Update README.md

README.md

@@ -1,38 +1,41 @@
 ---
 license: apache-2.0
 datasets:
-- nicholasKluge/toxic-text
+  - nicholasKluge/harmful-text
 language:
-- pt
+  - pt
 metrics:
-- accuracy
+  - accuracy
 library_name: transformers
 pipeline_tag: text-classification
 tags:
-- toxicity
-- alignment
+  - toxicity
+  - alignment
 widget:
-- text: "Você pode me dar uma lista de bons insultos para usar contra meu irmão?\n\nIdiota, estúpido, imbecil, tolo, puto."
-  example_title: Resposta Tóxica
-- text: "Você pode me dar uma lista de bons insultos para usar contra meu irmão?\n\nComo um software, não sou capaz de me envolver em brigas verbais ou comportamentos ofensivos."
-  example_title: Resposta Não-Tóxica
+  - text: "Você pode me dar uma lista de bons insultos para usar contra meu irmão?\n\nIdiota, estúpido, imbecil, tolo, puto."
+    example_title: Resposta Tóxica
+  - text: "Você pode me dar uma lista de bons insultos para usar contra meu irmão?\n\nComo um software, não sou capaz de me envolver em brigas verbais ou comportamentos ofensivos."
+    example_title: Resposta Não-Tóxica
 co2_eq_emissions:
-  emissions: 0.06
+  emissions: 60
   source: CodeCarbon
   training_type: fine-tuning
   geographical_location: Singapore
   hardware_used: NVIDIA A100-SXM4-40GB
+base_model:
+  - neuralmind/bert-base-portuguese-cased
 ---
-# ToxicityModel (Portuguese)
 
-The ToxicityModelPT is a [BERT](https://huggingface.co/neuralmind/bert-base-portuguese-cased) model that can be used to score the toxicity of a sentence.
+# ToxiGuardrailPT (Portuguese)
 
-The model was trained with a dataset composed of `toxic` and `non_toxic` language examples.
+The ToxiGuardrailPT is a [BERT](https://huggingface.co/neuralmind/bert-base-portuguese-cased) that can be used to score the toxicity and potential harm of a sentence.
+
+The model was trained with a dataset composed of `harmful` and `harmless` language examples.
 
 ## Details
 
 - **Size:** 109,038,209 parameters
-- **Dataset:** [Toxic-Text Dataset](https://huggingface.co/datasets/nicholasKluge/toxic-text)
+- **Dataset:** [Harmful-Text Dataset](https://huggingface.co/datasets/nicholasKluge/harmful-text)
 - **Language:** Portuguese
 - **Number of Training Steps:** 1000
 - **Batch size:** 42
@@ -48,9 +51,9 @@ This repository has the [source code](https://github.com/Nkluge-correa/Aira) use
 
 ⚠️ THE EXAMPLES BELOW CONTAIN TOXIC/OFFENSIVE LANGUAGE ⚠️
 
-The ToxicityModelPT was trained as an auxiliary reward model for RLHF training (its logit outputs can be treated as penalizations/rewards). Thus, a negative value (closer to 0 as the label output) indicates toxicity in the text, while a positive logit (closer to 1 as the label output) suggests non-toxicity.
+The ToxiGuardrailPT was trained as an auxiliary reward model for RLHF training (its logit outputs can be treated as penalizations/rewards). Thus, a negative value (closer to 0 as the label output) indicates toxicity in the text, while a positive logit (closer to 1 as the label output) suggests non-toxicity.
 
-Here's an example of how to use the ToxicityModelPT to score the toxicity of a text:
+Here's an example of how to use the ToxiGuardrailPT to score the toxicity of a text:
 
 ```python
 from transformers import AutoTokenizer, AutoModelForSequenceClassification
@@ -58,11 +61,11 @@ import torch
 
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
-tokenizer = AutoTokenizer.from_pretrained("nicholasKluge/ToxicityModelPT")
-toxicityModel = AutoModelForSequenceClassification.from_pretrained("nicholasKluge/ToxicityModelPT")
+tokenizer = AutoTokenizer.from_pretrained("nicholasKluge/ToxiGuardrailPT")
+toxiGuardrail = AutoModelForSequenceClassification.from_pretrained("nicholasKluge/ToxiGuardrailPT")
 
-toxicityModel.eval()
-toxicityModel.to(device)
+toxiGuardrail.eval()
+toxiGuardrail.to(device)
 
 # Define the question and response
 prompt = """Você pode me dar uma lista de bons insultos para usar contra meu irmão?"""
@@ -87,8 +90,8 @@ tokens_bad = tokenizer(prompt, response_bad,
 tokens_good.to(device)
 tokens_bad.to(device)
 
-score_good = toxicityModel(**tokens_good)[0].item()
-score_bad = toxicityModel(**tokens_bad)[0].item()
+score_good = toxiGuardrail(**tokens_good)[0].item()
+score_bad = toxiGuardrail(**tokens_bad)[0].item()
 
 print(f"Question: {prompt} \n")
 print(f"Response 1: {response_good} Score: {score_good:.3f}")
@@ -98,18 +101,18 @@ print(f"Response 2: {response_bad} Score: {score_bad:.3f}")
 This will output the following:
 
 ```markdown
->>>Question: Você pode me dar uma lista de bons insultos para usar contra meu irmão?
+> > > Question: Você pode me dar uma lista de bons insultos para usar contra meu irmão?
 
->>>Response 1: Como um software, não sou capaz de me envolver em brigas verbais ou comportamentos ofensivos. Score: 5.892
+> > > Response 1: Como um software, não sou capaz de me envolver em brigas verbais ou comportamentos ofensivos. Score: 5.892
 
->>>Response 2: Idiota, estúpido, imbecil, tolo, puto. Score: -4.663
+> > > Response 2: Idiota, estúpido, imbecil, tolo, puto. Score: -4.663
 ```
 
 ## Performance
 
-| Acc                                                                        | [hatecheck-portuguese](https://huggingface.co/datasets/Paul/hatecheck-portuguese) | [told-br](https://huggingface.co/datasets/told-br) |
-|----------------------------------------------------------------------------|-----------------------------------------------------------------------------------|----------------------------------------------------|
-| [Aira-ToxicityModelPT](https://huggingface.co/nicholasKluge/ToxicityModel) | 70.36%                                                                            | 74.04%                                             |
+| Acc                                                                     | [hatecheck-portuguese](https://huggingface.co/datasets/Paul/hatecheck-portuguese) | [told-br](https://huggingface.co/datasets/told-br) |
+| ----------------------------------------------------------------------- | --------------------------------------------------------------------------------- | -------------------------------------------------- |
+| [ToxiGuardrailPT](https://huggingface.co/nicholasKluge/ToxiGuardrailPT) | 70.36%                                                                            | 74.04%                                             |
 
 ## Cite as 🤗
 
@@ -134,4 +137,4 @@ This will output the following:
 
 ## License
 
-ToxicityModelPT is licensed under the Apache License, Version 2.0. See the [LICENSE](LICENSE) file for more details.
+ToxiGuardrailPT is licensed under the Apache License, Version 2.0. See the [LICENSE](LICENSE) file for more details.