowej jest urządzeniem inteligentnym, to znaczy potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. Z lidara korzyść mamy potrójną: użyteczną, naukową I dydaktyczną.',
142
+ 'Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar? \nPROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym. DIAL - lidar absorbcji różnicowej jest urządzeniem inteligentnym, to znaczy potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, korzyść naukową - rozwijamy badania nad tym urządzeniem I korzyść dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.Nasz lidar ma większe możliwości niż stacje monitoringowe. Mierzy zanieczyszczenia nie tylko lokalnie, ale też ich rozkład w przestrzeni, z wysoką rozdzielczością przestrzenną i na odległość kilku kilometrów.',
143
+ 'Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar? \nPROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym.Tego typu lidar jest drogi, kosztuje około miliona marek niemieckich. DIAL - lidar absorbcji różnicowej jest urządzeniem inteligentnym, to znaczy potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, korzyść naukową - rozwijamy badania nad tym urządzeniem, staramy się m.in. rozszerzyć jego zastosowanie także na inne substancje występujące w atmosferze. I korzyść dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.Lidar typu DIAL jest oparty na pomiarze absorbcji różnicowej, czyli muszą być zastosowane dwie wiązki laserowe o dwóch różnych długościach fali, z których jedna jest absorbowana, a druga nie jest absorbowana przez substancję, którą chcemy wykryć. Cząsteczki, które wykrywamy mają pasma absorbcji w bliskim nadfiolecie.Nasz lidar ma większe możliwości niż stacje monitoringowe. Mierzy zanieczyszczenia nie tylko lokalnie, ale też ich rozkład w przestrzeni, z wysoką rozdzielczością przestrzenną i na odległość kilku kilometrów. Możemy zatem śledzić ewolucję rozprzestrzeniania się tych zanieczyszczeń, ich kierunek i zmiany spowodowane m.in. warunkami atmosferycznymi. Wyniki naszych pomiarów porównujemy z danymi uzyskanymi ze stacji monitoringowych.',
144
+ 'Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar? \nPROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, korzyść naukową i dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.',
145
+ 'Jutro odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar? \n\nPROF. KRZYSZTOF ERNST: urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.\nto najnowsza generacja tego typu lidarów. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. korzyść mamy potrójną: użyteczną, przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, naukową - rozwijamy badania nad urządzeniem I dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.\nNasze przedsięwzięcie nie ma charakteru komercyjnego. Chcemy np. mierzyć w Warszawie rozkłady koncentracji tlenków azotu. Koncentrujemy się głównie na Czarnym Trójkącie - obszarze u zbiegu granic: Polski, Niemiec i Czech, do niedawna uważanym za najbardziej zdegradowany region w Europie.',
146
+ 'Jutro odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar? \n\nPROF. KRZYSZTOF ERNST: urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.\n\nto kosztowne urządzenie będzie służyło tylko naukowcom?\n\nlidar jest rzeczywiście drogi. to najnowsza generacja tego typu lidarów. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. korzyść mamy potrójną: użyteczną, przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, naukową - rozwijamy badania nad tym urządzeniem I dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.\n\nCzy wszystkie zanieczyszczenia będzie można wykryć za pomocą lidaru?\n\nNie ma takiego jednostkowego urządzenia, które by wykrywało i mierzyło wszystkie szkodliwe gazy w atmosferze. Ale prowadzimy badania mające na celu rozszerzenie możliwości lidaru o taką substancję jak fosgen.\n\nstać nas będzie na prowadzenie pomiarów ozonu w miastach? \n\nNasze przedsięwzięcie nie ma charakteru komercyjnego. Chcemy np. mierzyć w Warszawie rozkłady koncentracji tlenków azotu, ich ewolucję czasową nad różnymi arteriami miasta. Koncentrujemy się głównie na Czarnym Trójkącie - obszarze u zbiegu granic: Polski, Niemiec i Czech, do niedawna uważanym za najbardziej zdegradowany region w Europie. zanieczyszczenie było dużo większe niż obecnie i wszystko wskazuje na to, że będzie dalej spadać.\nDIAL dzisiaj ma zdecydowanie największe wzięcie w ochronie środowiska. \n\nFizycy dotychczas nie zajmowali się ochroną środowiska?\n\nTaka specjalizacja powstala na Wydziale Fizyki UW dwa lata temu.',
147
+ 'Co to jest lidar? \n\nPROF. KRZYSZTOF ERNST: urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.\nto najnowsza generacja tego typu lidarów. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. korzyść mamy potrójną: użyteczną, wykonujemy pomiary skażeń atmosferycznych, naukową - rozwijamy badania nad urządzeniem I dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.',
148
+ 'Co to jest lidar? \nPROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. staramy się rozszerzyć jego zastosowanie na inne substancje występujące w atmosferze. Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej. zamierzamy zakończyć pomiary skażeń atmosferycznych nad granicą polsko-niemiecką. Nasz lidar ma większe możliwości niż stacje monitoringowe. Możemy śledzić ewolucję rozprzestrzeniania się zanieczyszczeń, ich kierunek i zmiany. Gwarancją sukcesu naszego programu dydaktyczno-badawczego jest udział w nim zakładów należących do Instytutu Fizyki Doświadczalnej UW, Pracowni Przetwarzania Informacji Instytutu Geofizyki i współpraca z Freie Universität Berlin.',
149
+ "Co to jest lidar? \nPROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. DIAL - lidar absorbcji różnicowej potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. staramy się rozszerzyć jego zastosowanie także na inne substancje występujące w atmosferze. Pakiet software'u wzbogacamy o nowe algorytmy, które potrafią dokładniej rozszyfrowywać sygnał lidarowy, a w konsekwencji skażenia. Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej. \n\nChcemy mierzyć w Warszawie rozkłady koncentracji tlenków azotu, ich ewolucję czasową nad różnymi arteriami miasta. zamierzamy zakończyć pomiary skażeń atmosferycznych nad granicą polsko-niemiecką. Nasz lidar ma większe możliwości niż stacje monitoringowe. Możemy śledzić ewolucję rozprzestrzeniania się zanieczyszczeń, ich kierunek i zmiany spowodowane m.in. warunkami atmosferycznymi. \n\nDIAL jest tym typem lidara, który dzisiaj ma największe wzięcie w ochronie środowiska. Z lidarów korzysta meteorologia. W Europie takich lidarów jak nasz jest zaledwie kilka. Nasz lidar jest najnowocześniejszym w Polsce. Ponadto jest lidarem ruchomym, zainstalowanym na samochodzie. \n\nFizycy dotychczas nie zajmowali się ochroną środowiska?\nTaka specjalizacja powstala na Wydziale Fizyki UW dwa lata temu. Gwarancją sukcesu naszego programu dydaktyczno-badawczego jest udział w nim zakładów należących do Instytutu Fizyki Doświadczalnej UW, Pracowni Przetwarzania Informacji Instytutu Geofizyki i współpraca z Freie Universität Berlin.",
150
+ 'Co to jest lidar? \nPROF. KRZYSZTOF ERNST: to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. zamierzamy zakończyć pomiary skażeń atmosferycznych nad granicą polsko-niemiecką. Nasz lidar ma większe możliwości niż stacje monitoringowe. Możemy śledzić ewolucję rozprzestrzeniania się zanieczyszczeń, ich kierunek i zmiany.'],
151
+ 'ratio': [10, 20, 5, 10, 20, 5, 10, 20, 5, 10, 20, 5, 10, 20, 5],
152
+ 'spans': [{'end': [244, 396, 457, 867, 922, 1022, 1103, 1877],
153
+ 'span_text': ['Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar?',
154
+ 'PROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym.',
155
+ 'Czy to kosztowne urządzenie będzie służyło tylko naukowcom?',
156
+ 'Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych,',
157
+ 'naukową - rozwijamy badania nad tym urządzeniem',
158
+ '.',
159
+ 'I korzyść dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.',
160
+ 'Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej, dzięki której ten lidar u nas zaistniał i dla której w ramach naszych zobowiązań wykonujemy pomiary zanieczyszczeń nad naszą wspólną granicą.'],
161
+ 'start': [153, 247, 398, 760, 875, 1020, 1023, 1631]},
162
+ {'end': [244,
163
+ 396,
164
+ 457,
165
+ 867,
166
+ 922,
167
+ 1022,
168
+ 1103,
169
+ 1878,
170
+ 2132,
171
+ 2296,
172
+ 2969,
173
+ 6225,
174
+ 6985,
175
+ 7047,
176
+ 7282,
177
+ 7326,
178
+ 7383],
179
+ 'span_text': ['Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar?',
180
+ 'PROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym.',
181
+ 'Czy to kosztowne urządzenie będzie służyło tylko naukowcom?',
182
+ 'Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych,',
183
+ 'naukową - rozwijamy badania nad tym urządzeniem',
184
+ '.',
185
+ 'I korzyść dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.',
186
+ 'Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej, dzięki której ten lidar u nas zaistniał i dla której w ramach naszych zobowiązań wykonujemy pomiary zanieczyszczeń nad naszą wspólną granicą.',
187
+ 'Czy wszystkie zanieczyszczenia będzie można wykryć za pomocą lidaru?',
188
+ 'Nie ma takiego jednostkowego urządzenia, które by wykrywało i mierzyło wszystkie szkodliwe gazy w atmosferze łącznie z dostarczeniem informacji o ich rozkładzie.',
189
+ 'Możemy np. badać zawartość ozonu w troposferze.',
190
+ 'W Europie takich lidarów jak nasz jest zaledwie kilka. Większość z nich mierzy ozon, dwutlenek siarki i tlenek azotu.',
191
+ '',
192
+ 'Fizycy dotychczas nie zajmowali się ochroną środowiska?',
193
+ 'Taka specjalizacja powstala na Wydziale Fizyki UW dwa lata temu. Gwarancją sukcesu naszego programu dydaktyczno-badawczego jest udział w nim zakładów należących do Instytutu Fizyki Doświadczalnej UW, Pracowni Przetwarzania Informacji',
194
+ 'Instytutu Geofizyki i',
195
+ 'współpraca z Freie Universität Berlin.'],
196
+ 'start': [153,
197
+ 247,
198
+ 398,
199
+ 760,
200
+ 875,
201
+ 1020,
202
+ 1023,
203
+ 1631,
204
+ 2064,
205
+ 2134,
206
+ 2921,
207
+ 6108,
208
+ 6984,
209
+ 6992,
210
+ 7049,
211
+ 7304,
212
+ 7344]},
213
+ {'end': [244, 396, 1103, 1774, 1877],
214
+ 'span_text': ['Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar?',
215
+ 'PROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym.',
216
+ '',
217
+ 'Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej, dzięki której ten lidar u nas zaistniał',
218
+ '.'],
219
+ 'start': [153, 247, 1102, 1631, 1876]},
220
+ {'end': [159,
221
+ 227,
222
+ 243,
223
+ 360,
224
+ 804,
225
+ 882,
226
+ 1025,
227
+ 1044,
228
+ 1103,
229
+ 1454,
230
+ 1540,
231
+ 1629,
232
+ 2848],
233
+ 'span_text': ['Jutro',
234
+ 'odbędzie sie pokaz nowego polskiego lidara typu DIAL.',
235
+ 'lidar',
236
+ 'Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.',
237
+ 'DIAL - lidar absorbcji różnicowej jest urządzeniem inteligentnym, to znaczy potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. Z lidara korzyść mamy potrójną: użyteczną,',
238
+ 'naukową',
239
+ 'I',
240
+ 'dydaktyczną',
241
+ '.',
242
+ 'Żeby przetworzyć',
243
+ 'sygnał lidarowy, czyli to co wraca po rozproszeniu światła do układu, i otrzymać',
244
+ 'dane dotyczące rozkładu koncentracji - trzeba dokonać skomplikowanych operacji.',
245
+ 'muszą być zastosowane dwie wiązki laserowe o dwóch różnych długościach fali, z których jedna jest absorbowana, a druga nie jest absorbowana przez substancję, którą chcemy wykryć.'],
246
+ 'start': [153,
247
+ 173,
248
+ 238,
249
+ 270,
250
+ 591,
251
+ 875,
252
+ 1022,
253
+ 1033,
254
+ 1101,
255
+ 1437,
256
+ 1459,
257
+ 1549,
258
+ 2670]},
259
+ {'end': [159, 227, 243, 396, 922, 1103, 1629, 2062, 2582, 2848],
260
+ 'span_text': ['Jutro',
261
+ 'odbędzie sie pokaz nowego polskiego lidara typu DIAL.',
262
+ 'lidar',
263
+ 'Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym.',
264
+ 'Jest to najnowsza generacja tego typu lidarów. DIAL - lidar absorbcji różnicowej jest urządzeniem inteligentnym, to znaczy potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, korzyść naukową - rozwijamy badania nad tym urządzeniem',
265
+ '. I korzyść dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.',
266
+ 'Żeby przetworzyć tzw. sygnał lidarowy, czyli to co wraca po rozproszeniu światła do układu, i otrzymać rozsądne dane dotyczące rozkładu koncentracji - trzeba dokonać skomplikowanych operacji.',
267
+ 'Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej, dzięki której ten lidar u nas zaistniał i dla której w ramach naszych zobowiązań wykonujemy pomiary zanieczyszczeń nad naszą wspólną granicą. Zasadniczy koszt jego budowy pokryła uzyskana od Fundacji dotacja. Część pieniędzy przekazał też Narodowy Fundusz Ochrony Środowiska i Gospodarki Wodnej oraz Komitet Badań Naukowych.',
268
+ '',
269
+ 'Lidar typu DIAL jest oparty na pomiarze absorbcji różnicowej, czyli muszą być zastosowane dwie wiązki laserowe o dwóch różnych długościach fali, z których jedna jest absorbowana, a druga nie jest absorbowana przez substancję, którą chcemy wykryć.'],
270
+ 'start': [153, 173, 238, 270, 542, 1020, 1437, 1631, 2581, 2602]},
271
+ {'end': [159, 227, 243, 360, 804, 882, 1025, 1044, 1102],
272
+ 'span_text': ['Jutro',
273
+ 'odbędzie sie pokaz nowego polskiego lidara typu DIAL.',
274
+ 'lidar',
275
+ 'Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.',
276
+ 'DIAL - lidar absorbcji różnicowej jest urządzeniem inteligentnym, to znaczy potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. Z lidara korzyść mamy potrójną: użyteczną,',
277
+ 'naukową',
278
+ 'I',
279
+ 'dydaktyczną',
280
+ '.'],
281
+ 'start': [153, 173, 238, 270, 591, 875, 1022, 1033, 1101]},
282
+ {'end': [246, 396, 922, 1102, 4763],
283
+ 'span_text': ['Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar?',
284
+ 'PROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym.',
285
+ 'DIAL - lidar absorbcji różnicowej jest urządzeniem inteligentnym, to znaczy potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, korzyść naukową - rozwijamy badania nad tym urządzeniem',
286
+ 'I korzyść dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.',
287
+ 'Nasz lidar ma większe możliwości niż stacje monitoringowe. Mierzy zanieczyszczenia nie tylko lokalnie, ale też ich rozkład w przestrzeni, z wysoką rozdzielczością przestrzenną i na odległość kilku kilometrów.'],
288
+ 'start': [153, 247, 590, 1022, 4555]},
289
+ {'end': [246, 396, 480, 542, 1021, 1102, 2920, 4989],
290
+ 'span_text': ['Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar?',
291
+ 'PROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi. Nazywane też jest radarem laserowym.',
292
+ 'Tego typu lidar jest',
293
+ 'drogi, kosztuje około miliona marek niemieckich.',
294
+ 'DIAL - lidar absorbcji różnicowej jest urządzeniem inteligentnym, to znaczy potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen. Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, korzyść naukową - rozwijamy badania nad tym urządzeniem, staramy się m.in. rozszerzyć jego zastosowanie także na inne substancje występujące w atmosferze.',
295
+ 'I korzyść dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.',
296
+ 'Lidar typu DIAL jest oparty na pomiarze absorbcji różnicowej, czyli muszą być zastosowane dwie wiązki laserowe o dwóch różnych długościach fali, z których jedna jest absorbowana, a druga nie jest absorbowana przez substancję, którą chcemy wykryć. Cząsteczki, które wykrywamy mają pasma absorbcji w bliskim nadfiolecie.',
297
+ 'Nasz lidar ma większe możliwości niż stacje monitoringowe. Mierzy zanieczyszczenia nie tylko lokalnie, ale też ich rozkład w przestrzeni, z wysoką rozdzielczością przestrzenną i na odległość kilku kilometrów. Możemy zatem śledzić ewolucję rozprzestrzeniania się tych zanieczyszczeń, ich kierunek i zmiany spowodowane m.in. warunkami atmosferycznymi. Wyniki naszych pomiarów porównujemy z danymi uzyskanymi ze stacji monitoringowych.'],
298
+ 'start': [153, 247, 459, 493, 590, 1022, 2602, 4555]},
299
+ {'end': [246, 360, 626, 883, 920, 1102],
300
+ 'span_text': ['Jutro w Instytucie odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar?',
301
+ 'PROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.',
302
+ '',
303
+ 'Z lidara korzyść mamy potrójną: użyteczną, bo przy jego pomocy wykonujemy pomiary skażeń atmosferycznych, korzyść naukową',
304
+ 'i',
305
+ 'dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.'],
306
+ 'start': [153, 247, 625, 760, 919, 1032]},
307
+ {'end': [158,
308
+ 262,
309
+ 271,
310
+ 359,
311
+ 397,
312
+ 590,
313
+ 761,
314
+ 803,
315
+ 867,
316
+ 907,
317
+ 922,
318
+ 1025,
319
+ 1102,
320
+ 3311,
321
+ 3516,
322
+ 3595,
323
+ 3623,
324
+ 3675,
325
+ 4226,
326
+ 4332],
327
+ 'span_text': ['Jutro',
328
+ 'odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar? \n\nPROF. KRZYSZTOF',
329
+ 'ERNST:',
330
+ 'urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.',
331
+ '',
332
+ 'to najnowsza generacja tego typu lidarów.',
333
+ 'Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen.',
334
+ 'korzyść mamy potrójną: użyteczną,',
335
+ 'przy jego pomocy wykonujemy pomiary skażeń atmosferycznych,',
336
+ 'naukową - rozwijamy badania nad',
337
+ 'urządzeniem',
338
+ 'I',
339
+ 'dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.',
340
+ '',
341
+ 'Nasze przedsięwzięcie nie ma charakteru komercyjnego.',
342
+ 'Chcemy np. mierzyć w Warszawie rozkłady',
343
+ 'koncentracji tlenków azotu',
344
+ '.',
345
+ 'Koncentrujemy się głównie na Czarnym Trójkącie - obszarze u zbiegu',
346
+ 'granic: Polski, Niemiec i Czech, do niedawna uważanym za najbardziej zdegradowany region w Europie.'],
347
+ 'start': [153,
348
+ 172,
349
+ 263,
350
+ 279,
351
+ 396,
352
+ 548,
353
+ 699,
354
+ 769,
355
+ 806,
356
+ 875,
357
+ 911,
358
+ 1022,
359
+ 1033,
360
+ 3310,
361
+ 3462,
362
+ 3556,
363
+ 3596,
364
+ 3674,
365
+ 4158,
366
+ 4233]},
367
+ {'end': [158,
368
+ 262,
369
+ 271,
370
+ 359,
371
+ 398,
372
+ 459,
373
+ 498,
374
+ 543,
375
+ 590,
376
+ 761,
377
+ 803,
378
+ 867,
379
+ 922,
380
+ 1025,
381
+ 1102,
382
+ 2242,
383
+ 2300,
384
+ 2406,
385
+ 3247,
386
+ 3311,
387
+ 3516,
388
+ 3595,
389
+ 3675,
390
+ 4226,
391
+ 4333,
392
+ 5130,
393
+ 5241,
394
+ 5439,
395
+ 5661,
396
+ 5756,
397
+ 7113],
398
+ 'span_text': ['Jutro',
399
+ 'odbędzie sie pokaz nowego polskiego lidara typu DIAL. Co to jest lidar? \n\nPROF. KRZYSZTOF',
400
+ 'ERNST:',
401
+ 'urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.',
402
+ '',
403
+ 'to kosztowne urządzenie będzie służyło tylko naukowcom?',
404
+ 'lidar jest rzeczywiście drogi',
405
+ '.',
406
+ 'to najnowsza generacja tego typu lidarów.',
407
+ 'Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen.',
408
+ 'korzyść mamy potrójną: użyteczną,',
409
+ 'przy jego pomocy wykonujemy pomiary skażeń atmosferycznych,',
410
+ 'naukową - rozwijamy badania nad tym urządzeniem',
411
+ 'I',
412
+ 'dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.',
413
+ 'Czy wszystkie zanieczyszczenia będzie można wykryć za pomocą lidaru?\n\nNie ma takiego jednostkowego urządzenia, które by wykrywało i mierzyło wszystkie szkodliwe gazy w atmosferze',
414
+ '. Ale',
415
+ 'prowadzimy badania mające na celu rozszerzenie możliwości lidaru o taką substancję jak fosgen.',
416
+ '',
417
+ 'stać nas będzie na prowadzenie pomiarów ozonu w miastach?',
418
+ 'Nasze przedsięwzięcie nie ma charakteru komercyjnego.',
419
+ 'Chcemy np. mierzyć w Warszawie rozkłady',
420
+ 'koncentracji tlenków azotu, ich ewolucję czasową nad różnymi arteriami miasta.',
421
+ 'Koncentrujemy się głównie na Czarnym Trójkącie - obszarze u zbiegu',
422
+ 'granic: Polski, Niemiec i Czech, do niedawna uważanym za najbardziej zdegradowany region w Europie.',
423
+ 'zanieczyszczenie',
424
+ 'było dużo większe niż obecnie i wszystko wskazuje na to, że będzie dalej spadać.',
425
+ '',
426
+ 'DIAL',
427
+ 'dzisiaj ma zdecydowanie największe wzięcie w ochronie środowiska.',
428
+ 'Fizycy dotychczas nie zajmowali się ochroną środowiska?\n\nTaka specjalizacja powstala na Wydziale Fizyki UW dwa lata temu.'],
429
+ 'start': [153,
430
+ 172,
431
+ 263,
432
+ 279,
433
+ 396,
434
+ 402,
435
+ 469,
436
+ 541,
437
+ 548,
438
+ 699,
439
+ 769,
440
+ 806,
441
+ 875,
442
+ 1022,
443
+ 1033,
444
+ 2062,
445
+ 2294,
446
+ 2312,
447
+ 3245,
448
+ 3251,
449
+ 3462,
450
+ 3556,
451
+ 3596,
452
+ 4158,
453
+ 4233,
454
+ 5114,
455
+ 5160,
456
+ 5438,
457
+ 5656,
458
+ 5690,
459
+ 6990]},
460
+ {'end': [262, 271, 359, 397, 590, 761, 803, 807, 867, 907, 922, 1025, 1102],
461
+ 'span_text': ['Co to jest lidar? \n\nPROF. KRZYSZTOF',
462
+ 'ERNST:',
463
+ 'urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.',
464
+ '',
465
+ 'to najnowsza generacja tego typu lidarów.',
466
+ 'Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen.',
467
+ 'korzyść mamy potrójną: użyteczną,',
468
+ '',
469
+ 'wykonujemy pomiary skażeń atmosferycznych,',
470
+ 'naukową - rozwijamy badania nad',
471
+ 'urządzeniem',
472
+ 'I',
473
+ 'dydaktyczną - szkolimy studentów zainteresowanych ochroną środowiska.'],
474
+ 'start': [227,
475
+ 263,
476
+ 279,
477
+ 396,
478
+ 548,
479
+ 699,
480
+ 769,
481
+ 806,
482
+ 824,
483
+ 875,
484
+ 911,
485
+ 1022,
486
+ 1033]},
487
+ {'end': [245,
488
+ 360,
489
+ 761,
490
+ 936,
491
+ 971,
492
+ 1022,
493
+ 1733,
494
+ 1878,
495
+ 4159,
496
+ 4614,
497
+ 4772,
498
+ 4818,
499
+ 4860,
500
+ 4906,
501
+ 7283,
502
+ 7326,
503
+ 7383],
504
+ 'span_text': ['Co to jest lidar?',
505
+ 'PROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.',
506
+ 'Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen.',
507
+ 'staramy się',
508
+ 'rozszerzyć jego zastosowanie',
509
+ 'na inne substancje występujące w atmosferze.',
510
+ 'Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej',
511
+ '.',
512
+ 'zamierzamy zakończyć pomiary skażeń atmosferycznych nad granicą polsko-niemiecką.',
513
+ 'Nasz lidar ma większe możliwości niż stacje monitoringowe.',
514
+ 'Możemy',
515
+ 'śledzić ewolucję rozprzestrzeniania się',
516
+ 'zanieczyszczeń, ich kierunek i zmiany',
517
+ '.',
518
+ 'Gwarancją sukcesu naszego programu dydaktyczno-badawczego jest udział w nim zakładów należących do Instytutu Fizyki Doświadczalnej UW, Pracowni Przetwarzania Informacji',
519
+ 'Instytutu Geofizyki i',
520
+ 'współpraca z Freie Universität Berlin.'],
521
+ 'start': [227,
522
+ 246,
523
+ 699,
524
+ 924,
525
+ 942,
526
+ 977,
527
+ 1631,
528
+ 1876,
529
+ 4076,
530
+ 4555,
531
+ 4765,
532
+ 4778,
533
+ 4823,
534
+ 4904,
535
+ 7114,
536
+ 7305,
537
+ 7344]},
538
+ {'end': [245,
539
+ 360,
540
+ 625,
541
+ 761,
542
+ 936,
543
+ 1022,
544
+ 1311,
545
+ 1357,
546
+ 1436,
547
+ 1733,
548
+ 1878,
549
+ 3247,
550
+ 3311,
551
+ 3563,
552
+ 3676,
553
+ 4159,
554
+ 4614,
555
+ 4772,
556
+ 4818,
557
+ 4906,
558
+ 5410,
559
+ 5439,
560
+ 5701,
561
+ 5789,
562
+ 6163,
563
+ 6364,
564
+ 6472,
565
+ 7048,
566
+ 7283,
567
+ 7326,
568
+ 7383],
569
+ 'span_text': ['Co to jest lidar?',
570
+ 'PROF. KRZYSZTOF ERNST: Jest to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.',
571
+ 'DIAL - lidar absorbcji różnicowej',
572
+ 'potrafi rozróżnić, co mierzy. Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen.',
573
+ 'staramy się',
574
+ 'rozszerzyć jego zastosowanie także na inne substancje występujące w atmosferze.',
575
+ "Pakiet software'u",
576
+ 'wzbogacamy o nowe algorytmy, które potrafią',
577
+ 'dokładniej rozszyfrowywać sygnał lidarowy, a w konsekwencji skażenia.',
578
+ 'Badania, które prowadzimy, są zainicjowane i finansowane przez Fundację Współpracy Polsko-Niemieckiej',
579
+ '.',
580
+ '',
581
+ '',
582
+ 'Chcemy',
583
+ 'mierzyć w Warszawie rozkłady koncentracji tlenków azotu, ich ewolucję czasową nad różnymi arteriami miasta.',
584
+ 'zamierzamy zakończyć pomiary skażeń atmosferycznych nad granicą polsko-niemiecką.',
585
+ 'Nasz lidar ma większe możliwości niż stacje monitoringowe.',
586
+ 'Możemy',
587
+ 'śledzić ewolucję rozprzestrzeniania się',
588
+ 'zanieczyszczeń, ich kierunek i zmiany spowodowane m.in. warunkami atmosferycznymi.',
589
+ '',
590
+ '',
591
+ 'DIAL jest tym typem lidara, który dzisiaj ma',
592
+ 'największe wzięcie w ochronie środowiska. Z lidarów korzysta meteorologia.',
593
+ 'W Europie takich lidarów jak nasz jest zaledwie kilka.',
594
+ 'Nasz lidar',
595
+ 'jest najnowocześniejszym w Polsce. Ponadto jest lidarem ruchomym, zainstalowanym na samochodzie.',
596
+ 'Fizycy dotychczas nie zajmowali się ochroną środowiska?',
597
+ 'Taka specjalizacja powstala na Wydziale Fizyki UW dwa lata temu. Gwarancją sukcesu naszego programu dydaktyczno-badawczego jest udział w nim zakładów należących do Instytutu Fizyki Doświadczalnej UW, Pracowni Przetwarzania Informacji',
598
+ 'Instytutu Geofizyki i',
599
+ 'współpraca z Freie Universität Berlin.'],
600
+ 'start': [227,
601
+ 246,
602
+ 591,
603
+ 668,
604
+ 924,
605
+ 942,
606
+ 1293,
607
+ 1313,
608
+ 1366,
609
+ 1631,
610
+ 1876,
611
+ 3246,
612
+ 3310,
613
+ 3556,
614
+ 3567,
615
+ 4076,
616
+ 4555,
617
+ 4765,
618
+ 4778,
619
+ 4823,
620
+ 5409,
621
+ 5438,
622
+ 5656,
623
+ 5714,
624
+ 6108,
625
+ 6353,
626
+ 6374,
627
+ 6990,
628
+ 7049,
629
+ 7305,
630
+ 7344]},
631
+ {'end': [245, 271, 360, 761, 4159, 4614, 4772, 4818, 4860, 4905],
632
+ 'span_text': ['Co to jest lidar?',
633
+ 'PROF. KRZYSZTOF ERNST:',
634
+ 'to urządzenie pozwalające wyznaczać zanieczyszczenia atmosfery metodami optycznymi.',
635
+ 'Wykrywa ozon, dwutlenek siarki, tlenki azotu, benzen, toluen.',
636
+ 'zamierzamy zakończyć pomiary skażeń atmosferycznych nad granicą polsko-niemiecką.',
637
+ 'Nasz lidar ma większe możliwości niż stacje monitoringowe.',
638
+ 'Możemy',
639
+ 'śledzić ewolucję rozprzestrzeniania się',
640
+ 'zanieczyszczeń, ich kierunek i zmiany',
641
+ '.'],
642
+ 'start': [227, 246, 276, 699, 4076, 4555, 4765, 4778, 4823, 4904]}],
643
+ 'type': ['extract',
644
+ 'extract',
645
+ 'extract',
646
+ 'extract',
647
+ 'extract',
648
+ 'extract',
649
+ 'extract',
650
+ 'extract',
651
+ 'extract',
652
+ 'extract',
653
+ 'extract',
654
+ 'extract',
655
+ 'extract',
656
+ 'extract',
657
+ 'extract']},
658
+ 'title': 'Lidarowe oczy'}
659
+ ```
660
+
661
+ ### Data Fields
662
+
663
+ - `id`: a `string` example identifier
664
+ - `date`: date of the original article (`string`)
665
+ - `title`: title of the original article (`string`)
666
+ - `section`: the section of the newspaper the original article belonged to (`string`)
667
+ - `authors`: original article authors (`string`)
668
+ - `body`: original article body (list of `string`s)
669
+ - `summaries`: a dictionary feature containing summaries of the original article with the following attributes:
670
+ - `ratio`: ratio of summary - percentage of the original article (list of `int32`s)
671
+ - `type`: type of summary - extractive (`extract`) or abstractive (`abstract`) (list of `string`s)
672
+ - `author`: acronym of summary author (list of `string`)
673
+ - `body`: body of summary (list of `string`)
674
+ - `spans`: a list containing spans for extractive summaries (empty for abstractive summaries):
675
+ - `start`: start of span (`int32`)
676
+ - `end`: end of span (`int32`)
677
+ - `span_text`: span text (`string`)
678
+
679
+ ### Data Splits
680
+
681
+ Single train split
682
+
683
+ ## Dataset Creation
684
+
685
+ ### Curation Rationale
686
+
687
+ [Needs More Information]
688
+
689
+ ### Source Data
690
+
691
+ #### Initial Data Collection and Normalization
692
+
693
+ [Needs More Information]
694
+
695
+ #### Who are the source language producers?
696
+
697
+ [Needs More Information]
698
+
699
+ ### Annotations
700
+
701
+ #### Annotation process
702
+
703
+ [Needs More Information]
704
+
705
+ #### Who are the annotators?
706
+
707
+ [Needs More Information]
708
+
709
+ ### Personal and Sensitive Information
710
+
711
+ [Needs More Information]
712
+
713
+ ## Considerations for Using the Data
714
+
715
+ ### Social Impact of Dataset
716
+
717
+ [Needs More Information]
718
+
719
+ ### Discussion of Biases
720
+
721
+ [Needs More Information]
722
+
723
+ ### Other Known Limitations
724
+
725
+ [Needs More Information]
726
+
727
+ ## Additional Information
728
+
729
+ ### Dataset Curators
730
+
731
+ [Needs More Information]
732
+
733
+ ### Licensing Information
734
+
735
+ [Needs More Information]
736
+
737
+ ### Citation Information
738
+ ```
739
+ @inproceedings{
740
+ ogro:kop:14:lrec,
741
+ author = "Ogrodniczuk, Maciej and Kopeć, Mateusz",
742
+ pdf = "http://nlp.ipipan.waw.pl/Bib/ogro:kop:14:lrec.pdf",
743
+ title = "The {P}olish {S}ummaries {C}orpus",
744
+ pages = "3712--3715",
745
+ crossref = "lrec:14"
746
+ }
747
+ @proceedings{
748
+ lrec:14,
749
+ editor = "Calzolari, Nicoletta and Choukri, Khalid and Declerck, Thierry and Loftsson, Hrafn and Maegaard, Bente and Mariani, Joseph and Moreno, Asuncion and Odijk, Jan and Piperidis, Stelios",
750
+ isbn = "978-2-9517408-8-4",
751
+ title = "Proceedings of the Ninth International {C}onference on {L}anguage {R}esources and {E}valuation, {LREC}~2014",
752
+ url = "http://www.lrec-conf.org/proceedings/lrec2014/index.html",
753
+ booktitle = "Proceedings of the Ninth International {C}onference on {L}anguage {R}esources and {E}valuation, {LREC}~2014",
754
+ address = "Reykjavík, Iceland",
755
+ key = "LREC",
756
+ year = "2014",
757
+ organization = "European Language Resources Association (ELRA)"
758
+ }
759
+ ```
760
+ ### Contributions
761
+
762
+ Thanks to [@kldarek](https://github.com/kldarek) for adding this dataset.
huggingface_dataset/Dataset_Card/readerbench_ro-fb-offense.md ADDED
@@ -0,0 +1,178 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ annotations_creators:
3
+ - expert-generated
4
+ language_creators:
5
+ - found
6
+ language:
7
+ - ro
8
+ license:
9
+ - apache-2.0
10
+ multilinguality:
11
+ - monolingual
12
+ size_categories:
13
+ - 1K<n<10K
14
+ source_datasets:
15
+ - original
16
+ task_categories:
17
+ - text-classification
18
+ task_ids:
19
+ - hate-speech-detection
20
+ pretty_name: RO-FB-Offense
21
+ extra_gated_prompt: 'Warning: this repository contains harmful content (abusive language,
22
+ hate speech).'
23
+ tags:
24
+ - hate-speech-detection
25
+ ---
26
+
27
+ # Dataset Card for "RO-FB-Offense"
28
+
29
+ ## Table of Contents
30
+ - [Dataset Description](#dataset-description)
31
+ - [Dataset Summary](#dataset-summary)
32
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
33
+ - [Languages](#languages)
34
+ - [Dataset Structure](#dataset-structure)
35
+ - [Data Instances](#data-instances)
36
+ - [Data Fields](#data-fields)
37
+ - [Data Splits](#data-splits)
38
+ - [Dataset Creation](#dataset-creation)
39
+ - [Curation Rationale](#curation-rationale)
40
+ - [Source Data](#source-data)
41
+ - [Annotations](#annotations)
42
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
43
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
44
+ - [Social Impact of Dataset](#social-impact-of-dataset)
45
+ - [Discussion of Biases](#discussion-of-biases)
46
+ - [Other Known Limitations](#other-known-limitations)
47
+ - [Additional Information](#additional-information)
48
+ - [Dataset Curators](#dataset-curators)
49
+ - [Licensing Information](#licensing-information)
50
+ - [Citation Information](#citation-information)
51
+ - [Contributions](#contributions)
52
+
53
+ ## Dataset Description
54
+
55
+ - **Homepage:** [https://github.com/readerbench/ro-fb-offense](https://github.com/readerbench/ro-fb-offense)
56
+ - **Repository:** [https://github.com/readerbench/ro-fb-offense](https://github.com/readerbench/ro-fb-offense)
57
+ - **Paper:** FB-RO-Offense – A Romanian Dataset and Baseline Models for detecting Offensive Language in Facebook Comments
58
+ - **Point of Contact:** [Andrei Paraschiv](https://github.com/AndyTheFactory)
59
+
60
+ ### Dataset Summary
61
+
62
+ FB-RO-Offense corpus, an offensive speech dataset containing 4,455 user-generated comments from Facebook live broadcasts available in Romanian
63
+
64
+ The annotation follows the hierarchical tagset proposed in the Germeval 2018 Dataset.
65
+ The following Classes are available:
66
+ * OTHER: Non-Offensive Language
67
+ * OFFENSIVE:
68
+ - PROFANITY
69
+ - INSULT
70
+ - ABUSE
71
+
72
+ ### Languages
73
+
74
+ Romanian
75
+
76
+ ## Dataset Structure
77
+
78
+
79
+ ### Data Instances
80
+
81
+
82
+ An example of 'train' looks as follows.
83
+
84
+ ```
85
+ {
86
+ 'sender': '$USER1208',
87
+ 'no_reacts': 1,
88
+ 'text': 'PLACEHOLDER TEXT',
89
+ 'label': OTHER,
90
+ }
91
+ ```
92
+
93
+
94
+ ### Data Fields
95
+
96
+ - `sender`: a `string` feature.
97
+ - 'no_reacts': a `integer`
98
+ - `text`: a `string`.
99
+ - `label`: categorical `OTHER`, `PROFANITY`, `INSULT`, `ABUSE`
100
+
101
+
102
+ ### Data Splits
103
+
104
+ | name |train|test|
105
+ |---------|----:|---:|
106
+ |ro|x|x|
107
+
108
+
109
+ ## Dataset Creation
110
+
111
+ ### Curation Rationale
112
+
113
+ Collecting data for abusive language classification for Romanian Language.
114
+
115
+ ### Source Data
116
+
117
+ Facebook comments
118
+
119
+ #### Initial Data Collection and Normalization
120
+
121
+
122
+
123
+ #### Who are the source language producers?
124
+
125
+ Social media users
126
+
127
+ ### Annotations
128
+
129
+ #### Annotation process
130
+
131
+
132
+
133
+ #### Who are the annotators?
134
+
135
+ Native speakers
136
+
137
+ ### Personal and Sensitive Information
138
+
139
+ The data was public at the time of collection. No PII removal has been performed.
140
+
141
+ ## Considerations for Using the Data
142
+
143
+ ### Social Impact of Dataset
144
+
145
+ The data definitely contains abusive language. The data could be used to develop and propagate offensive language against every target group involved, i.e. ableism, racism, sexism, ageism, and so on.
146
+
147
+ ### Discussion of Biases
148
+
149
+
150
+ ### Other Known Limitations
151
+
152
+
153
+ ## Additional Information
154
+
155
+
156
+ ### Dataset Curators
157
+
158
+
159
+ ### Licensing Information
160
+
161
+ This data is available and distributed under Apache-2.0 license
162
+
163
+ ### Citation Information
164
+
165
+ ```
166
+ @inproceedings{busuioc2022fb-ro-offense,
167
+ title={FB-RO-Offense – A Romanian Dataset and Baseline Models for detecting Offensive Language in Facebook Comments},
168
+ author={ Busuioc, Gabriel-Razvan and Paraschiv, Andrei and Dascalu, Mihai},
169
+ booktitle={International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC) 2022},
170
+ year={2022}
171
+ }
172
+
173
+ ```
174
+
175
+
176
+ ### Contributions
177
+
178
+