Update README.md

README.md

@@ -8,6 +8,7 @@ pipeline_tag: text-classification
 
 ## BETO
 
+*WARNING*: Don't use it on the side textbox! Try our [demo]((https://huggingface.co/spaces/piubamas/discurso-de-odio) instead.
 Model trained to detect hate speech comments in news articles. Base model is BETO, a Spanish BERT pre-trained model. The task the model was trained on is a multilabel classification problem, where each input have a label for each of the considered groups:
 
 | Label | Description |
@@ -52,9 +53,7 @@ Hay que matarlos a todos!!! Nos infectaron con su virus! [SEP] China prohibió l
 
 Sadly, the `huggingface` pipeline does not support multi-label classification, so this model cannot be tested directly in the side widget.
 
-To use it, you can try our [Demo](https://huggingface.co/spaces/piubamas/discurso-de-odio).
-
-If you want to use it with your own code, you can use the following snippet:
+To use it, you can try our [demo](https://huggingface.co/spaces/piubamas/discurso-de-odio). If you want to use it with your own code, use the following snippet:
 
 ```python