Update README.md

README.md

@@ -6,13 +6,15 @@ colorTo: green
 sdk: static
 pinned: false
 ---
+This repo contains the models and datasets used in DReAMy, an open-source toolkit to automatically analyse (for now only) textual dream reports. 
+Annotations are currently based exclusively on the [Hall & Van de Castle](https://link.springer.com/chapter/10.1007/978-1-4899-0298-6_2) framework, 
+and are divided into two main tasks: sentiment analysis, with both a multi-label classification and/or generation format, and character identification (i.e., NER), 
+For more details on the training and evaluation procedure of the models see [Bertolini et al., 23](https://arxiv.org/abs/2302.14828). 
 
+DReAMy can be easily installed via `pip install dreamy` – we do recommend using a virtual env running python 3.9. 
+Please refer to the git-repo for code, examples and tutorials. If you wish to test/query (a set of) DReAMy's model, 
+try out a Space below. 
 
-DReAMy is an open-source toolkit to automatically analyse (for now only) textual dream reports. At the moment, annotations are exclusively based on the [Hall & Van de Castle](https://link.springer.com/chapter/10.1007/978-1-4899-0298-6_2) emotions framework, but we are looking forward to expanding our set applications. As for the theoretical backbone, DReAMy and its model are based on a fruitful collaboration between NLP and sleep research. For more details, please refer to the [Bertolini et al., 23](https://arxiv.org/abs/2302.14828) pre-print.
-
-Here you can find DReAMy's models, tuned to analyse dream reports, as well as two datasets containing (non-annotated) dream reports from [DreamBank](https://www.dreambank.net/). Specialised models focus on two main tasks: sentiment analysis and character identification (i.e., NER), with both a multi-label classification and/or generation format.
-
-DReAMy can be easily installed via `pip install dreamy` – we do recommend using a virtual env running python 3.9. Please refer  to the git-repo for code, examples and tutorials. If you wish to test/query (a set of) DReAMy's model, try out a Space below. 
 
 Lastly, if you use a model in your work, please cite the pre-print.
 ```bibtex