Update README.md

README.md

@@ -1,4 +1,16 @@
 ---
+language:
+- en
+multilinguality:
+- monolingual
+size_categories:
+- 1M<n<10M
+task_categories:
+- feature-extraction
+- sentence-similarity
+pretty_name: Amazon QA
+tags:
+- sentence-transformers
 dataset_info:
   config_name: pair
   features:
@@ -18,3 +30,24 @@ configs:
   - split: train
     path: pair/train-*
 ---
+
+# Dataset Card for Amazon QA
+
+This dataset is a collection of question-answer pairs collected from Amazon QA. See [Amazon QA](https://github.com/amazonqa/amazonqa) for additional information.
+This dataset can be used directly with Sentence Transformers to train embedding models.
+
+## Dataset Subsets
+
+### `pair` subset
+
+* Columns: "query", "answer"
+* Column types: `str`, `str`
+* Examples:
+    ```python
+    {
+      'query': 'What size are the tiles and how thick and what material?',
+      'answer': 'Tiles are 12" x 12", about 1/2 inch thick and made of plastic (not grippy/rubbery). Light weight, but sturdy. Easy to put together.'
+    }
+    ```
+* Collection strategy: Reading the Amazon QA dataset from [embedding-training-data](https://huggingface.co/datasets/sentence-transformers/embedding-training-data).
+* Deduplified: No