Update files from the datasets library (from 1.3.0)

Release notes: https://github.com/huggingface/datasets/releases/tag/1.3.0

.gitattributes

@@ -0,0 +1,27 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bin.* filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zstandard filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,390 @@
+---
+annotations_creators:
+- expert-generated
+language_creators:
+- expert-generated
+languages:
+- en
+licenses:
+- cc-by-sa-4-0
+multilinguality:
+- monolingual
+size_categories:
+- 10K<n<100K
+source_datasets:
+- original
+task_categories:
+- sequence-modeling
+- text-classification
+- text-scoring
+task_ids:
+  dyda_da:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-dialogue-act-classification
+  dyda_e:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-emotion-classification
+  iemocap:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-emotion-classification
+  maptask:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-dialogue-act-classification
+  meld_e:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-emotion-classification
+  meld_s:
+  - dialogue-modeling
+  - language-modeling
+  - sentiment-classification
+  mrda:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-dialogue-act-classification
+  oasis:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-dialogue-act-classification
+  sem:
+  - dialogue-modeling
+  - language-modeling
+  - sentiment-classification
+  swda:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-dialogue-act-classification
+---
+
+# Dataset Card for SILICONE Benchmark
+
+## Table of Contents
+- [Dataset Description](#dataset-description)
+  - [Dataset Summary](#dataset-summary)
+  - [Supported Tasks](#supported-tasks-and-leaderboards)
+  - [Languages](#languages)
+- [Dataset Structure](#dataset-structure)
+  - [Data Instances](#data-instances)
+  - [Data Fields](#data-fields)
+  - [Data Splits](#data-splits)
+- [Dataset Creation](#dataset-creation)
+  - [Curation Rationale](#curation-rationale)
+  - [Source Data](#source-data)
+  - [Annotations](#annotations)
+  - [Personal and Sensitive Information](#personal-and-sensitive-information)
+- [Considerations for Using the Data](#considerations-for-using-the-data)
+  - [Social Impact of Dataset](#social-impact-of-dataset)
+  - [Discussion of Biases](#discussion-of-biases)
+  - [Other Known Limitations](#other-known-limitations)
+- [Additional Information](#additional-information)
+  - [Dataset Curators](#dataset-curators)
+  - [Licensing Information](#licensing-information)
+  - [Citation Information](#citation-information)
+  - [Contributions](#contributions)
+
+## Dataset Description
+
+- **Homepage:** [N/A]
+- **Repository:** https://github.com/eusip/SILICONE-benchmark
+- **Paper:** https://arxiv.org/abs/2009.11152
+- **Leaderboard:** [N/A]
+- **Point of Contact:** Ebenge Usip, [email protected]
+
+### Dataset Summary
+
+The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection of resources for training, evaluating, and analyzing natural language understanding systems specifically designed for spoken language. All datasets are in the English language and covers a variety of domains including daily life, scripted scenarios, joint task completion, phone call conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant labels.
+
+### Supported Tasks and Leaderboards
+
+[More Information Needed]
+
+### Languages
+
+English.
+
+## Dataset Structure
+
+### Data Instances
+
+#### DailyDialog Act Corpus (Dialogue Act)
+For the `dyda_da` configuration one example from the dataset is:
+```
+{
+  'Utterance': "the taxi drivers are on strike again .",
+  'Dialogue_Act': 2, # "inform"
+  'Dialogue_ID': "2"
+}
+```
+
+#### DailyDialog Act Corpus (Emotion)
+For the `dyda_e` configuration one example from the dataset is:
+```
+{
+  'Utterance': "'oh , breaktime flies .'",
+  'Emotion': 5, # "sadness"
+  'Dialogue_ID': "997"
+}
+```
+
+#### Interactive Emotional Dyadic Motion Capture (IEMOCAP) database
+For the `iemocap` configuration one example from the dataset is:
+```
+{
+  'Dialogue_ID': "Ses04F_script03_2",
+  'Utterance_ID': "Ses04F_script03_2_F025",
+  'Utterance': "You're quite insufferable.  I expect it's because you're drunk.",
+  'Emotion': 0, # "ang"
+}
+```
+
+#### HCRC MapTask Corpus
+For the `maptask` configuration one example from the dataset is:
+```
+{
+  'Speaker': "f",
+  'Utterance': "i think that would bring me over the crevasse",
+  'Dialogue_Act': 4, # "explain"
+}
+```
+
+
+#### Multimodal EmotionLines Dataset (Emotion)
+For the `meld_e` configuration one example from the dataset is:
+```
+{
+  'Utterance': "'Push 'em out , push 'em out , harder , harder .'",
+  'Speaker': "Joey",
+  'Emotion': 3, # "joy"
+  'Dialogue_ID': "1",
+  'Utterance_ID': "2"
+}
+```
+
+#### Multimodal EmotionLines Dataset (Sentiment)
+For the `meld_s` configuration one example from the dataset is:
+```
+{
+  'Utterance': "'Okay , y'know what ? There is no more left , left !'",
+  'Speaker': "Rachel",
+  'Sentiment': 0, # "negative"
+  'Dialogue_ID': "2",
+  'Utterance_ID': "4"
+}
+```
+
+#### ICSI MRDA Corpus
+For the `mrda` configuration one example from the dataset is:
+```
+{
+  'Utterance_ID': "Bed006-c2_0073656_0076706",
+  'Dialogue_Act': 0, # "s"
+  'Channel_ID': "Bed006-c2",
+  'Speaker': "mn015",
+  'Dialogue_ID': "Bed006",
+  'Utterance': "keith is not technically one of us yet ."
+}
+```
+
+#### BT OASIS Corpus
+For the `oasis` configuration one example from the dataset is:
+```
+{
+  'Speaker': "b",
+  'Utterance': "when i rang up um when i rang to find out why she said oh well your card's been declined",
+  'Dialogue_Act': 21, # "inform"
+}
+```
+
+#### SEMAINE database
+For the `sem` configuration one example from the dataset is:
+```
+{
+  'Utterance': "can you think of somebody who is like that ?",
+  'NbPairInSession': "11",
+  'Dialogue_ID': "59",
+  'SpeechTurn': "674",
+  'Speaker': "Agent",
+  'Sentiment': 1, # "Neutral"
+}
+```
+
+#### Switchboard Dialog Act (SwDA) Corpus
+For the `swda` configuration one example from the dataset is:
+```
+{
+  'Utterance': "but i 'd probably say that 's roughly right .",
+  'Dialogue_Act': 33, # "aap_am"
+  'From_Caller': "1255",
+  'To_Caller': "1087",
+  'Topic': "CRIME",
+  'Dialogue_ID': "818",
+  'Conv_ID': "sw2836",
+}
+```
+
+### Data Fields
+
+For the `dyda_da` configuration, the different fields are:
+- `Utterance`: Utterance as a string.
+- `Dialogue_Act`: Dialog act label of the utterance. It can be one of "commissive" (0), "directive" (1), "inform" (2) or "question" (3).
+- `Dialogue_ID`: identifier of the dialogue as a string.
+
+For the `dyda_e` configuration, the different fields are:
+- `Utterance`: Utterance as a string.
+- `Dialogue_Act`: Dialog act label of the utterance. It can be one of "anger" (0), "disgust" (1), "fear" (2), "happiness" (3), "no emotion" (4), "sadness" (5) or "surprise" (6).
+- `Dialogue_ID`: identifier of the dialogue as a string.
+
+For the `iemocap` configuration, the different fields are:
+- `Dialogue_ID`: identifier of the dialogue as a string.
+- `Utterance_ID`: identifier of the utterance as a string.
+- `Utterance`: Utterance as a string.
+- `Emotion`: Emotion label of the utterance. It can be one of "Anger" (0), "Disgust" (1), "Excitement" (2), "Fear" (3), "Frustration" (4), "Happiness" (5), "Neutral" (6), "Other" (7), "Sadness" (8), "Surprise" (9) or "Unknown" (10).
+
+For the `maptask` configuration, the different fields are:
+- `Speaker`: identifier of the speaker as a string.
+- `Utterance`: Utterance as a string.
+- `Dialogue_Act`: Dialog act label of the utterance. It can be one of "acknowledge" (0), "align" (1), "check" (2), "clarify" (3), "explain" (4), "instruct" (5), "query_w" (6), "query_yn" (7), "ready" (8), "reply_n" (9), "reply_w" (10) or "reply_y" (11).
+
+For the `meld_e` configuration, the different fields are:
+- `Utterance`: Utterance as a string.
+- `Speaker`: Speaker as a string.
+- `Emotion`: Emotion label of the utterance. It can be one of "anger" (0), "disgust" (1), "fear" (2), "joy" (3), "neutral" (4), "sadness" (5) or "surprise" (6).
+- `Dialogue_ID`: identifier of the dialogue as a string.
+- `Utterance_ID`: identifier of the utterance as a string.
+
+For the `meld_s` configuration, the different fields are:
+- `Utterance`: Utterance as a string.
+- `Speaker`: Speaker as a string.
+- `Sentiment`: Sentiment label of the utterance. It can be one of "negative" (0), "neutral" (1) or "positive" (2).
+- `Dialogue_ID`: identifier of the dialogue as a string.
+- `Utterance_ID`: identifier of the utterance as a string.
+
+For the `mrda` configuration, the different fields are:
+- `Utterance_ID`: identifier of the utterance as a string.
+- `Dialogue_Act`: Dialog act label of the utterance. It can be one of "s" (0) [Statement/Subjective Statement], "d" (1) [Declarative Question], "b" (2) [Backchannel], "f" (3) [Follow-me] or "q" (4) [Question].
+- `Channel_ID`: identifier of the channel as a string.
+- `Speaker`: identifier of the speaker as a string.
+- `Dialogue_ID`: identifier of the channel as a string.
+- `Utterance`: Utterance as a string.
+
+For the `oasis` configuration, the different fields are:
+- `Speaker`: identifier of the speaker as a string.
+- `Utterance`: Utterance as a string.
+- `Dialogue_Act`: Dialog act label of the utterance. It can be one of "accept" (0), "ackn" (1), "answ" (2), "answElab" (3), "appreciate" (4), "backch" (5), "bye" (6), "complete" (7), "confirm" (8), "correct" (9), "direct" (10), "directElab" (11), "echo" (12), "exclaim" (13), "expressOpinion"(14), "expressPossibility" (15), "expressRegret" (16), "expressWish" (17), "greet" (18), "hold" (19),
+"identifySelf" (20), "inform" (21), "informCont" (22), "informDisc" (23), "informIntent" (24), "init" (25), "negate" (26), "offer" (27), "pardon" (28), "raiseIssue" (29), "refer" (30), "refuse" (31), "reqDirect" (32), "reqInfo" (33), "reqModal" (34), "selfTalk" (35), "suggest" (36), "thank" (37), "informIntent-hold" (38), "correctSelf" (39), "expressRegret-inform" (40) or "thank-identifySelf" (41).
+
+For the `sem` configuration, the different fields are:
+- `Utterance`: Utterance as a string.
+- `NbPairInSession`: number of utterance pairs in a dialogue.
+- `Dialogue_ID`: identifier of the dialogue as a string.
+- `SpeechTurn`: SpeakerTurn as a string.
+- `Speaker`: Speaker as a string.
+- `Sentiment`: Sentiment label of the utterance. It can be "Negative", "Neutral" or "Positive".
+
+For the `swda` configuration, the different fields are:
+`Utterance`: Utterance as a string.
+`Dialogue_Act`: Dialogue act label of the utterance. It can be "sd" (0) [Statement-non-opinion], "b" (1) [Acknowledge (Backchannel)], "sv" (2) [Statement-opinion], "%" (3) [Uninterpretable], "aa" (4) [Agree/Accept], "ba" (5) [Appreciation], "fc" (6) [Conventional-closing], "qw" (7) [Wh-Question], "nn" (8) [No Answers], "bk" (9) [Response Acknowledgement], "h" (10) [Hedge], "qy^d" (11) [Declarative Yes-No-Question], "bh" (12) [Backchannel in Question Form], "^q" (13) [Quotation], "bf" (14) [Summarize/Reformulate], 'fo_o_fw_"_by_bc' (15) [Other], 'fo_o_fw_by_bc_"' (16) [Other], "na" (17) [Affirmative Non-yes Answers], "ad" (18) [Action-directive], "^2" (19) [Collaborative Completion], "b^m" (20) [Repeat-phrase], "qo" (21) [Open-Question], "qh" (22) [Rhetorical-Question], "^h" (23) [Hold Before Answer/Agreement], "ar" (24) [Reject], "ng" (25) [Negative Non-no Answers], "br" (26) [Signal-non-understanding], "no" (27) [Other Answers], "fp" (28) [Conventional-opening], "qrr" (29) [Or-Clause], "arp_nd" (30) [Dispreferred Answers], "t3" (31) [3rd-party-talk], "oo_co_cc" (32) [Offers, Options Commits], "aap_am" (33) [Maybe/Accept-part], "t1" (34) [Downplayer], "bd" (35) [Self-talk], "^g" (36) [Tag-Question], "qw^d" (37) [Declarative Wh-Question], "fa" (38) [Apology], "ft" (39) [Thanking], "+" (40) [Unknown], "x" (41) [Unknown], "ny" (42) [Unknown], "sv_fx" (43) [Unknown], "qy_qr" (44) [Unknown] or "ba_fe" (45) [Unknown].
+`From_Caller`: identifier of the from caller as a string.
+`To_Caller`: identifier of the to caller as a string.
+`Topic`: Topic as a string.
+`Dialogue_ID`: identifier of the dialogue as a string.
+`Conv_ID`: identifier of the conversation as a string.
+
+### Data Splits
+
+| Dataset name | Train  | Valid | Test  |
+| ------------ | -----  | ----- | ----  |
+| dyda_da      | 87170  | 8069  | 7740  |
+| dyda_e       | 87170  | 8069  | 7740  |
+| iemocap      | 7213   | 805   | 2021  |
+| maptask      | 20905  | 2963  | 2894  |
+| meld_e       | 9989   | 1109  | 2610  |
+| meld_s       | 9989   | 1109  | 2610  |
+| mrda         | 83944  | 9815  | 15470 |
+| oasis        | 12076  | 1513  | 1478  |
+| sem          | 4264   | 485   | 878   |
+| swda         | 190709 | 21203 | 2714  |
+
+## Dataset Creation
+
+### Curation Rationale
+
+[More Information Needed]
+
+### Source Data
+
+#### Initial Data Collection and Normalization
+
+[More Information Needed]
+
+#### Who are the source language producers?
+
+[More Information Needed]
+
+### Annotations
+
+#### Annotation process
+
+[More Information Needed]
+
+#### Who are the annotators?
+
+[More Information Needed]
+
+### Personal and Sensitive Information
+
+[More Information Needed]
+
+## Considerations for Using the Data
+
+### Social Impact of Dataset
+
+[More Information Needed]
+
+### Discussion of Biases
+
+[More Information Needed]
+
+### Other Known Limitations
+
+[More Information Needed]
+
+## Additional Information
+
+### Benchmark Curators
+
+Emile Chapuis, Pierre Colombo, Ebenge Usip.
+
+### Licensing Information
+
+This work is licensed under a [Creative Commons Attribution-NonCommercial-ShareAlike 4.0 Unported License](https://creativecommons.org/licenses/by-sa/4.0/).
+
+### Citation Information
+
+```
+@inproceedings{chapuis-etal-2020-hierarchical,
+    title = "Hierarchical Pre-training for Sequence Labelling in Spoken Dialog",
+    author = "Chapuis, Emile  and
+      Colombo, Pierre  and
+      Manica, Matteo  and
+      Labeau, Matthieu  and
+      Clavel, Chlo{\'e}",
+    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2020",
+    month = nov,
+    year = "2020",
+    address = "Online",
+    publisher = "Association for Computational Linguistics",
+    url = "https://www.aclweb.org/anthology/2020.findings-emnlp.239",
+    doi = "10.18653/v1/2020.findings-emnlp.239",
+    pages = "2636--2648",
+    abstract = "Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a key component of spoken dialog systems. In this work, we propose a new approach to learn generic representations adapted to spoken dialog, which we evaluate on a new benchmark we call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE). SILICONE is model-agnostic and contains 10 different datasets of various sizes. We obtain our representations with a hierarchical encoder based on transformer architectures, for which we extend two well-known pre-training objectives. Pre-training is performed on OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We demonstrate how hierarchical encoders achieve competitive results with consistently fewer parameters compared to state-of-the-art models and we show their importance for both pre-training and fine-tuning.",
+}
+```
+
+### Contributions
+
+Thanks to [@eusip](https://github.com/eusip) and [@lhoestq](https://github.com/lhoestq) for adding this dataset.

dataset_infos.json

@@ -0,0 +1 @@
+{"dyda_da": {"description": "The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection\n of resources for training, evaluating, and analyzing natural language understanding systems\n specifically designed for spoken language. All datasets are in the English language and cover a\n variety of domains including daily life, scripted scenarios, joint task completion, phone call\n conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant\n labels.\n", "citation": "@InProceedings{li2017dailydialog,\nauthor = {Li, Yanran and Su, Hui and Shen, Xiaoyu and Li, Wenjie and Cao, Ziqiang and Niu, Shuzi},\ntitle = {DailyDialog: A Manually Labelled Multi-turn Dialogue Dataset},\nbooktitle = {Proceedings of The 8th International Joint Conference on Natural Language Processing (IJCNLP 2017)},\nyear = {2017}\n}\n@inproceedings{chapuis-etal-2020-hierarchical,\n    title = \"Hierarchical Pre-training for Sequence Labelling in Spoken Dialog\",\n    author = \"Chapuis, Emile  and\n      Colombo, Pierre  and\n      Manica, Matteo  and\n      Labeau, Matthieu  and\n      Clavel, Chlo{'e}\",\n    booktitle = \"Findings of the Association for Computational Linguistics: EMNLP 2020\",\n    month = nov,\n    year = \"2020\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://www.aclweb.org/anthology/2020.findings-emnlp.239\",\n    doi = \"10.18653/v1/2020.findings-emnlp.239\",\n    pages = \"2636--2648\",\n    abstract = \"Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a\n        key component of spoken dialog systems. In this work, we propose a new approach to learn\n        generic representations adapted to spoken dialog, which we evaluate on a new benchmark we\n        call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE).\n        SILICONE is model-agnostic and contains 10 different datasets of various sizes.\n        We obtain our representations with a hierarchical encoder based on transformer architectures,\n        for which we extend two well-known pre-training objectives. Pre-training is performed on\n        OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We\n        demonstrate how hierarchical encoders achieve competitive results with consistently fewer\n        parameters compared to state-of-the-art models and we show their importance for both\n        pre-training and fine-tuning.\",\n}\n", "homepage": "http://yanran.li/dailydialog.html", "license": "", "features": {"Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_Act": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 4, "names": ["commissive", "directive", "inform", "question"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "silicone", "config_name": "dyda_da", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 8346638, "num_examples": 87170, "dataset_name": "silicone"}, "validation": {"name": "validation", "num_bytes": 764277, "num_examples": 8069, "dataset_name": "silicone"}, "test": {"name": "test", "num_bytes": 740226, "num_examples": 7740, "dataset_name": "silicone"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/dyda/train.csv": {"num_bytes": 7519783, "checksum": "9710558dd08351e1544424fcf43f8aed5a3e99daba3e1f133e86acf2373fb4c6"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/dyda/dev.csv": {"num_bytes": 688589, "checksum": "66be9ec11f0234686cd338d5e7a374e1706ffd998b54fd3e6b72ef18598470b2"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/dyda/test.csv": {"num_bytes": 666553, "checksum": "3b711ebac2a1a64067aea5252c6bbbb1abe1f9f1434d2e94f9dd4a99706fc2ce"}}, "download_size": 8874925, "post_processing_size": null, "dataset_size": 9851141, "size_in_bytes": 18726066}, "dyda_e": {"description": "The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection\n of resources for training, evaluating, and analyzing natural language understanding systems\n specifically designed for spoken language. All datasets are in the English language and cover a\n variety of domains including daily life, scripted scenarios, joint task completion, phone call\n conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant\n labels.\n", "citation": "@InProceedings{li2017dailydialog,\nauthor = {Li, Yanran and Su, Hui and Shen, Xiaoyu and Li, Wenjie and Cao, Ziqiang and Niu, Shuzi},\ntitle = {DailyDialog: A Manually Labelled Multi-turn Dialogue Dataset},\nbooktitle = {Proceedings of The 8th International Joint Conference on Natural Language Processing (IJCNLP 2017)},\nyear = {2017}\n}\n@inproceedings{chapuis-etal-2020-hierarchical,\n    title = \"Hierarchical Pre-training for Sequence Labelling in Spoken Dialog\",\n    author = \"Chapuis, Emile  and\n      Colombo, Pierre  and\n      Manica, Matteo  and\n      Labeau, Matthieu  and\n      Clavel, Chlo{'e}\",\n    booktitle = \"Findings of the Association for Computational Linguistics: EMNLP 2020\",\n    month = nov,\n    year = \"2020\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://www.aclweb.org/anthology/2020.findings-emnlp.239\",\n    doi = \"10.18653/v1/2020.findings-emnlp.239\",\n    pages = \"2636--2648\",\n    abstract = \"Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a\n        key component of spoken dialog systems. In this work, we propose a new approach to learn\n        generic representations adapted to spoken dialog, which we evaluate on a new benchmark we\n        call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE).\n        SILICONE is model-agnostic and contains 10 different datasets of various sizes.\n        We obtain our representations with a hierarchical encoder based on transformer architectures,\n        for which we extend two well-known pre-training objectives. Pre-training is performed on\n        OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We\n        demonstrate how hierarchical encoders achieve competitive results with consistently fewer\n        parameters compared to state-of-the-art models and we show their importance for both\n        pre-training and fine-tuning.\",\n}\n", "homepage": "http://yanran.li/dailydialog.html", "license": "", "features": {"Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Emotion": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 7, "names": ["anger", "disgust", "fear", "happiness", "no emotion", "sadness", "surprise"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "silicone", "config_name": "dyda_e", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 8547111, "num_examples": 87170, "dataset_name": "silicone"}, "validation": {"name": "validation", "num_bytes": 781445, "num_examples": 8069, "dataset_name": "silicone"}, "test": {"name": "test", "num_bytes": 757670, "num_examples": 7740, "dataset_name": "silicone"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/dyda/train.csv": {"num_bytes": 7519783, "checksum": "9710558dd08351e1544424fcf43f8aed5a3e99daba3e1f133e86acf2373fb4c6"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/dyda/dev.csv": {"num_bytes": 688589, "checksum": "66be9ec11f0234686cd338d5e7a374e1706ffd998b54fd3e6b72ef18598470b2"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/dyda/test.csv": {"num_bytes": 666553, "checksum": "3b711ebac2a1a64067aea5252c6bbbb1abe1f9f1434d2e94f9dd4a99706fc2ce"}}, "download_size": 8874925, "post_processing_size": null, "dataset_size": 10086226, "size_in_bytes": 18961151}, "iemocap": {"description": "The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection\n of resources for training, evaluating, and analyzing natural language understanding systems\n specifically designed for spoken language. All datasets are in the English language and cover a\n variety of domains including daily life, scripted scenarios, joint task completion, phone call\n conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant\n labels.\n", "citation": "@article{busso2008iemocap,\ntitle={IEMOCAP: Interactive emotional dyadic motion capture database},\nauthor={Busso, Carlos and Bulut, Murtaza and Lee, Chi-Chun and Kazemzadeh, Abe and Mower,\nEmily and Kim, Samuel and Chang, Jeannette N and Lee, Sungbok and Narayanan, Shrikanth S},\njournal={Language resources and evaluation},\nvolume={42},\nnumber={4},\npages={335},\nyear={2008},\npublisher={Springer}\n}\n@inproceedings{chapuis-etal-2020-hierarchical,\n    title = \"Hierarchical Pre-training for Sequence Labelling in Spoken Dialog\",\n    author = \"Chapuis, Emile  and\n      Colombo, Pierre  and\n      Manica, Matteo  and\n      Labeau, Matthieu  and\n      Clavel, Chlo{'e}\",\n    booktitle = \"Findings of the Association for Computational Linguistics: EMNLP 2020\",\n    month = nov,\n    year = \"2020\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://www.aclweb.org/anthology/2020.findings-emnlp.239\",\n    doi = \"10.18653/v1/2020.findings-emnlp.239\",\n    pages = \"2636--2648\",\n    abstract = \"Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a\n        key component of spoken dialog systems. In this work, we propose a new approach to learn\n        generic representations adapted to spoken dialog, which we evaluate on a new benchmark we\n        call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE).\n        SILICONE is model-agnostic and contains 10 different datasets of various sizes.\n        We obtain our representations with a hierarchical encoder based on transformer architectures,\n        for which we extend two well-known pre-training objectives. Pre-training is performed on\n        OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We\n        demonstrate how hierarchical encoders achieve competitive results with consistently fewer\n        parameters compared to state-of-the-art models and we show their importance for both\n        pre-training and fine-tuning.\",\n}\n", "homepage": "https://sail.usc.edu/iemocap/", "license": "", "features": {"Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Utterance_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Emotion": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 11, "names": ["ang", "dis", "exc", "fea", "fru", "hap", "neu", "oth", "sad", "sur", "xxx"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "silicone", "config_name": "iemocap", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 908180, "num_examples": 7213, "dataset_name": "silicone"}, "validation": {"name": "validation", "num_bytes": 100969, "num_examples": 805, "dataset_name": "silicone"}, "test": {"name": "test", "num_bytes": 254248, "num_examples": 2021, "dataset_name": "silicone"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/iemocap/train.csv": {"num_bytes": 834130, "checksum": "8bd85ac4c3081b9ab9a4607dc5726b563d7ace1dcc7b427ac751ed54d853086b"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/iemocap/dev.csv": {"num_bytes": 92272, "checksum": "8133acf20055d1e82733f70aa2dd381b01d2749909d77b2da8f32444c964eb51"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/iemocap/test.csv": {"num_bytes": 232376, "checksum": "7749583b50747f584ce4e8b7c137c546107f37c203c4c9a342ab03b8a4fa4406"}}, "download_size": 1158778, "post_processing_size": null, "dataset_size": 1263397, "size_in_bytes": 2422175}, "maptask": {"description": "The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection\n of resources for training, evaluating, and analyzing natural language understanding systems\n specifically designed for spoken language. All datasets are in the English language and cover a\n variety of domains including daily life, scripted scenarios, joint task completion, phone call\n conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant\n labels.\n", "citation": "@inproceedings{thompson1993hcrc,\ntitle={The HCRC map task corpus: natural dialogue for speech recognition},\nauthor={Thompson, Henry S and Anderson, Anne H and Bard, Ellen Gurman and Doherty-Sneddon,\nGwyneth and Newlands, Alison and Sotillo, Cathy},\nbooktitle={HUMAN LANGUAGE TECHNOLOGY: Proceedings of a Workshop Held at Plainsboro, New Jersey, March 21-24, 1993},\nyear={1993}\n}\n@inproceedings{chapuis-etal-2020-hierarchical,\n    title = \"Hierarchical Pre-training for Sequence Labelling in Spoken Dialog\",\n    author = \"Chapuis, Emile  and\n      Colombo, Pierre  and\n      Manica, Matteo  and\n      Labeau, Matthieu  and\n      Clavel, Chlo{'e}\",\n    booktitle = \"Findings of the Association for Computational Linguistics: EMNLP 2020\",\n    month = nov,\n    year = \"2020\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://www.aclweb.org/anthology/2020.findings-emnlp.239\",\n    doi = \"10.18653/v1/2020.findings-emnlp.239\",\n    pages = \"2636--2648\",\n    abstract = \"Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a\n        key component of spoken dialog systems. In this work, we propose a new approach to learn\n        generic representations adapted to spoken dialog, which we evaluate on a new benchmark we\n        call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE).\n        SILICONE is model-agnostic and contains 10 different datasets of various sizes.\n        We obtain our representations with a hierarchical encoder based on transformer architectures,\n        for which we extend two well-known pre-training objectives. Pre-training is performed on\n        OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We\n        demonstrate how hierarchical encoders achieve competitive results with consistently fewer\n        parameters compared to state-of-the-art models and we show their importance for both\n        pre-training and fine-tuning.\",\n}\n", "homepage": "http://groups.inf.ed.ac.uk/maptask/", "license": "", "features": {"Speaker": {"dtype": "string", "id": null, "_type": "Value"}, "Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_Act": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 12, "names": ["acknowledge", "align", "check", "clarify", "explain", "instruct", "query_w", "query_yn", "ready", "reply_n", "reply_w", "reply_y"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "silicone", "config_name": "maptask", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 1260413, "num_examples": 20905, "dataset_name": "silicone"}, "validation": {"name": "validation", "num_bytes": 178184, "num_examples": 2963, "dataset_name": "silicone"}, "test": {"name": "test", "num_bytes": 171806, "num_examples": 2894, "dataset_name": "silicone"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/maptask/train.txt": {"num_bytes": 821386, "checksum": "be52bd3de9c5fe134edba6d45ed112285f6db271dbe69a5a93557e6a0ef625fa"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/maptask/dev.txt": {"num_bytes": 115951, "checksum": "92f86ce1b15f1e05e16cee83732494ded4f572e0a474dd64ea952658601ccee7"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/maptask/test.txt": {"num_bytes": 111020, "checksum": "4758898044e6849c29084971bae9957ae06854e6606682b5d77984b5b861f87b"}}, "download_size": 1048357, "post_processing_size": null, "dataset_size": 1610403, "size_in_bytes": 2658760}, "meld_e": {"description": "The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection\n of resources for training, evaluating, and analyzing natural language understanding systems\n specifically designed for spoken language. All datasets are in the English language and cover a\n variety of domains including daily life, scripted scenarios, joint task completion, phone call\n conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant\n labels.\n", "citation": "@article{chen2018emotionlines,\ntitle={Emotionlines: An emotion corpus of multi-party conversations},\nauthor={Chen, Sheng-Yeh and Hsu, Chao-Chun and Kuo, Chuan-Chun and Ku, Lun-Wei and others},\njournal={arXiv preprint arXiv:1802.08379},\nyear={2018}\n}\n@inproceedings{chapuis-etal-2020-hierarchical,\n    title = \"Hierarchical Pre-training for Sequence Labelling in Spoken Dialog\",\n    author = \"Chapuis, Emile  and\n      Colombo, Pierre  and\n      Manica, Matteo  and\n      Labeau, Matthieu  and\n      Clavel, Chlo{'e}\",\n    booktitle = \"Findings of the Association for Computational Linguistics: EMNLP 2020\",\n    month = nov,\n    year = \"2020\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://www.aclweb.org/anthology/2020.findings-emnlp.239\",\n    doi = \"10.18653/v1/2020.findings-emnlp.239\",\n    pages = \"2636--2648\",\n    abstract = \"Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a\n        key component of spoken dialog systems. In this work, we propose a new approach to learn\n        generic representations adapted to spoken dialog, which we evaluate on a new benchmark we\n        call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE).\n        SILICONE is model-agnostic and contains 10 different datasets of various sizes.\n        We obtain our representations with a hierarchical encoder based on transformer architectures,\n        for which we extend two well-known pre-training objectives. Pre-training is performed on\n        OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We\n        demonstrate how hierarchical encoders achieve competitive results with consistently fewer\n        parameters compared to state-of-the-art models and we show their importance for both\n        pre-training and fine-tuning.\",\n}\n", "homepage": "https://affective-meld.github.io/", "license": "", "features": {"Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Speaker": {"dtype": "string", "id": null, "_type": "Value"}, "Emotion": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Utterance_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 7, "names": ["anger", "disgust", "fear", "joy", "neutral", "sadness", "surprise"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "silicone", "config_name": "meld_e", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 916337, "num_examples": 9989, "dataset_name": "silicone"}, "validation": {"name": "validation", "num_bytes": 100234, "num_examples": 1109, "dataset_name": "silicone"}, "test": {"name": "test", "num_bytes": 242352, "num_examples": 2610, "dataset_name": "silicone"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/meld/train.csv": {"num_bytes": 1131992, "checksum": "e734e07bba181798ac658d0a03d42d784a4fec0f7f9e48c996fc9bce7d3f7f20"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/meld/dev.csv": {"num_bytes": 123198, "checksum": "a69e5b1b50cd683432aefa7f1bdf355c581f05bd1b387bcec0dd2129a794ea1c"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/meld/test.csv": {"num_bytes": 297824, "checksum": "89d4948730222969a361909086d1d77a3c55d0dbb9e96da2ad90a5dc78b5077d"}}, "download_size": 1553014, "post_processing_size": null, "dataset_size": 1258923, "size_in_bytes": 2811937}, "meld_s": {"description": "The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection\n of resources for training, evaluating, and analyzing natural language understanding systems\n specifically designed for spoken language. All datasets are in the English language and cover a\n variety of domains including daily life, scripted scenarios, joint task completion, phone call\n conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant\n labels.\n", "citation": "@article{chen2018emotionlines,\ntitle={Emotionlines: An emotion corpus of multi-party conversations},\nauthor={Chen, Sheng-Yeh and Hsu, Chao-Chun and Kuo, Chuan-Chun and Ku, Lun-Wei and others},\njournal={arXiv preprint arXiv:1802.08379},\nyear={2018}\n}\n@inproceedings{chapuis-etal-2020-hierarchical,\n    title = \"Hierarchical Pre-training for Sequence Labelling in Spoken Dialog\",\n    author = \"Chapuis, Emile  and\n      Colombo, Pierre  and\n      Manica, Matteo  and\n      Labeau, Matthieu  and\n      Clavel, Chlo{'e}\",\n    booktitle = \"Findings of the Association for Computational Linguistics: EMNLP 2020\",\n    month = nov,\n    year = \"2020\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://www.aclweb.org/anthology/2020.findings-emnlp.239\",\n    doi = \"10.18653/v1/2020.findings-emnlp.239\",\n    pages = \"2636--2648\",\n    abstract = \"Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a\n        key component of spoken dialog systems. In this work, we propose a new approach to learn\n        generic representations adapted to spoken dialog, which we evaluate on a new benchmark we\n        call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE).\n        SILICONE is model-agnostic and contains 10 different datasets of various sizes.\n        We obtain our representations with a hierarchical encoder based on transformer architectures,\n        for which we extend two well-known pre-training objectives. Pre-training is performed on\n        OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We\n        demonstrate how hierarchical encoders achieve competitive results with consistently fewer\n        parameters compared to state-of-the-art models and we show their importance for both\n        pre-training and fine-tuning.\",\n}\n", "homepage": "https://affective-meld.github.io/", "license": "", "features": {"Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Speaker": {"dtype": "string", "id": null, "_type": "Value"}, "Sentiment": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Utterance_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 3, "names": ["negative", "neutral", "positive"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "silicone", "config_name": "meld_s", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 930405, "num_examples": 9989, "dataset_name": "silicone"}, "validation": {"name": "validation", "num_bytes": 101801, "num_examples": 1109, "dataset_name": "silicone"}, "test": {"name": "test", "num_bytes": 245873, "num_examples": 2610, "dataset_name": "silicone"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/meld/train.csv": {"num_bytes": 1131992, "checksum": "e734e07bba181798ac658d0a03d42d784a4fec0f7f9e48c996fc9bce7d3f7f20"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/meld/dev.csv": {"num_bytes": 123198, "checksum": "a69e5b1b50cd683432aefa7f1bdf355c581f05bd1b387bcec0dd2129a794ea1c"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/meld/test.csv": {"num_bytes": 297824, "checksum": "89d4948730222969a361909086d1d77a3c55d0dbb9e96da2ad90a5dc78b5077d"}}, "download_size": 1553014, "post_processing_size": null, "dataset_size": 1278079, "size_in_bytes": 2831093}, "mrda": {"description": "The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection\n of resources for training, evaluating, and analyzing natural language understanding systems\n specifically designed for spoken language. All datasets are in the English language and cover a\n variety of domains including daily life, scripted scenarios, joint task completion, phone call\n conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant\n labels.\n", "citation": "@techreport{shriberg2004icsi,\ntitle={The ICSI meeting recorder dialog act (MRDA) corpus},\nauthor={Shriberg, Elizabeth and Dhillon, Raj and Bhagat, Sonali and Ang, Jeremy and Carvey, Hannah},\nyear={2004},\ninstitution={INTERNATIONAL COMPUTER SCIENCE INST BERKELEY CA}\n}\n@inproceedings{chapuis-etal-2020-hierarchical,\n    title = \"Hierarchical Pre-training for Sequence Labelling in Spoken Dialog\",\n    author = \"Chapuis, Emile  and\n      Colombo, Pierre  and\n      Manica, Matteo  and\n      Labeau, Matthieu  and\n      Clavel, Chlo{'e}\",\n    booktitle = \"Findings of the Association for Computational Linguistics: EMNLP 2020\",\n    month = nov,\n    year = \"2020\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://www.aclweb.org/anthology/2020.findings-emnlp.239\",\n    doi = \"10.18653/v1/2020.findings-emnlp.239\",\n    pages = \"2636--2648\",\n    abstract = \"Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a\n        key component of spoken dialog systems. In this work, we propose a new approach to learn\n        generic representations adapted to spoken dialog, which we evaluate on a new benchmark we\n        call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE).\n        SILICONE is model-agnostic and contains 10 different datasets of various sizes.\n        We obtain our representations with a hierarchical encoder based on transformer architectures,\n        for which we extend two well-known pre-training objectives. Pre-training is performed on\n        OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We\n        demonstrate how hierarchical encoders achieve competitive results with consistently fewer\n        parameters compared to state-of-the-art models and we show their importance for both\n        pre-training and fine-tuning.\",\n}\n", "homepage": "https://www.aclweb.org/anthology/W04-2319", "license": "", "features": {"Utterance_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_Act": {"dtype": "string", "id": null, "_type": "Value"}, "Channel_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Speaker": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 5, "names": ["s", "d", "b", "f", "q"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "silicone", "config_name": "mrda", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 9998857, "num_examples": 83943, "dataset_name": "silicone"}, "validation": {"name": "validation", "num_bytes": 1143286, "num_examples": 9815, "dataset_name": "silicone"}, "test": {"name": "test", "num_bytes": 1807462, "num_examples": 15470, "dataset_name": "silicone"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/mrda/train.csv": {"num_bytes": 7977039, "checksum": "d7a963aac70eb80d315b76b9e71d82a7888532ef8c6752c84487e34dfce3b7eb"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/mrda/dev.csv": {"num_bytes": 903718, "checksum": "dc795c60b3825645a20dbb0700e279271fc52eef4bc297564a0227f608a19619"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/mrda/test.csv": {"num_bytes": 1425091, "checksum": "8426417d316cb0aa57f13e46e866e8964eed71f719212ab1da1c3cbaa23ea414"}}, "download_size": 10305848, "post_processing_size": null, "dataset_size": 12949605, "size_in_bytes": 23255453}, "oasis": {"description": "The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection\n of resources for training, evaluating, and analyzing natural language understanding systems\n specifically designed for spoken language. All datasets are in the English language and cover a\n variety of domains including daily life, scripted scenarios, joint task completion, phone call\n conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant\n labels.\n", "citation": "@inproceedings{leech2003generic,\ntitle={Generic speech act annotation for task-oriented dialogues},\nauthor={Leech, Geoffrey and Weisser, Martin},\nbooktitle={Proceedings of the corpus linguistics 2003 conference},\nvolume={16},\npages={441--446},\nyear={2003},\norganization={Lancaster: Lancaster University}\n}\n@inproceedings{chapuis-etal-2020-hierarchical,\n    title = \"Hierarchical Pre-training for Sequence Labelling in Spoken Dialog\",\n    author = \"Chapuis, Emile  and\n      Colombo, Pierre  and\n      Manica, Matteo  and\n      Labeau, Matthieu  and\n      Clavel, Chlo{'e}\",\n    booktitle = \"Findings of the Association for Computational Linguistics: EMNLP 2020\",\n    month = nov,\n    year = \"2020\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://www.aclweb.org/anthology/2020.findings-emnlp.239\",\n    doi = \"10.18653/v1/2020.findings-emnlp.239\",\n    pages = \"2636--2648\",\n    abstract = \"Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a\n        key component of spoken dialog systems. In this work, we propose a new approach to learn\n        generic representations adapted to spoken dialog, which we evaluate on a new benchmark we\n        call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE).\n        SILICONE is model-agnostic and contains 10 different datasets of various sizes.\n        We obtain our representations with a hierarchical encoder based on transformer architectures,\n        for which we extend two well-known pre-training objectives. Pre-training is performed on\n        OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We\n        demonstrate how hierarchical encoders achieve competitive results with consistently fewer\n        parameters compared to state-of-the-art models and we show their importance for both\n        pre-training and fine-tuning.\",\n}\n", "homepage": "http://groups.inf.ed.ac.uk/oasis/", "license": "", "features": {"Speaker": {"dtype": "string", "id": null, "_type": "Value"}, "Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_Act": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 42, "names": ["accept", "ackn", "answ", "answElab", "appreciate", "backch", "bye", "complete", "confirm", "correct", "direct", "directElab", "echo", "exclaim", "expressOpinion", "expressPossibility", "expressRegret", "expressWish", "greet", "hold", "identifySelf", "inform", "informCont", "informDisc", "informIntent", "init", "negate", "offer", "pardon", "raiseIssue", "refer", "refuse", "reqDirect", "reqInfo", "reqModal", "selfTalk", "suggest", "thank", "informIntent-hold", "correctSelf", "expressRegret-inform", "thank-identifySelf"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "silicone", "config_name": "oasis", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 887018, "num_examples": 12076, "dataset_name": "silicone"}, "validation": {"name": "validation", "num_bytes": 112185, "num_examples": 1513, "dataset_name": "silicone"}, "test": {"name": "test", "num_bytes": 119254, "num_examples": 1478, "dataset_name": "silicone"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/oasis/train.txt": {"num_bytes": 633398, "checksum": "5a6a42cc47e0276afde1189dbc604fb8c0cc46afab2b49de1c2c6d9b61d5ce16"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/oasis/dev.txt": {"num_bytes": 80400, "checksum": "a8d4e7d6fa8e582c80c7e3aef0d4a73fc9ac38a18f9aab10bac723ee45489c8e"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/oasis/test.txt": {"num_bytes": 88204, "checksum": "a22790b90f99b92d28a5beda6d0a8a7d107c195cb19bdfc4ce3f1df4f2ca901b"}}, "download_size": 802002, "post_processing_size": null, "dataset_size": 1118457, "size_in_bytes": 1920459}, "sem": {"description": "The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection\n of resources for training, evaluating, and analyzing natural language understanding systems\n specifically designed for spoken language. All datasets are in the English language and cover a\n variety of domains including daily life, scripted scenarios, joint task completion, phone call\n conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant\n labels.\n", "citation": "@article{mckeown2011semaine,\ntitle={The semaine database: Annotated multimodal records of emotionally colored conversations\nbetween a person and a limited agent},\nauthor={McKeown, Gary and Valstar, Michel and Cowie, Roddy and Pantic, Maja and Schroder, Marc},\njournal={IEEE transactions on affective computing},\nvolume={3},\nnumber={1},\npages={5--17},\nyear={2011},\npublisher={IEEE}\n}\n@inproceedings{chapuis-etal-2020-hierarchical,\n    title = \"Hierarchical Pre-training for Sequence Labelling in Spoken Dialog\",\n    author = \"Chapuis, Emile  and\n      Colombo, Pierre  and\n      Manica, Matteo  and\n      Labeau, Matthieu  and\n      Clavel, Chlo{'e}\",\n    booktitle = \"Findings of the Association for Computational Linguistics: EMNLP 2020\",\n    month = nov,\n    year = \"2020\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://www.aclweb.org/anthology/2020.findings-emnlp.239\",\n    doi = \"10.18653/v1/2020.findings-emnlp.239\",\n    pages = \"2636--2648\",\n    abstract = \"Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a\n        key component of spoken dialog systems. In this work, we propose a new approach to learn\n        generic representations adapted to spoken dialog, which we evaluate on a new benchmark we\n        call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE).\n        SILICONE is model-agnostic and contains 10 different datasets of various sizes.\n        We obtain our representations with a hierarchical encoder based on transformer architectures,\n        for which we extend two well-known pre-training objectives. Pre-training is performed on\n        OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We\n        demonstrate how hierarchical encoders achieve competitive results with consistently fewer\n        parameters compared to state-of-the-art models and we show their importance for both\n        pre-training and fine-tuning.\",\n}\n", "homepage": "https://ieeexplore.ieee.org/document/5959155", "license": "", "features": {"Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "NbPairInSession": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "SpeechTurn": {"dtype": "string", "id": null, "_type": "Value"}, "Speaker": {"dtype": "string", "id": null, "_type": "Value"}, "Sentiment": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 3, "names": ["Negative", "Neutral", "Positive"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "silicone", "config_name": "sem", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 496168, "num_examples": 4264, "dataset_name": "silicone"}, "validation": {"name": "validation", "num_bytes": 57896, "num_examples": 485, "dataset_name": "silicone"}, "test": {"name": "test", "num_bytes": 100072, "num_examples": 878, "dataset_name": "silicone"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/sem/train.csv": {"num_bytes": 389633, "checksum": "9fa9fc0851babf987b2f0c09c507eb46b5e9fab1a8680d8789a4c24af53cde9c"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/sem/dev.csv": {"num_bytes": 45859, "checksum": "43d93b46664e9e250ef59f551e02fdef54501b1bab5d2932bc3a97bf32d6c365"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/sem/test.csv": {"num_bytes": 78197, "checksum": "c8024e6e431776c9b838ad3f23290cb33b4a9d00bdc820e090bd62ef7b4a63db"}}, "download_size": 513689, "post_processing_size": null, "dataset_size": 654136, "size_in_bytes": 1167825}, "swda": {"description": "The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection\n of resources for training, evaluating, and analyzing natural language understanding systems\n specifically designed for spoken language. All datasets are in the English language and cover a\n variety of domains including daily life, scripted scenarios, joint task completion, phone call\n conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant\n labels.\n", "citation": "@article{stolcke2000dialogue,\ntitle={Dialogue act modeling for automatic tagging and recognition of conversational speech},\nauthor={Stolcke, Andreas and Ries, Klaus and Coccaro, Noah and Shriberg, Elizabeth and\nBates, Rebecca and Jurafsky, Daniel and Taylor, Paul and Martin, Rachel and Ess-Dykema,\nCarol Van and Meteer, Marie},\njournal={Computational linguistics},\nvolume={26},\nnumber={3},\npages={339--373},\nyear={2000},\npublisher={MIT Press}\n}\n@inproceedings{chapuis-etal-2020-hierarchical,\n    title = \"Hierarchical Pre-training for Sequence Labelling in Spoken Dialog\",\n    author = \"Chapuis, Emile  and\n      Colombo, Pierre  and\n      Manica, Matteo  and\n      Labeau, Matthieu  and\n      Clavel, Chlo{'e}\",\n    booktitle = \"Findings of the Association for Computational Linguistics: EMNLP 2020\",\n    month = nov,\n    year = \"2020\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://www.aclweb.org/anthology/2020.findings-emnlp.239\",\n    doi = \"10.18653/v1/2020.findings-emnlp.239\",\n    pages = \"2636--2648\",\n    abstract = \"Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a\n        key component of spoken dialog systems. In this work, we propose a new approach to learn\n        generic representations adapted to spoken dialog, which we evaluate on a new benchmark we\n        call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE).\n        SILICONE is model-agnostic and contains 10 different datasets of various sizes.\n        We obtain our representations with a hierarchical encoder based on transformer architectures,\n        for which we extend two well-known pre-training objectives. Pre-training is performed on\n        OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We\n        demonstrate how hierarchical encoders achieve competitive results with consistently fewer\n        parameters compared to state-of-the-art models and we show their importance for both\n        pre-training and fine-tuning.\",\n}\n", "homepage": "https://web.stanford.edu/~jurafsky/ws97/", "license": "", "features": {"Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_Act": {"dtype": "string", "id": null, "_type": "Value"}, "From_Caller": {"dtype": "string", "id": null, "_type": "Value"}, "To_Caller": {"dtype": "string", "id": null, "_type": "Value"}, "Topic": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Conv_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 46, "names": ["sd", "b", "sv", "%", "aa", "ba", "fc", "qw", "nn", "bk", "h", "qy^d", "bh", "^q", "bf", "fo_o_fw_\"_by_bc", "fo_o_fw_by_bc_\"", "na", "ad", "^2", "b^m", "qo", "qh", "^h", "ar", "ng", "br", "no", "fp", "qrr", "arp_nd", "t3", "oo_co_cc", "aap_am", "t1", "bd", "^g", "qw^d", "fa", "ft", "+", "x", "ny", "sv_fx", "qy_qr", "ba_fe"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "silicone", "config_name": "swda", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 20499788, "num_examples": 190709, "dataset_name": "silicone"}, "validation": {"name": "validation", "num_bytes": 2265898, "num_examples": 21203, "dataset_name": "silicone"}, "test": {"name": "test", "num_bytes": 291471, "num_examples": 2714, "dataset_name": "silicone"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/swda/train.csv": {"num_bytes": 14420223, "checksum": "37475282ef24b5d53a761f0576619f7c1fb1520b1af7dfaf306f13b6c9b60d57"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/swda/dev.csv": {"num_bytes": 1589677, "checksum": "000a1f0a805f275936748d6a4d61f928b18457c8c9e1e5149c56888d95c866b5"}, "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main/swda/test.csv": {"num_bytes": 217600, "checksum": "0584658976bd4b218351630ec282145238ce02d127d76cb31c1dd47ecdfa34f4"}}, "download_size": 16227500, "post_processing_size": null, "dataset_size": 23057157, "size_in_bytes": 39284657}}

dummy/dyda_da/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f4ca75f38c6924640a3b5845ff777e2ef292c5b9a266c3be46d7e66636f7229
+size 1070

dummy/dyda_e/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e371a39a0f6c2a07aea72a6836b1146b5abb69e02db33b81d386437e55cbb601
+size 1070

dummy/iemocap/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56f934b9b70bdeaf136430bbe21f772862a1baaf0c09ddca63f25b3b2b6a7269
+size 1180

dummy/maptask/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f035f797f9667738a8e2f3650bdc861d7ea6706dca420bb5095801501c628f0
+size 854

dummy/meld_e/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94fcf11e0c7b54caf1c73c17e6f3db9b504a9d440e8e30103c9cd3f094ddb7ed
+size 1486

dummy/meld_s/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09471cc7ad152c8e8e689d345d49268674a830d38dea6caf02565f693fbb6c94
+size 1486

dummy/mrda/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1cb6d7024fc6f67a1c116125b84cb152f843cb726a93d31037e13cbbf5a857d
+size 1115

dummy/oasis/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fe94e06b4853e1133906b418617460a7a5e2030b576db6a89aaacc8e5885daf
+size 1586

dummy/sem/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1cf1c4875acca9d68cba85eaf8ef7b73d287a62531b66e6ff05f6ce8901ad827
+size 1197

dummy/swda/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d9fea104d15d142749dc903db1a859c48116d8ca3a24a279343cdf3a6e5f2f9
+size 1064

silicone.py

@@ -0,0 +1,660 @@
+# coding=utf-8
+# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Lint as: python3
+"""The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark."""
+
+from __future__ import absolute_import, division, print_function
+
+import textwrap
+
+import pandas as pd
+import six
+
+import datasets
+
+
+_SILICONE_CITATION = """\
+@inproceedings{chapuis-etal-2020-hierarchical,
+    title = "Hierarchical Pre-training for Sequence Labelling in Spoken Dialog",
+    author = "Chapuis, Emile  and
+      Colombo, Pierre  and
+      Manica, Matteo  and
+      Labeau, Matthieu  and
+      Clavel, Chlo{\'e}",
+    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2020",
+    month = nov,
+    year = "2020",
+    address = "Online",
+    publisher = "Association for Computational Linguistics",
+    url = "https://www.aclweb.org/anthology/2020.findings-emnlp.239",
+    doi = "10.18653/v1/2020.findings-emnlp.239",
+    pages = "2636--2648",
+    abstract = "Sequence labelling tasks like Dialog Act and Emotion/Sentiment identification are a
+        key component of spoken dialog systems. In this work, we propose a new approach to learn
+        generic representations adapted to spoken dialog, which we evaluate on a new benchmark we
+        call Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE benchmark (SILICONE).
+        SILICONE is model-agnostic and contains 10 different datasets of various sizes.
+        We obtain our representations with a hierarchical encoder based on transformer architectures,
+        for which we extend two well-known pre-training objectives. Pre-training is performed on
+        OpenSubtitles: a large corpus of spoken dialog containing over 2.3 billion of tokens. We
+        demonstrate how hierarchical encoders achieve competitive results with consistently fewer
+        parameters compared to state-of-the-art models and we show their importance for both
+        pre-training and fine-tuning.",
+}
+"""
+
+_SILICONE_DESCRIPTION = """\
+The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark is a collection
+ of resources for training, evaluating, and analyzing natural language understanding systems
+ specifically designed for spoken language. All datasets are in the English language and cover a
+ variety of domains including daily life, scripted scenarios, joint task completion, phone call
+ conversations, and televsion dialogue. Some datasets additionally include emotion and/or sentimant
+ labels.
+"""
+
+_URL = "https://raw.githubusercontent.com/eusip/SILICONE-benchmark/main"
+
+SWDA_DA_DESCRIPTION = {
+    "sd": "Statement-non-opinion",
+    "b": "Acknowledge (Backchannel)",
+    "sv": "Statement-opinion",
+    "%": "Uninterpretable",
+    "aa": "Agree/Accept",
+    "ba": "Appreciation",
+    "fc": "Conventional-closing",
+    "qw": "Wh-Question",
+    "nn": "No Answers",
+    "bk": "Response Acknowledgement",
+    "h": "Hedge",
+    "qy^d": "Declarative Yes-No-Question",
+    "bh": "Backchannel in Question Form",
+    "^q": "Quotation",
+    "bf": "Summarize/Reformulate",
+    'fo_o_fw_"_by_bc': "Other",
+    'fo_o_fw_by_bc_"': "Other",
+    "na": "Affirmative Non-yes Answers",
+    "ad": "Action-directive",
+    "^2": "Collaborative Completion",
+    "b^m": "Repeat-phrase",
+    "qo": "Open-Question",
+    "qh": "Rhetorical-Question",
+    "^h": "Hold Before Answer/Agreement",
+    "ar": "Reject",
+    "ng": "Negative Non-no Answers",
+    "br": "Signal-non-understanding",
+    "no": "Other Answers",
+    "fp": "Conventional-opening",
+    "qrr": "Or-Clause",
+    "arp_nd": "Dispreferred Answers",
+    "t3": "3rd-party-talk",
+    "oo_co_cc": "Offers, Options Commits",
+    "aap_am": "Maybe/Accept-part",
+    "t1": "Downplayer",
+    "bd": "Self-talk",
+    "^g": "Tag-Question",
+    "qw^d": "Declarative Wh-Question",
+    "fa": "Apology",
+    "ft": "Thanking",
+    "+": "Unknown",
+    "x": "Unknown",
+    "ny": "Unknown",
+    "sv_fx": "Unknown",
+    "qy_qr": "Unknown",
+    "ba_fe": "Unknown",
+}
+
+MRDA_DA_DESCRIPTION = {
+    "s": "Statement/Subjective Statement",
+    "d": "Declarative Question",
+    "b": "Backchannel",
+    "f": '"Follow-me"',
+    "q": "Question",
+}
+
+IEMOCAP_E_DESCRIPTION = {
+    "ang": "Anger",
+    "dis": "Disgust",
+    "exc": "Excitement",
+    "fea": "Fear",
+    "fru": "Frustration",
+    "hap": "Happiness",
+    "neu": "Neutral",
+    "oth": "Other",
+    "sad": "Sadness",
+    "sur": "Surprise",
+    "xxx": "Unknown",
+}
+
+
+class SiliconeConfig(datasets.BuilderConfig):
+    """BuilderConfig for SILICONE."""
+
+    def __init__(
+        self,
+        text_features,
+        label_column,
+        data_url,
+        citation,
+        url,
+        label_classes=None,
+        **kwargs,
+    ):
+        """BuilderConfig for SILICONE.
+        Args:
+          text_features: `dict[string, string]`, map from the name of the feature
+            dict for each text field to the name of the column in the tsv file
+          label_column: `string`, name of the column in the csv/txt file corresponding
+            to the label
+          data_url: `string`, url to download the csv/text file from
+          citation: `string`, citation for the data set
+          url: `string`, url for information about the data set
+          label_classes: `list[string]`, the list of classes if the label is
+            categorical. If not provided, then the label will be of type
+            `datasets.Value('float32')`.
+          **kwargs: keyword arguments forwarded to super.
+        """
+        super(SiliconeConfig, self).__init__(version=datasets.Version("1.0.0", ""), **kwargs)
+        self.text_features = text_features
+        self.label_column = label_column
+        self.label_classes = label_classes
+        self.data_url = data_url
+        self.citation = citation
+        self.url = url
+
+
+class Silicone(datasets.GeneratorBasedBuilder):
+    """The Sequence labellIng evaLuatIon benChmark fOr spoken laNguagE (SILICONE) benchmark."""
+
+    BUILDER_CONFIGS = [
+        SiliconeConfig(
+            name="dyda_da",
+            description=textwrap.dedent(
+                """\
+            The DailyDialog Act Corpus contains multi-turn dialogues and is supposed to reflect daily
+            communication by covering topics about daily life. The dataset is manually labelled with
+             dialog act and emotions. It is the third biggest corpus of SILICONE with 102k utterances."""
+            ),
+            text_features={
+                "Utterance": "Utterance",
+                "Dialogue_Act": "Dialogue_Act",
+                "Dialogue_ID": "Dialogue_ID",
+            },
+            label_classes=["commissive", "directive", "inform", "question"],
+            label_column="Dialogue_Act",
+            data_url={
+                "train": _URL + "/dyda/train.csv",
+                "dev": _URL + "/dyda/dev.csv",
+                "test": _URL + "/dyda/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @InProceedings{li2017dailydialog,
+            author = {Li, Yanran and Su, Hui and Shen, Xiaoyu and Li, Wenjie and Cao, Ziqiang and Niu, Shuzi},
+            title = {DailyDialog: A Manually Labelled Multi-turn Dialogue Dataset},
+            booktitle = {Proceedings of The 8th International Joint Conference on Natural Language Processing (IJCNLP 2017)},
+            year = {2017}
+            }"""
+            ),
+            url="http://yanran.li/dailydialog.html",
+        ),
+        SiliconeConfig(
+            name="dyda_e",
+            description=textwrap.dedent(
+                """\
+            The DailyDialog Act Corpus contains multi-turn dialogues and is supposed to reflect daily
+            communication by covering topics about daily life. The dataset is manually labelled with
+             dialog act and emotions. It is the third biggest corpus of SILICONE with 102k utterances."""
+            ),
+            text_features={
+                "Utterance": "Utterance",
+                "Emotion": "Emotion",
+                "Dialogue_ID": "Dialogue_ID",
+            },
+            label_classes=["anger", "disgust", "fear", "happiness", "no emotion", "sadness", "surprise"],
+            label_column="Emotion",
+            data_url={
+                "train": _URL + "/dyda/train.csv",
+                "dev": _URL + "/dyda/dev.csv",
+                "test": _URL + "/dyda/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @InProceedings{li2017dailydialog,
+            author = {Li, Yanran and Su, Hui and Shen, Xiaoyu and Li, Wenjie and Cao, Ziqiang and Niu, Shuzi},
+            title = {DailyDialog: A Manually Labelled Multi-turn Dialogue Dataset},
+            booktitle = {Proceedings of The 8th International Joint Conference on Natural Language Processing (IJCNLP 2017)},
+            year = {2017}
+            }"""
+            ),
+            url="http://yanran.li/dailydialog.html",
+        ),
+        SiliconeConfig(
+            name="iemocap",
+            description=textwrap.dedent(
+                """\
+            The IEMOCAP database is a multi-modal database of ten speakers. It consists of dyadic
+            sessions where actors perform improvisations or scripted scenarios. Emotion categories
+            are: anger, happiness, sadness, neutral, excitement, frustration, fear, surprise, and other.
+            There is no official split of this dataset."""
+            ),
+            text_features={
+                "Dialogue_ID": "Dialogue_ID",
+                "Utterance_ID": "Utterance_ID",
+                "Utterance": "Utterance",
+                "Emotion": "Emotion",
+            },
+            label_classes=list(six.iterkeys(IEMOCAP_E_DESCRIPTION)),
+            label_column="Emotion",
+            data_url={
+                "train": _URL + "/iemocap/train.csv",
+                "dev": _URL + "/iemocap/dev.csv",
+                "test": _URL + "/iemocap/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @article{busso2008iemocap,
+            title={IEMOCAP: Interactive emotional dyadic motion capture database},
+            author={Busso, Carlos and Bulut, Murtaza and Lee, Chi-Chun and Kazemzadeh, Abe and Mower,
+            Emily and Kim, Samuel and Chang, Jeannette N and Lee, Sungbok and Narayanan, Shrikanth S},
+            journal={Language resources and evaluation},
+            volume={42},
+            number={4},
+            pages={335},
+            year={2008},
+            publisher={Springer}
+            }"""
+            ),
+            url="https://sail.usc.edu/iemocap/",
+        ),
+        SiliconeConfig(
+            name="maptask",
+            description=textwrap.dedent(
+                """\
+            The HCRC MapTask Corpus was constructed through the verbal collaboration of participants
+             in order to construct a map route. This corpus is small (27k utterances). As there is
+             no standard train/dev/test split performance depends on the split."""
+            ),
+            text_features={
+                "Speaker": "Speaker",
+                "Utterance": "Utterance",
+                "Dialogue_Act": "Dialogue_Act",
+            },
+            label_classes=[
+                "acknowledge",
+                "align",
+                "check",
+                "clarify",
+                "explain",
+                "instruct",
+                "query_w",
+                "query_yn",
+                "ready",
+                "reply_n",
+                "reply_w",
+                "reply_y",
+            ],
+            label_column="Dialogue_Act",
+            data_url={
+                "train": _URL + "/maptask/train.txt",
+                "dev": _URL + "/maptask/dev.txt",
+                "test": _URL + "/maptask/test.txt",
+            },
+            citation=textwrap.dedent(
+                """\
+            @inproceedings{thompson1993hcrc,
+            title={The HCRC map task corpus: natural dialogue for speech recognition},
+            author={Thompson, Henry S and Anderson, Anne H and Bard, Ellen Gurman and Doherty-Sneddon,
+            Gwyneth and Newlands, Alison and Sotillo, Cathy},
+            booktitle={HUMAN LANGUAGE TECHNOLOGY: Proceedings of a Workshop Held at Plainsboro, New Jersey, March 21-24, 1993},
+            year={1993}
+            }"""
+            ),
+            url="http://groups.inf.ed.ac.uk/maptask/",
+        ),
+        SiliconeConfig(
+            name="meld_e",
+            description=textwrap.dedent(
+                """\
+            The Multimodal EmotionLines Dataset enhances and extends the EmotionLines dataset where
+            multiple speakers participate in the dialogue."""
+            ),
+            text_features={
+                "Utterance": "Utterance",
+                "Speaker": "Speaker",
+                "Emotion": "Emotion",
+                "Dialogue_ID": "Dialogue_ID",
+                "Utterance_ID": "Utterance_ID",
+            },
+            label_classes=["anger", "disgust", "fear", "joy", "neutral", "sadness", "surprise"],
+            label_column="Emotion",
+            data_url={
+                "train": _URL + "/meld/train.csv",
+                "dev": _URL + "/meld/dev.csv",
+                "test": _URL + "/meld/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @article{chen2018emotionlines,
+            title={Emotionlines: An emotion corpus of multi-party conversations},
+            author={Chen, Sheng-Yeh and Hsu, Chao-Chun and Kuo, Chuan-Chun and Ku, Lun-Wei and others},
+            journal={arXiv preprint arXiv:1802.08379},
+            year={2018}
+            }"""
+            ),
+            url="https://affective-meld.github.io/",
+        ),
+        SiliconeConfig(
+            name="meld_s",
+            description=textwrap.dedent(
+                """\
+            The Multimodal EmotionLines Dataset enhances and extends the EmotionLines dataset where
+            multiple speakers participate in the dialogue."""
+            ),
+            text_features={
+                "Utterance": "Utterance",
+                "Speaker": "Speaker",
+                "Sentiment": "Sentiment",
+                "Dialogue_ID": "Dialogue_ID",
+                "Utterance_ID": "Utterance_ID",
+            },
+            label_classes=["negative", "neutral", "positive"],
+            label_column="Sentiment",
+            data_url={
+                "train": _URL + "/meld/train.csv",
+                "dev": _URL + "/meld/dev.csv",
+                "test": _URL + "/meld/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @article{chen2018emotionlines,
+            title={Emotionlines: An emotion corpus of multi-party conversations},
+            author={Chen, Sheng-Yeh and Hsu, Chao-Chun and Kuo, Chuan-Chun and Ku, Lun-Wei and others},
+            journal={arXiv preprint arXiv:1802.08379},
+            year={2018}
+            }"""
+            ),
+            url="https://affective-meld.github.io/",
+        ),
+        SiliconeConfig(
+            name="mrda",
+            description=textwrap.dedent(
+                """\
+            ICSI MRDA Corpus consist of transcripts of multi-party meetings hand-annotated with dialog
+            acts. It is the second biggest dataset with around 110k utterances."""
+            ),
+            text_features={
+                "Utterance_ID": "Utterance_ID",
+                "Dialogue_Act": "Dialogue_Act",
+                "Channel_ID": "Channel_ID",
+                "Speaker": "Speaker",
+                "Dialogue_ID": "Dialogue_ID",
+                "Utterance": "Utterance",
+            },
+            label_classes=list(six.iterkeys(MRDA_DA_DESCRIPTION)),
+            label_column="Dialogue_Act",
+            data_url={
+                "train": _URL + "/mrda/train.csv",
+                "dev": _URL + "/mrda/dev.csv",
+                "test": _URL + "/mrda/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @techreport{shriberg2004icsi,
+            title={The ICSI meeting recorder dialog act (MRDA) corpus},
+            author={Shriberg, Elizabeth and Dhillon, Raj and Bhagat, Sonali and Ang, Jeremy and Carvey, Hannah},
+            year={2004},
+            institution={INTERNATIONAL COMPUTER SCIENCE INST BERKELEY CA}
+            }"""
+            ),
+            url="https://www.aclweb.org/anthology/W04-2319",
+        ),
+        SiliconeConfig(
+            name="oasis",
+            description=textwrap.dedent(
+                """\
+            The Bt Oasis Corpus (Oasis) contains the transcripts of live calls made to the BT and
+            operator services. This corpus is rather small (15k utterances). There is no standard
+            train/dev/test split."""
+            ),
+            text_features={
+                "Speaker": "Speaker",
+                "Utterance": "Utterance",
+                "Dialogue_Act": "Dialogue_Act",
+            },
+            label_classes=[
+                "accept",
+                "ackn",
+                "answ",
+                "answElab",
+                "appreciate",
+                "backch",
+                "bye",
+                "complete",
+                "confirm",
+                "correct",
+                "direct",
+                "directElab",
+                "echo",
+                "exclaim",
+                "expressOpinion",
+                "expressPossibility",
+                "expressRegret",
+                "expressWish",
+                "greet",
+                "hold",
+                "identifySelf",
+                "inform",
+                "informCont",
+                "informDisc",
+                "informIntent",
+                "init",
+                "negate",
+                "offer",
+                "pardon",
+                "raiseIssue",
+                "refer",
+                "refuse",
+                "reqDirect",
+                "reqInfo",
+                "reqModal",
+                "selfTalk",
+                "suggest",
+                "thank",
+                "informIntent-hold",
+                "correctSelf",
+                "expressRegret-inform",
+                "thank-identifySelf",
+            ],
+            label_column="Dialogue_Act",
+            data_url={
+                "train": _URL + "/oasis/train.txt",
+                "dev": _URL + "/oasis/dev.txt",
+                "test": _URL + "/oasis/test.txt",
+            },
+            citation=textwrap.dedent(
+                """\
+            @inproceedings{leech2003generic,
+            title={Generic speech act annotation for task-oriented dialogues},
+            author={Leech, Geoffrey and Weisser, Martin},
+            booktitle={Proceedings of the corpus linguistics 2003 conference},
+            volume={16},
+            pages={441--446},
+            year={2003},
+            organization={Lancaster: Lancaster University}
+            }"""
+            ),
+            url="http://groups.inf.ed.ac.uk/oasis/",
+        ),
+        SiliconeConfig(
+            name="sem",
+            description=textwrap.dedent(
+                """\
+            The SEMAINE database comes from the Sustained Emotionally coloured Human-Machine Interaction
+            using Nonverbal Expression project. This dataset has been annotated on three sentiments
+            labels: positive, negative and neutral. It is built on Multimodal Wizard of Oz experiment
+            where participants held conversations with an operator who adopted various roles designed
+            to evoke emotional reactions. There is no official split on this dataset."""
+            ),
+            text_features={
+                "Utterance": "Utterance",
+                "NbPairInSession": "NbPairInSession",
+                "Dialogue_ID": "Dialogue_ID",
+                "SpeechTurn": "SpeechTurn",
+                "Speaker": "Speaker",
+                "Sentiment": "Sentiment",
+            },
+            label_classes=["Negative", "Neutral", "Positive"],
+            label_column="Sentiment",
+            data_url={
+                "train": _URL + "/sem/train.csv",
+                "dev": _URL + "/sem/dev.csv",
+                "test": _URL + "/sem/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @article{mckeown2011semaine,
+            title={The semaine database: Annotated multimodal records of emotionally colored conversations
+            between a person and a limited agent},
+            author={McKeown, Gary and Valstar, Michel and Cowie, Roddy and Pantic, Maja and Schroder, Marc},
+            journal={IEEE transactions on affective computing},
+            volume={3},
+            number={1},
+            pages={5--17},
+            year={2011},
+            publisher={IEEE}
+            }"""
+            ),
+            url="https://ieeexplore.ieee.org/document/5959155",
+        ),
+        SiliconeConfig(
+            name="swda",
+            description=textwrap.dedent(
+                """\
+            Switchboard Dialog Act Corpus (SwDA) is a telephone speech corpus consisting of two-sided
+            telephone conversations with provided topics. This dataset includes additional features
+            such as speaker id and topic information."""
+            ),
+            text_features={
+                "Utterance": "Utterance",
+                "Dialogue_Act": "Dialogue_Act",
+                "From_Caller": "From_Caller",
+                "To_Caller": "To_Caller",
+                "Topic": "Topic",
+                "Dialogue_ID": "Dialogue_ID",
+                "Conv_ID": "Conv_ID",
+            },
+            label_classes=list(six.iterkeys(SWDA_DA_DESCRIPTION)),
+            label_column="Dialogue_Act",
+            data_url={
+                "train": _URL + "/swda/train.csv",
+                "dev": _URL + "/swda/dev.csv",
+                "test": _URL + "/swda/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @article{stolcke2000dialogue,
+            title={Dialogue act modeling for automatic tagging and recognition of conversational speech},
+            author={Stolcke, Andreas and Ries, Klaus and Coccaro, Noah and Shriberg, Elizabeth and
+            Bates, Rebecca and Jurafsky, Daniel and Taylor, Paul and Martin, Rachel and Ess-Dykema,
+            Carol Van and Meteer, Marie},
+            journal={Computational linguistics},
+            volume={26},
+            number={3},
+            pages={339--373},
+            year={2000},
+            publisher={MIT Press}
+            }"""
+            ),
+            url="https://web.stanford.edu/~jurafsky/ws97/",
+        ),
+    ]
+
+    def _info(self):
+        features = {text_feature: datasets.Value("string") for text_feature in six.iterkeys(self.config.text_features)}
+        if self.config.label_classes:
+            features["Label"] = datasets.features.ClassLabel(names=self.config.label_classes)
+        features["Idx"] = datasets.Value("int32")
+        return datasets.DatasetInfo(
+            description=_SILICONE_DESCRIPTION,
+            features=datasets.Features(features),
+            homepage=self.config.url,
+            citation=self.config.citation + "\n" + _SILICONE_CITATION,
+        )
+
+    def _split_generators(self, dl_manager):
+        data_files = dl_manager.download(self.config.data_url)
+        splits = []
+        splits.append(
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={
+                    "data_file": data_files["train"],
+                    "split": "train",
+                },
+            )
+        )
+        splits.append(
+            datasets.SplitGenerator(
+                name=datasets.Split.VALIDATION,
+                gen_kwargs={
+                    "data_file": data_files["dev"],
+                    "split": "dev",
+                },
+            )
+        )
+        splits.append(
+            datasets.SplitGenerator(
+                name=datasets.Split.TEST,
+                gen_kwargs={
+                    "data_file": data_files["test"],
+                    "split": "test",
+                },
+            )
+        )
+        return splits
+
+    def _generate_examples(self, data_file, split):
+        if self.config.name not in ("maptask", "iemocap", "oasis"):
+            df = pd.read_csv(data_file, delimiter=",", header=0, quotechar='"', dtype=str)[
+                six.iterkeys(self.config.text_features)
+            ]
+
+        if self.config.name == "iemocap":
+            df = pd.read_csv(
+                data_file,
+                delimiter=",",
+                header=0,
+                quotechar='"',
+                names=["Dialogue_ID", "Utterance_ID", "Utterance", "Emotion", "Valence", "Activation", "Dominance"],
+                dtype=str,
+            )[six.iterkeys(self.config.text_features)]
+
+        if self.config.name in ("maptask", "oasis"):
+            df = pd.read_csv(data_file, delimiter="|", names=["Speaker", "Utterance", "Dialogue_Act"], dtype=str)[
+                six.iterkeys(self.config.text_features)
+            ]
+
+        rows = df.to_dict(orient="records")
+
+        for n, row in enumerate(rows):
+            example = row
+            example["Idx"] = n
+
+            if self.config.label_column in example:
+                label = example[self.config.label_column]
+                example["Label"] = label
+
+            yield example["Idx"], example