Update files from the datasets library (from 1.6.0)

Release notes: https://github.com/huggingface/datasets/releases/tag/1.6.0

.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 +1,285 @@
+---
+annotations_creators:
+- expert-generated
+language_creators:
+- expert-generated
+languages:
+  dihana:
+    - es
+  ilisten:
+    - it
+  loria:
+    - fr
+  maptask:
+    - en
+  vm2:
+    - de
+licenses:
+- cc-by-sa-4-0
+multilinguality:
+- multilingual
+size_categories:
+- 10K<n<100K
+source_datasets:
+- original
+task_categories:
+- sequence-modeling
+- text-classification
+task_ids:
+  dihana:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-dialogue-act-classification
+  ilisten:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-dialogue-act-classification
+  loria:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-dialogue-act-classification
+  maptask:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-dialogue-act-classification
+  vm2:
+  - dialogue-modeling
+  - language-modeling
+  - text-classification-other-dialogue-act-classification
+---
+
+# Dataset Card for MIAM
+
+## Table of Contents
+- [Dataset Card for MIAM](#dataset-card-for-miam)
+  - [Table of Contents](#table-of-contents)
+  - [Dataset Description](#dataset-description)
+    - [Dataset Summary](#dataset-summary)
+    - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
+    - [Languages](#languages)
+  - [Dataset Structure](#dataset-structure)
+    - [Data Instances](#data-instances)
+      - [Dihana Corpus](#dihana-corpus)
+      - [iLISTEN Corpus](#ilisten-corpus)
+      - [LORIA Corpus](#loria-corpus)
+      - [HCRC MapTask Corpus](#hcrc-maptask-corpus)
+      - [VERBMOBIL](#verbmobil)
+    - [Data Fields](#data-fields)
+    - [Data Splits](#data-splits)
+  - [Dataset Creation](#dataset-creation)
+    - [Curation Rationale](#curation-rationale)
+    - [Source Data](#source-data)
+      - [Initial Data Collection and Normalization](#initial-data-collection-and-normalization)
+      - [Who are the source language producers?](#who-are-the-source-language-producers)
+    - [Annotations](#annotations)
+      - [Annotation process](#annotation-process)
+      - [Who are the annotators?](#who-are-the-annotators)
+    - [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)
+    - [Benchmark Curators](#benchmark-curators)
+    - [Licensing Information](#licensing-information)
+    - [Citation Information](#citation-information)
+
+## Dataset Description
+
+- **Homepage:** [N/A]
+- **Repository:** [N/A]
+- **Paper:** [N/A]
+- **Leaderboard:** [N/A]
+- **Point of Contact:** [N/A]
+
+### Dataset Summary
+
+Multilingual dIalogAct benchMark is a collection of resources for training, evaluating, and
+analyzing natural language understanding systems specifically designed for spoken language. Datasets
+are in English, French, German, Italian and Spanish. They cover a variety of domains including
+spontaneous speech, scripted scenarios, and joint task completion. All datasets contain dialogue act
+labels.
+
+### Supported Tasks and Leaderboards
+
+[More Information Needed]
+
+### Languages
+
+English, French, German, Italian, Spanish.
+
+## Dataset Structure
+
+### Data Instances
+
+#### Dihana Corpus
+For the `dihana` configuration one example from the dataset is:
+```
+{
+  'Speaker': 'U',
+  'Utterance': 'Hola , quería obtener el horario para ir a Valencia',
+  'Dialogue_Act': 9,  # 'Pregunta' ('Request')
+  'Dialogue_ID': '0',
+  'File_ID': 'B209_BA5c3',
+}
+```
+
+#### iLISTEN Corpus
+For the `ilisten` configuration one example from the dataset is:
+```
+{
+  'Speaker': 'T_11_U11',
+  'Utterance': 'ok, grazie per le informazioni',
+  'Dialogue_Act': 6,  # 'KIND-ATTITUDE_SMALL-TALK'
+  'Dialogue_ID': '0',
+}
+```
+
+#### LORIA Corpus
+For the `loria` configuration one example from the dataset is:
+```
+{
+  'Speaker': 'Samir',
+  'Utterance': 'Merci de votre visite, bonne chance, et à la prochaine !',
+  'Dialogue_Act': 21,  # 'quit'
+  'Dialogue_ID': '5',
+  'File_ID': 'Dial_20111128_113927',
+}
+```
+
+#### HCRC MapTask Corpus
+For the `maptask` configuration one example from the dataset is:
+```
+{
+  'Speaker': 'f',
+  'Utterance': 'is it underneath the rope bridge or to the left',
+  'Dialogue_Act': 6,  # 'query_w'
+  'Dialogue_ID': '0',
+  'File_ID': 'q4ec1',
+}
+```
+
+#### VERBMOBIL
+For the `vm2` configuration one example from the dataset is:
+```
+{
+  'Utterance': 'ja was sind viereinhalb Stunden Bahngerüttel gegen siebzig Minuten Turbulenzen im Flugzeug',
+  'Utterance': 'Utterance',
+  'Dialogue_Act': 'Dialogue_Act',  # 'INFORM'
+  'Speaker': 'A',
+  'Dialogue_ID': '66',
+}
+```
+
+### Data Fields
+
+For the `dihana` 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 'Afirmacion' (0) [Feedback_positive], 'Apertura' (1) [Opening], 'Cierre' (2) [Closing], 'Confirmacion' (3) [Acknowledge], 'Espera' (4) [Hold], 'Indefinida' (5) [Undefined], 'Negacion' (6) [Feedback_negative], 'No_entendido' (7) [Request_clarify], 'Nueva_consulta' (8) [New_request], 'Pregunta' (9) [Request] or 'Respuesta' (10) [Reply].
+- `Dialogue_ID`: identifier of the dialogue as a string.
+- `File_ID`: identifier of the source file as a string.
+
+For the `ilisten` 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 'AGREE' (0), 'ANSWER' (1), 'CLOSING' (2), 'ENCOURAGE-SORRY' (3), 'GENERIC-ANSWER' (4), 'INFO-REQUEST' (5), 'KIND-ATTITUDE_SMALL-TALK' (6), 'OFFER-GIVE-INFO' (7), 'OPENING' (8), 'PERSUASION-SUGGEST' (9), 'QUESTION' (10), 'REJECT' (11), 'SOLICITATION-REQ_CLARIFICATION' (12), 'STATEMENT' (13) or 'TALK-ABOUT-SELF' (14).
+- `Dialogue_ID`: identifier of the dialogue as a string.
+
+For the `loria` 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 'ack' (0), 'ask' (1), 'find_mold' (2), 'find_plans' (3), 'first_step' (4), 'greet' (5), 'help' (6), 'inform' (7), 'inform_engine' (8), 'inform_job' (9), 'inform_material_space' (10), 'informer_conditioner' (11), 'informer_decoration' (12), 'informer_elcomps' (13), 'informer_end_manufacturing' (14), 'kindAtt' (15), 'manufacturing_reqs' (16), 'next_step' (17), 'no' (18), 'other' (19), 'quality_control' (20), 'quit' (21), 'reqRep' (22), 'security_policies' (23), 'staff_enterprise' (24), 'staff_job' (25), 'studies_enterprise' (26), 'studies_job' (27), 'todo_failure' (28), 'todo_irreparable' (29), 'yes' (30)
+- `Dialogue_ID`: identifier of the dialogue as a string.
+- `File_ID`: identifier of the source file as a string.
+
+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).
+- `Dialogue_ID`: identifier of the dialogue as a string.
+- `File_ID`: identifier of the source file as a string.
+
+For the `vm2` configuration, the different fields are:
+- `Utterance`: Utterance as a string.
+- `Dialogue_Act`: Dialogue act label of the utterance. It can be one of 'ACCEPT' (0), 'BACKCHANNEL' (1), 'BYE' (2), 'CLARIFY' (3), 'CLOSE' (4), 'COMMIT' (5), 'CONFIRM' (6), 'DEFER' (7), 'DELIBERATE' (8), 'DEVIATE_SCENARIO' (9), 'EXCLUDE' (10), 'EXPLAINED_REJECT' (11), 'FEEDBACK' (12), 'FEEDBACK_NEGATIVE' (13), 'FEEDBACK_POSITIVE' (14), 'GIVE_REASON' (15), 'GREET' (16), 'INFORM' (17), 'INIT' (18), 'INTRODUCE' (19), 'NOT_CLASSIFIABLE' (20), 'OFFER' (21), 'POLITENESS_FORMULA' (22), 'REJECT' (23), 'REQUEST' (24), 'REQUEST_CLARIFY' (25), 'REQUEST_COMMENT' (26), 'REQUEST_COMMIT' (27), 'REQUEST_SUGGEST' (28), 'SUGGEST' (29), 'THANK' (30).
+- `Speaker`: Speaker as a string.
+- `Dialogue_ID`: identifier of the dialogue as a string.
+
+### Data Splits
+
+| Dataset name | Train | Valid | Test |
+| ------------ | ----- | ----- | ---- |
+| dihana       | 19063 | 2123  | 2361 |
+| ilisten      | 1986  | 230   | 971  |
+| loria        | 8465  | 942   | 1047 |
+| maptask      | 25382 | 5221  | 5335 |
+| vm2          | 25060 | 2860  | 2855 |
+
+## 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
+
+Anonymous
+
+### 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
+
+```
+@unpublished{
+anonymous2021cross-lingual,
+title={Cross-Lingual Pretraining Methods for Spoken Dialog},
+author={Anonymous},
+journal={OpenReview Preprint},
+year={2021},
+url{https://openreview.net/forum?id=c1oDhu_hagR},
+note={anonymous preprint under review}
+}
+```

dataset_infos.json

@@ -0,0 +1 @@
+{"dihana": {"description": "Multilingual dIalogAct benchMark is a collection of resources for training, evaluating, and\nanalyzing natural language understanding systems specifically designed for spoken language. Datasets\nare in English, French, German, Italian and Spanish. They cover a variety of domains including\nspontaneous speech, scripted scenarios, and joint task completion. Some datasets additionally include\nemotion and/or sentimant labels.\n", "citation": "@inproceedings{benedi2006design,\ntitle={Design and acquisition of a telephone spontaneous speech dialogue corpus in Spanish: DIHANA},\nauthor={Bened{\\i}, Jos{'e}-Miguel and Lleida, Eduardo and Varona, Amparo and Castro, Mar{\\i}a-Jos{'e} and Galiano, Isabel and Justo, Raquel and L{'o}pez, I and Miguel, Antonio},\nbooktitle={Fifth International Conference on Language Resources and Evaluation (LREC)},\npages={1636--1639},\nyear={2006}\n}\n@inproceedings{post2013improved,\ntitle={Improved speech-to-text translation with the Fisher and Callhome Spanish--English speech translation corpus},\nauthor={Post, Matt and Kumar, Gaurav and Lopez, Adam and Karakos, Damianos and Callison-Burch, Chris and Khudanpur, Sanjeev},\nbooktitle={Proc. IWSLT},\nyear={2013}\n}\n@article{coria2005predicting,\ntitle={Predicting obligation dialogue acts from prosodic and speaker infomation},\nauthor={Coria, S and Pineda, L},\njournal={Research on Computing Science (ISSN 1665-9899), Centro de Investigacion en Computacion, Instituto Politecnico Nacional, Mexico City},\nyear={2005}\n}\n@inproceedings{anonymous,\n    title = \"Cross-Lingual Pretraining Methods for Spoken Dialog\",\n    author = \"Anonymous\",\n    booktitle = \"Transactions of the Association for Computational Linguistics\",\n    month = ,\n    year = \"\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"\",\n    doi = \"\",\n    pages = \"\",\n    abstract = \"There has been an increasing interest among NLP researchers towards learning generic\n    representations. However, in the field of multilingual spoken dialogue systems, this problem\n    remains overlooked. Indeed most of the pre-training methods focus on learning representations\n    for written and non-conversational data or are restricted to the monolingual setting. In this\n    work we (1) generalise existing losses to the multilingual setting, (2) develop a new set of\n    losses to leverage parallel conversations when available. These losses improve the learning of\n    representations by fostering the deep encoder to better learn contextual dependencies. The\n    pre-training relies on OpenSubtitles, a huge multilingual corpus that is composed of 24.3G tokens;\n    a by-product of the pre-processing includes multilingual aligned conversations. We also introduce\n    two new multilingual tasks and a new benchmark on multilingual dialogue act labels called MIAM.\n    We validate our pre-training on the three aforementioned tasks and show that our model using our\n    newly designed losses achieves better performances than existing models. Our implementation will\n    be available on github.com and pre-processed data will be available in Datasets (Wolf et al., 2020).\",\n}\n", "homepage": "", "license": "", "features": {"Speaker": {"dtype": "string", "id": null, "_type": "Value"}, "Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_Act": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "File_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 11, "names": ["Afirmacion", "Apertura", "Cierre", "Confirmacion", "Espera", "Indefinida", "Negacion", "No_entendido", "Nueva_consulta", "Pregunta", "Respuesta"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "miam", "config_name": "dihana", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 1946735, "num_examples": 19063, "dataset_name": "miam"}, "validation": {"name": "validation", "num_bytes": 216498, "num_examples": 2123, "dataset_name": "miam"}, "test": {"name": "test", "num_bytes": 238446, "num_examples": 2361, "dataset_name": "miam"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/MIAM/main/dihana/train.csv": {"num_bytes": 1441183, "checksum": "4702276f3494926fa1c751492b5385530d49ab4b8d4e583f89d5c1ecc9b69311"}, "https://raw.githubusercontent.com/eusip/MIAM/main/dihana/dev.csv": {"num_bytes": 160244, "checksum": "fa267399dbb66f8b096134a2a3f51d71dc48bfa5332dbc0c7b96b2eb0bd91097"}, "https://raw.githubusercontent.com/eusip/MIAM/main/dihana/test.csv": {"num_bytes": 175840, "checksum": "27e45d5f8f0655ed310589777fa9f9eda6a0727dbae277586c086e937c9aca28"}}, "download_size": 1777267, "post_processing_size": null, "dataset_size": 2401679, "size_in_bytes": 4178946}, "ilisten": {"description": "Multilingual dIalogAct benchMark is a collection of resources for training, evaluating, and\nanalyzing natural language understanding systems specifically designed for spoken language. Datasets\nare in English, French, German, Italian and Spanish. They cover a variety of domains including\nspontaneous speech, scripted scenarios, and joint task completion. Some datasets additionally include\nemotion and/or sentimant labels.\n", "citation": "@article{basile2018overview,\ntitle={Overview of the Evalita 2018itaLIan Speech acT labEliNg (iLISTEN) Task},\nauthor={Basile, Pierpaolo and Novielli, Nicole},\njournal={EVALITA Evaluation of NLP and Speech Tools for Italian},\nvolume={12},\npages={44},\nyear={2018}\n}\n@inproceedings{anonymous,\n    title = \"Cross-Lingual Pretraining Methods for Spoken Dialog\",\n    author = \"Anonymous\",\n    booktitle = \"Transactions of the Association for Computational Linguistics\",\n    month = ,\n    year = \"\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"\",\n    doi = \"\",\n    pages = \"\",\n    abstract = \"There has been an increasing interest among NLP researchers towards learning generic\n    representations. However, in the field of multilingual spoken dialogue systems, this problem\n    remains overlooked. Indeed most of the pre-training methods focus on learning representations\n    for written and non-conversational data or are restricted to the monolingual setting. In this\n    work we (1) generalise existing losses to the multilingual setting, (2) develop a new set of\n    losses to leverage parallel conversations when available. These losses improve the learning of\n    representations by fostering the deep encoder to better learn contextual dependencies. The\n    pre-training relies on OpenSubtitles, a huge multilingual corpus that is composed of 24.3G tokens;\n    a by-product of the pre-processing includes multilingual aligned conversations. We also introduce\n    two new multilingual tasks and a new benchmark on multilingual dialogue act labels called MIAM.\n    We validate our pre-training on the three aforementioned tasks and show that our model using our\n    newly designed losses achieves better performances than existing models. Our implementation will\n    be available on github.com and pre-processed data will be available in Datasets (Wolf et al., 2020).\",\n}\n", "homepage": "", "license": "", "features": {"Speaker": {"dtype": "string", "id": null, "_type": "Value"}, "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": 15, "names": ["AGREE", "ANSWER", "CLOSING", "ENCOURAGE-SORRY", "GENERIC-ANSWER", "INFO-REQUEST", "KIND-ATTITUDE_SMALL-TALK", "OFFER-GIVE-INFO", "OPENING", "PERSUASION-SUGGEST", "QUESTION", "REJECT", "SOLICITATION-REQ_CLARIFICATION", "STATEMENT", "TALK-ABOUT-SELF"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "miam", "config_name": "ilisten", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 244336, "num_examples": 1986, "dataset_name": "miam"}, "validation": {"name": "validation", "num_bytes": 33988, "num_examples": 230, "dataset_name": "miam"}, "test": {"name": "test", "num_bytes": 145376, "num_examples": 971, "dataset_name": "miam"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/MIAM/main/ilisten/train.csv": {"num_bytes": 198219, "checksum": "75ea972e2ac1cac8b2e77f55574b18d218e355eb4585b9b95d6873ebd9bdcd04"}, "https://raw.githubusercontent.com/eusip/MIAM/main/ilisten/dev.csv": {"num_bytes": 28741, "checksum": "745faed68a4471a81ad65300e5476fba94d4740ec113df529a2bcd5dd2439971"}, "https://raw.githubusercontent.com/eusip/MIAM/main/ilisten/test.csv": {"num_bytes": 123033, "checksum": "6d2abc758426747b1d271766783d4126756c7372fd92ad3baa0bda67d1de0c77"}}, "download_size": 349993, "post_processing_size": null, "dataset_size": 423700, "size_in_bytes": 773693}, "loria": {"description": "Multilingual dIalogAct benchMark is a collection of resources for training, evaluating, and\nanalyzing natural language understanding systems specifically designed for spoken language. Datasets\nare in English, French, German, Italian and Spanish. They cover a variety of domains including\nspontaneous speech, scripted scenarios, and joint task completion. Some datasets additionally include\nemotion and/or sentimant labels.\n", "citation": "@inproceedings{barahona2012building,\ntitle={Building and exploiting a corpus of dialog interactions between french speaking virtual and human agents},\nauthor={Barahona, Lina Maria Rojas and Lorenzo, Alejandra and Gardent, Claire},\nbooktitle={The eighth international conference on Language Resources and Evaluation (LREC)},\npages={1428--1435},\nyear={2012}\n}\n@inproceedings{anonymous,\n    title = \"Cross-Lingual Pretraining Methods for Spoken Dialog\",\n    author = \"Anonymous\",\n    booktitle = \"Transactions of the Association for Computational Linguistics\",\n    month = ,\n    year = \"\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"\",\n    doi = \"\",\n    pages = \"\",\n    abstract = \"There has been an increasing interest among NLP researchers towards learning generic\n    representations. However, in the field of multilingual spoken dialogue systems, this problem\n    remains overlooked. Indeed most of the pre-training methods focus on learning representations\n    for written and non-conversational data or are restricted to the monolingual setting. In this\n    work we (1) generalise existing losses to the multilingual setting, (2) develop a new set of\n    losses to leverage parallel conversations when available. These losses improve the learning of\n    representations by fostering the deep encoder to better learn contextual dependencies. The\n    pre-training relies on OpenSubtitles, a huge multilingual corpus that is composed of 24.3G tokens;\n    a by-product of the pre-processing includes multilingual aligned conversations. We also introduce\n    two new multilingual tasks and a new benchmark on multilingual dialogue act labels called MIAM.\n    We validate our pre-training on the three aforementioned tasks and show that our model using our\n    newly designed losses achieves better performances than existing models. Our implementation will\n    be available on github.com and pre-processed data will be available in Datasets (Wolf et al., 2020).\",\n}\n", "homepage": "", "license": "", "features": {"Speaker": {"dtype": "string", "id": null, "_type": "Value"}, "Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_Act": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "File_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 31, "names": ["ack", "ask", "find_mold", "find_plans", "first_step", "greet", "help", "inform", "inform_engine", "inform_job", "inform_material_space", "informer_conditioner", "informer_decoration", "informer_elcomps", "informer_end_manufacturing", "kindAtt", "manufacturing_reqs", "next_step", "no", "other", "quality_control", "quit", "reqRep", "security_policies", "staff_enterprise", "staff_job", "studies_enterprise", "studies_job", "todo_failure", "todo_irreparable", "yes"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "miam", "config_name": "loria", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 1208730, "num_examples": 8465, "dataset_name": "miam"}, "validation": {"name": "validation", "num_bytes": 133829, "num_examples": 942, "dataset_name": "miam"}, "test": {"name": "test", "num_bytes": 149855, "num_examples": 1047, "dataset_name": "miam"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/MIAM/main/loria/train.csv": {"num_bytes": 989066, "checksum": "0cda3440cdf0157f1a70617374842f04a3b7cf8c5175cca9f3fe9a33a5105ddf"}, "https://raw.githubusercontent.com/eusip/MIAM/main/loria/dev.csv": {"num_bytes": 109364, "checksum": "02f99287f69dc869926aaf077d6c3d0a81cb2576c80721c9272e39a2b226d989"}, "https://raw.githubusercontent.com/eusip/MIAM/main/loria/test.csv": {"num_bytes": 122702, "checksum": "8e2f5e513761970aad138332fb83488ca52c943492153c08199ddc4ae8fe4209"}}, "download_size": 1221132, "post_processing_size": null, "dataset_size": 1492414, "size_in_bytes": 2713546}, "maptask": {"description": "Multilingual dIalogAct benchMark is a collection of resources for training, evaluating, and\nanalyzing natural language understanding systems specifically designed for spoken language. Datasets\nare in English, French, German, Italian and Spanish. They cover a variety of domains including\nspontaneous speech, scripted scenarios, and joint task completion. Some datasets additionally include\nemotion and/or sentimant 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{anonymous,\n    title = \"Cross-Lingual Pretraining Methods for Spoken Dialog\",\n    author = \"Anonymous\",\n    booktitle = \"Transactions of the Association for Computational Linguistics\",\n    month = ,\n    year = \"\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"\",\n    doi = \"\",\n    pages = \"\",\n    abstract = \"There has been an increasing interest among NLP researchers towards learning generic\n    representations. However, in the field of multilingual spoken dialogue systems, this problem\n    remains overlooked. Indeed most of the pre-training methods focus on learning representations\n    for written and non-conversational data or are restricted to the monolingual setting. In this\n    work we (1) generalise existing losses to the multilingual setting, (2) develop a new set of\n    losses to leverage parallel conversations when available. These losses improve the learning of\n    representations by fostering the deep encoder to better learn contextual dependencies. The\n    pre-training relies on OpenSubtitles, a huge multilingual corpus that is composed of 24.3G tokens;\n    a by-product of the pre-processing includes multilingual aligned conversations. We also introduce\n    two new multilingual tasks and a new benchmark on multilingual dialogue act labels called MIAM.\n    We validate our pre-training on the three aforementioned tasks and show that our model using our\n    newly designed losses achieves better performances than existing models. Our implementation will\n    be available on github.com and pre-processed data will be available in Datasets (Wolf et al., 2020).\",\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"}, "Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "File_ID": {"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": "miam", "config_name": "maptask", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 1910120, "num_examples": 25382, "dataset_name": "miam"}, "validation": {"name": "validation", "num_bytes": 389879, "num_examples": 5221, "dataset_name": "miam"}, "test": {"name": "test", "num_bytes": 396947, "num_examples": 5335, "dataset_name": "miam"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/MIAM/main/maptask/train.csv": {"num_bytes": 1226569, "checksum": "76ce790d9c5100d2f2b1f535edd8e8b5a6d88d6bbaa7f3a948dd95bff1e2e798"}, "https://raw.githubusercontent.com/eusip/MIAM/main/maptask/dev.csv": {"num_bytes": 249215, "checksum": "895ca2761e8224b02f963df6836912c0a913362f5d44dfc4813391b51919f147"}, "https://raw.githubusercontent.com/eusip/MIAM/main/maptask/test.csv": {"num_bytes": 253237, "checksum": "e11a1fbaa4ffc74c0b438b2d3e6f17f991adee50e368c2c9c57f98ef6a2dd0c3"}}, "download_size": 1729021, "post_processing_size": null, "dataset_size": 2696946, "size_in_bytes": 4425967}, "vm2": {"description": "Multilingual dIalogAct benchMark is a collection of resources for training, evaluating, and\nanalyzing natural language understanding systems specifically designed for spoken language. Datasets\nare in English, French, German, Italian and Spanish. They cover a variety of domains including\nspontaneous speech, scripted scenarios, and joint task completion. Some datasets additionally include\nemotion and/or sentimant labels.\n", "citation": "@book{kay1992verbmobil,\ntitle={Verbmobil: A translation system for face-to-face dialog},\nauthor={Kay, Martin},\nyear={1992},\npublisher={University of Chicago Press}\n}\n@inproceedings{anonymous,\n    title = \"Cross-Lingual Pretraining Methods for Spoken Dialog\",\n    author = \"Anonymous\",\n    booktitle = \"Transactions of the Association for Computational Linguistics\",\n    month = ,\n    year = \"\",\n    address = \"Online\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"\",\n    doi = \"\",\n    pages = \"\",\n    abstract = \"There has been an increasing interest among NLP researchers towards learning generic\n    representations. However, in the field of multilingual spoken dialogue systems, this problem\n    remains overlooked. Indeed most of the pre-training methods focus on learning representations\n    for written and non-conversational data or are restricted to the monolingual setting. In this\n    work we (1) generalise existing losses to the multilingual setting, (2) develop a new set of\n    losses to leverage parallel conversations when available. These losses improve the learning of\n    representations by fostering the deep encoder to better learn contextual dependencies. The\n    pre-training relies on OpenSubtitles, a huge multilingual corpus that is composed of 24.3G tokens;\n    a by-product of the pre-processing includes multilingual aligned conversations. We also introduce\n    two new multilingual tasks and a new benchmark on multilingual dialogue act labels called MIAM.\n    We validate our pre-training on the three aforementioned tasks and show that our model using our\n    newly designed losses achieves better performances than existing models. Our implementation will\n    be available on github.com and pre-processed data will be available in Datasets (Wolf et al., 2020).\",\n}\n", "homepage": "", "license": "", "features": {"Utterance": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_Act": {"dtype": "string", "id": null, "_type": "Value"}, "Speaker": {"dtype": "string", "id": null, "_type": "Value"}, "Dialogue_ID": {"dtype": "string", "id": null, "_type": "Value"}, "Label": {"num_classes": 31, "names": ["ACCEPT", "BACKCHANNEL", "BYE", "CLARIFY", "CLOSE", "COMMIT", "CONFIRM", "DEFER", "DELIBERATE", "DEVIATE_SCENARIO", "EXCLUDE", "EXPLAINED_REJECT", "FEEDBACK", "FEEDBACK_NEGATIVE", "FEEDBACK_POSITIVE", "GIVE_REASON", "GREET", "INFORM", "INIT", "INTRODUCE", "NOT_CLASSIFIABLE", "OFFER", "POLITENESS_FORMULA", "REJECT", "REQUEST", "REQUEST_CLARIFY", "REQUEST_COMMENT", "REQUEST_COMMIT", "REQUEST_SUGGEST", "SUGGEST", "THANK"], "names_file": null, "id": null, "_type": "ClassLabel"}, "Idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "miam", "config_name": "vm2", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 1869254, "num_examples": 25060, "dataset_name": "miam"}, "validation": {"name": "validation", "num_bytes": 209390, "num_examples": 2860, "dataset_name": "miam"}, "test": {"name": "test", "num_bytes": 209032, "num_examples": 2855, "dataset_name": "miam"}}, "download_checksums": {"https://raw.githubusercontent.com/eusip/MIAM/main/vm2/train.csv": {"num_bytes": 1342990, "checksum": "5b5ddbd333a57f033a01e9b4135c7675d56a3a819258f85cee40e6adef53f7f7"}, "https://raw.githubusercontent.com/eusip/MIAM/main/vm2/dev.csv": {"num_bytes": 149358, "checksum": "30c5ba7219d113cdb4260488b0f7e515347f8c06ca5e0345b701ccfc229c41be"}, "https://raw.githubusercontent.com/eusip/MIAM/main/vm2/test.csv": {"num_bytes": 149105, "checksum": "a94db1a8ece862084624ffb0774b9b22d180a30a778a19613b963e4c244c9da3"}}, "download_size": 1641453, "post_processing_size": null, "dataset_size": 2287676, "size_in_bytes": 3929129}}

dummy/dihana/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8809b1becaafced1402ccedf528ab36a8c4bae3de01cad9b992250a418590069
+size 1101

dummy/ilisten/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:95ed08efe28bd23b4af687b0af96c48b817fe93d0928938fb39fd8752bbe875e
+size 1367

dummy/loria/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd15bcbe33011fc79d8ac8434978e3375069a9e3b28b3a3cdf8343d2741d5eaf
+size 1266

dummy/maptask/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1bd0b4d17d0e03ca14e1b4e6fd95fa7c951c47e46663ecaf17dcf35870f4321d
+size 900

dummy/vm2/1.0.0/dummy_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:902af73b121bb8c2896412035d4e8d2e2669fbac3d060b40ff76cf12688beff8
+size 1114

miam.py

@@ -0,0 +1,436 @@
+# 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 Multilingual dIalogAct benchMark."""
+
+
+import textwrap
+
+import pandas as pd
+
+import datasets
+
+
+_MIAM_CITATION = """\
+@unpublished{
+anonymous2021cross-lingual,
+title={Cross-Lingual Pretraining Methods for Spoken Dialog},
+author={Anonymous},
+journal={OpenReview Preprint},
+year={2021},
+url{https://openreview.net/forum?id=c1oDhu_hagR},
+note={anonymous preprint under review}
+}
+"""
+
+_MIAM_DESCRIPTION = """\
+Multilingual dIalogAct benchMark is a collection of resources for training, evaluating, and
+analyzing natural language understanding systems specifically designed for spoken language. Datasets
+are in English, French, German, Italian and Spanish. They cover a variety of domains including
+spontaneous speech, scripted scenarios, and joint task completion. Some datasets additionally include
+emotion and/or sentimant labels.
+"""
+
+_URL = "https://raw.githubusercontent.com/eusip/MIAM/main"
+
+DIHANA_DA_DESCRIPTION = {
+    "Afirmacion": "Feedback_positive",
+    "Apertura": "Opening",
+    "Cierre": "Closing",
+    "Confirmacion": "Acknowledge",
+    "Espera": "Hold",
+    "Indefinida": "Undefined",
+    "Negacion": "Feedback_negative",
+    "No_entendido": "Request_clarify",
+    "Nueva_consulta": "New_request",
+    "Pregunta": "Request",
+    "Respuesta": "Reply",
+}
+
+
+class MiamConfig(datasets.BuilderConfig):
+    """BuilderConfig for MIAM."""
+
+    def __init__(
+        self,
+        text_features,
+        label_column,
+        data_url,
+        citation,
+        url,
+        label_classes=None,
+        **kwargs,
+    ):
+        """BuilderConfig for MIAM.
+        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(MiamConfig, 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 Miam(datasets.GeneratorBasedBuilder):
+    """The Multilingual dIalogAct benchMark."""
+
+    BUILDER_CONFIGS = [
+        MiamConfig(
+            name="dihana",
+            description=textwrap.dedent(
+                """\
+            The Dihana corpus primarily consists of spontaneous speech. The corpus is annotated
+            using three different levels of labels. The first level is dedicated to the generic
+            task-independent DA and the two additional are made with task-specific information. We
+            focus on the 11 first level tags."""
+            ),
+            text_features={
+                "Speaker": "Speaker",
+                "Utterance": "Utterance",
+                "Dialogue_Act": "Dialogue_Act",
+                "Dialogue_ID": "Dialogue_ID",
+                "File_ID": "File_ID",
+            },
+            label_classes=list(DIHANA_DA_DESCRIPTION.keys()),
+            label_column="Dialogue_Act",
+            data_url={
+                "train": _URL + "/dihana/train.csv",
+                "dev": _URL + "/dihana/dev.csv",
+                "test": _URL + "/dihana/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @inproceedings{benedi2006design,
+            title={Design and acquisition of a telephone spontaneous speech dialogue corpus in Spanish: DIHANA},
+            author={Bened{\'i}, Jos{\'e}-Miguel and Lleida, Eduardo and Varona, Amparo and Castro, Mar{\'i}a-Jos{\'e} and Galiano, Isabel and Justo, Raquel and L{\'o}pez, I and Miguel, Antonio},
+            booktitle={Fifth International Conference on Language Resources and Evaluation (LREC)},
+            pages={1636--1639},
+            year={2006}
+            }
+            @inproceedings{post2013improved,
+            title={Improved speech-to-text translation with the Fisher and Callhome Spanish--English speech translation corpus},
+            author={Post, Matt and Kumar, Gaurav and Lopez, Adam and Karakos, Damianos and Callison-Burch, Chris and Khudanpur, Sanjeev},
+            booktitle={Proc. IWSLT},
+            year={2013}
+            }
+            @article{coria2005predicting,
+            title={Predicting obligation dialogue acts from prosodic and speaker infomation},
+            author={Coria, S and Pineda, L},
+            journal={Research on Computing Science (ISSN 1665-9899), Centro de Investigacion en Computacion, Instituto Politecnico Nacional, Mexico City},
+            year={2005}
+            }"""
+            ),
+            url="",
+        ),
+        MiamConfig(
+            name="ilisten",
+            description=textwrap.dedent(
+                """\
+            "itaLIan Speech acT labEliNg" (iLISTEN) is a corpus of annotated dialogue turns labeled
+            for speech acts."""
+            ),
+            text_features={
+                "Speaker": "Speaker",
+                "Utterance": "Utterance",
+                "Dialogue_Act": "Dialogue_Act",
+                "Dialogue_ID": "Dialogue_ID",
+            },
+            label_classes=[
+                "AGREE",
+                "ANSWER",
+                "CLOSING",
+                "ENCOURAGE-SORRY",
+                "GENERIC-ANSWER",
+                "INFO-REQUEST",
+                "KIND-ATTITUDE_SMALL-TALK",
+                "OFFER-GIVE-INFO",
+                "OPENING",
+                "PERSUASION-SUGGEST",
+                "QUESTION",
+                "REJECT",
+                "SOLICITATION-REQ_CLARIFICATION",
+                "STATEMENT",
+                "TALK-ABOUT-SELF",
+            ],
+            label_column="Dialogue_Act",
+            data_url={
+                "train": _URL + "/ilisten/train.csv",
+                "dev": _URL + "/ilisten/dev.csv",
+                "test": _URL + "/ilisten/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @article{basile2018overview,
+            title={Overview of the Evalita 2018itaLIan Speech acT labEliNg (iLISTEN) Task},
+            author={Basile, Pierpaolo and Novielli, Nicole},
+            journal={EVALITA Evaluation of NLP and Speech Tools for Italian},
+            volume={12},
+            pages={44},
+            year={2018}
+            }"""
+            ),
+            url="",
+        ),
+        MiamConfig(
+            name="loria",
+            description=textwrap.dedent(
+                """\
+            The LORIA Nancy dialog corpus is derived from human-machine interactions in a serious
+            game setting."""
+            ),
+            text_features={
+                "Speaker": "Speaker",
+                "Utterance": "Utterance",
+                "Dialogue_Act": "Dialogue_Act",
+                "Dialogue_ID": "Dialogue_ID",
+                "File_ID": "File_ID",
+            },
+            label_classes=[
+                "ack",
+                "ask",
+                "find_mold",
+                "find_plans",
+                "first_step",
+                "greet",
+                "help",
+                "inform",
+                "inform_engine",
+                "inform_job",
+                "inform_material_space",
+                "informer_conditioner",
+                "informer_decoration",
+                "informer_elcomps",
+                "informer_end_manufacturing",
+                "kindAtt",
+                "manufacturing_reqs",
+                "next_step",
+                "no",
+                "other",
+                "quality_control",
+                "quit",
+                "reqRep",
+                "security_policies",
+                "staff_enterprise",
+                "staff_job",
+                "studies_enterprise",
+                "studies_job",
+                "todo_failure",
+                "todo_irreparable",
+                "yes",
+            ],
+            label_column="Dialogue_Act",
+            data_url={
+                "train": _URL + "/loria/train.csv",
+                "dev": _URL + "/loria/dev.csv",
+                "test": _URL + "/loria/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @inproceedings{barahona2012building,
+            title={Building and exploiting a corpus of dialog interactions between french speaking virtual and human agents},
+            author={Barahona, Lina Maria Rojas and Lorenzo, Alejandra and Gardent, Claire},
+            booktitle={The eighth international conference on Language Resources and Evaluation (LREC)},
+            pages={1428--1435},
+            year={2012}
+            }"""
+            ),
+            url="",
+        ),
+        MiamConfig(
+            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",
+                "Dialogue_ID": "Dialogue_ID",
+                "File_ID": "File_ID",
+            },
+            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.csv",
+                "dev": _URL + "/maptask/dev.csv",
+                "test": _URL + "/maptask/test.csv",
+            },
+            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/",
+        ),
+        MiamConfig(
+            name="vm2",
+            description=textwrap.dedent(
+                """\
+            The VERBMOBIL 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": "Utterance",
+                "Dialogue_Act": "Dialogue_Act",
+                "Speaker": "Speaker",
+                "Dialogue_ID": "Dialogue_ID",
+            },
+            label_classes=[
+                "ACCEPT",
+                "BACKCHANNEL",
+                "BYE",
+                "CLARIFY",
+                "CLOSE",
+                "COMMIT",
+                "CONFIRM",
+                "DEFER",
+                "DELIBERATE",
+                "DEVIATE_SCENARIO",
+                "EXCLUDE",
+                "EXPLAINED_REJECT",
+                "FEEDBACK",
+                "FEEDBACK_NEGATIVE",
+                "FEEDBACK_POSITIVE",
+                "GIVE_REASON",
+                "GREET",
+                "INFORM",
+                "INIT",
+                "INTRODUCE",
+                "NOT_CLASSIFIABLE",
+                "OFFER",
+                "POLITENESS_FORMULA",
+                "REJECT",
+                "REQUEST",
+                "REQUEST_CLARIFY",
+                "REQUEST_COMMENT",
+                "REQUEST_COMMIT",
+                "REQUEST_SUGGEST",
+                "SUGGEST",
+                "THANK",
+            ],
+            label_column="Dialogue_Act",
+            data_url={
+                "train": _URL + "/vm2/train.csv",
+                "dev": _URL + "/vm2/dev.csv",
+                "test": _URL + "/vm2/test.csv",
+            },
+            citation=textwrap.dedent(
+                """\
+            @book{kay1992verbmobil,
+            title={Verbmobil: A translation system for face-to-face dialog},
+            author={Kay, Martin},
+            year={1992},
+            publisher={University of Chicago Press}
+            }"""
+            ),
+            url="",
+        ),
+    ]
+
+    def _info(self):
+        features = {text_feature: datasets.Value("string") for text_feature in self.config.text_features.keys()}
+        if self.config.label_classes:
+            features["Label"] = datasets.features.ClassLabel(names=self.config.label_classes)
+        features["Idx"] = datasets.Value("int32")
+        return datasets.DatasetInfo(
+            description=_MIAM_DESCRIPTION,
+            features=datasets.Features(features),
+            homepage=self.config.url,
+            citation=self.config.citation + "\n" + _MIAM_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):
+        df = pd.read_csv(data_file, delimiter=",", header=0, quotechar='"', dtype=str)[
+            self.config.text_features.keys()
+        ]
+
+        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