Update README.md

README.md

@@ -1,2 +1,27 @@
+---
+datasets:
+- jigsaw_unintended_bias
+language:
+- en
+pipeline_tag: fill-mask
+tags:
+- code
+- bert
+- non-toxic
+- nontoxicbert
+---
+
 Bert-Based-Uncased finetuned with the non-toxic only from the Jigsaw-unintended-bias dataset, to make the BERT 
-prediction less toxic.
+prediction less toxic.
+
+I am still working to make this model even less toxic.
+
+
+
+You can use it directly using the transformers:
+
+from transformers import AutoTokenizer, AutoModelForMaskedLM
+
+tokenizer = AutoTokenizer.from_pretrained("Ashokajou51/NonToxicCivilBert")
+
+model = AutoModelForMaskedLM.from_pretrained("Ashokajou51/NonToxicCivilBert")