Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
This dataset is made up of data taken from SQuAD v2.0, RACE, CoQA, and MSMARCO. Some examples have been filtered out of the original datasets and others have been modified.
|
2 |
+
|
3 |
+
There are two fields; question and text. The question field contains the question, and the text field contains both the answer and the context in the following format:
|
4 |
+
"<answer> (answer text) <context> (context text)"
|
5 |
+
|
6 |
+
The <answer> and <context> are included as special tokens in the question generator's tokenizer.
|
7 |
+
|
8 |
+
This dataset is intended to be used with the [question_generator repo](https://github.com/AMontgomerie/question_generator) to train the question generator model.
|