readme: add initial version

README.md

@@ -1,3 +1,22 @@
 ---
 license: cc-by-4.0
+base_model: benjamin/wtp-canine-s-12l
 ---
+
+# Where's the Point? A `wtpsplit` model for Historical German
+
+This repository hosts a model that was trained using the awesome [`wtpsplit`](https://github.com/bminixhofer/wtpsplit) library.
+
+The model was trained on OCR'ed German books from 1819 to 1944 extracted from the [NCC](https://huggingface.co/datasets/NbAiLab/NCC) corpus.
+
+Please note: the model is constantly updated and work in progress.
+
+# Usage
+
+The following snippet shows how to use this model:
+
+```python
+from wtpsplit import WtP
+
+model.split("Das ist einer schöner Tag... Wie geht es dir?", lang_code="de", style="digibok")
+```