Update README.md

README.md

@@ -16,6 +16,9 @@ co2_eq_emissions:
 
 - Problem type: Binary Classification
 - Model ID: 2489276793
+- Base model: BertForSequenceClassification
+- Parameters: 109M
+- Model size: 416MB
 - CO2 Emissions (in grams): 4.4298
 
 ## Validation Metrics
@@ -29,6 +32,9 @@ co2_eq_emissions:
 
 ## Usage
 
+This model was trained on a random subset of the [told_br](https://huggingface.co/datasets/told-br) dataset (1/3 of the original size). Our main objective is to provide a small
+model that can be used to classify Brazilian Portuguese tweets in a binary way ('toxic' or 'non toxic').
+
 You can use cURL to access this model:
 
 ```
@@ -40,9 +46,9 @@ Or Python API:
 ```
 from transformers import AutoModelForSequenceClassification, AutoTokenizer
 
-model = AutoModelForSequenceClassification.from_pretrained("alexandreteles/autotrain-told_br_binary_sm-2489276793", use_auth_token=True)
+model = AutoModelForSequenceClassification.from_pretrained("alexandreteles/told_br_binary_sm", use_auth_token=True)
 
-tokenizer = AutoTokenizer.from_pretrained("alexandreteles/autotrain-told_br_binary_sm-2489276793", use_auth_token=True)
+tokenizer = AutoTokenizer.from_pretrained("alexandreteles/told_br_binary_sm", use_auth_token=True)
 
 inputs = tokenizer("I love AutoTrain", return_tensors="pt")