typo

app.py

@@ -28,7 +28,7 @@ with st.sidebar:
     st.header("How REBEL works")
     st.markdown("REBEL is a **text2text** model obtained by fine-tuning [**BART**](https://huggingface.co/docs/transformers/model_doc/bart) for translating a raw input sentence containing entities and implicit relations into a set of triplets that explicitly refer to those relations. You can find [REBEL in the Hugging Face Hub](https://huggingface.co/Babelscape/rebel-large).")
     st.header("Further steps")
-    st.markdown("Even though they are not visualized, the knowledge graph saves information about the provenience of each relation (e.g. from which articles it has been exrtacted and other metadata), along with Wikipedia data about each entity.")
+    st.markdown("Even though they are not visualized, the knowledge graph saves information about the provenience of each relation (e.g. from which articles it has been extracted and other metadata), along with Wikipedia data about each entity.")
     st.markdown("Other libraries used:\n- [wikipedia](https://pypi.org/project/wikipedia/): For validating extracted entities checking if they have a corresponding Wikipedia page.\n- [newspaper](https://github.com/codelucas/newspaper): For parsing articles from URLs.\n- [pyvis](https://pyvis.readthedocs.io/en/latest/index.html): For graphs visualizations.\n- [GoogleNews](https://github.com/Iceloof/GoogleNews): For reading Google News latest articles about a topic.")
     st.header("Considerations")
     st.markdown("If you look closely at the extracted knowledge graphs, some extracted relations are false. Indeed, relation extraction models are still far from perfect and require further steps in the pipeline to build reliable knowledge graphs. Consider this demo as a starting step!")