Readme

README.md

@@ -9,13 +9,14 @@ Having the same architecture of [BERTweet](https://huggingface.co/docs/transform
 
 ## Usage
 
+### Normalized Inputs
+
 ```python
 import torch
 from transformers import AutoModel, AutoTokenizer 
 
 model = AutoModel.from_pretrained('melll-uff/bertweetbr')
-
-tokenizer = AutoTokenizer.from_pretrained('melll-uff/bertweetbr')
+tokenizer = AutoTokenizer.from_pretrained('melll-uff/bertweetbr', normalization=False)
 
 # INPUT TWEET IS ALREADY NORMALIZED!
 line = "Tem vídeo novo no canal do @USER :rosto_sorridente_com_olhos_de_coração: Passem por lá e confiram : HTTPURL"
@@ -24,4 +25,22 @@ input_ids = tokenizer(line, return_tensors="pt")
 
 with torch.no_grad():
     features = model(**input_ids)  # Models outputs are now tuples
+```
+
+ ### Normalize raw input Tweets
+
+ ```python
+import torch
+from transformers import AutoModel, AutoTokenizer 
+
+
+ ```python
+from transformers import pipeline
+
+model_name = 'melll-uff/bertweetbr'
+tokenizer = AutoTokenizer.from_pretrained('melll-uff/bertweetbr', normalization=False)
+
+filler_mask = pipeline("fill-mask", model=model_name, tokenizer=tokenizer)
+
+filler_mask("Rio é a <mask> cidade do Brasil.", top_k=5)
 ```