update

README.md

@@ -0,0 +1,56 @@
+---
+languages:
+- en
+multilinguality:
+- monolingual
+size_categories:
+- 100K<n<1M
+task_categories:
+- conditional-text-generation
+task_ids:
+- summarization
+---
+
+# MediaSum dataset for summarization
+
+Summarization dataset copied from [MediaSum: A Large-scale Media Interview Dataset for Dialogue Summarization](https://github.com/zcgzcgzcg1/MediaSumR)
+
+This dataset is compatible with the [`run_summarization.py`](https://github.com/huggingface/transformers/tree/master/examples/pytorch/summarization) script from Transformers if you add this line to the `summarization_name_mapping` variable:
+```python
+"ccdv/mediasum": ("document", "summary")
+```
+
+# Configs
+4 possibles configs:
+- `roberta` will concatenate documents with "</s>" (default)
+- `newline` will concatenate documents with "\n"
+- `bert` will concatenate documents with "[SEP]"
+- `list` will return the list of documents instead of a single string
+
+### Data Fields
+
+- `id`: paper id
+- `document`: a string/list containing the body of a set of documents
+- `summary`: a string containing the abstract of the set
+
+### Data Splits
+
+This dataset has 3 splits: _train_, _validation_, and _test_. \
+Token counts are white space based.
+
+| Dataset Split | Number of Instances | 
+| ------------- | --------------------|
+| Train         | 443596              |
+| Validation    | 10000               |
+| Test          | 10000               |
+
+
+# Cite original article
+```
+@article{zhu2021mediasum,
+  title={MediaSum: A Large-scale Media Interview Dataset for Dialogue Summarization},
+  author={Zhu, Chenguang and Liu, Yang and Mei, Jie and Zeng, Michael},
+  journal={arXiv preprint arXiv:2103.06410},
+  year={2021}
+}
+```

mediasum.py

@@ -21,7 +21,7 @@ _CITATION = """\
 }
 """
 _ABSTRACT = "summary"
-_ARTICLE = "utt"
+_ARTICLE = "document"
 
 class MediaSumSummarizationConfig(datasets.BuilderConfig):
     """BuilderConfig for MediaSumSummarization."""
@@ -117,7 +117,9 @@ class MediaSumSummarizationDataset(datasets.GeneratorBasedBuilder):
                 'summary': str,
                 'document': List[str],
                 """
+                
                 document = data["utt"]
+                
                 if self.config.name != "list":
                     document = join_.join(document)
                 summary = data["summary"]

test_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:760b1efde09738c70fdb0449dbb9a4fa07f1d474cb790ddf9446650052df89d7
+size 33913771

train_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d29532f5f9d84b33a9007ebad450898540742edff55d5d058fa47e4c975a3b7
+size 1442774847

val_data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f466af07260fd2fa72aff5b065ac355e9bf677931b3e84c99764e6ebf2c9215
+size 34058060