Text Classification
Transformers
PyTorch
English
bert
misogyny detection
abusive language
hate speech
offensive language
text-embeddings-inference
Instructions to use MilaNLProc/bert-base-uncased-ear-misogyny with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use MilaNLProc/bert-base-uncased-ear-misogyny with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="MilaNLProc/bert-base-uncased-ear-misogyny")# Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("MilaNLProc/bert-base-uncased-ear-misogyny") model = AutoModelForSequenceClassification.from_pretrained("MilaNLProc/bert-base-uncased-ear-misogyny") - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -32,7 +32,7 @@ We trained a total of three versions for Italian and English.
|
|
| 32 |
| Model | Download |
|
| 33 |
| ------ | -------------------------|
|
| 34 |
| `bert-base-uncased-ear-misogyny` | [Link](https://huggingface.co/MilaNLProc/bert-base-uncased-ear-misogyny) |
|
| 35 |
-
| `bert-base-uncased-ear-mlma` | [Link]() |
|
| 36 |
| `bert-base-uncased-ear-misogyny-italian` | [Link](https://huggingface.co/MilaNLProc/bert-base-uncased-ear-misogyny-italian) |
|
| 37 |
|
| 38 |
# Authors
|
|
@@ -66,7 +66,8 @@ Please use the following BibTeX entry if you use this model in your project:
|
|
| 66 |
|
| 67 |
# Limitations
|
| 68 |
|
| 69 |
-
Entropy-Attention Regularization mitigates lexical overfitting but does not completely remove it. We expect the model still to show biases, e.g., peculiar keywords that induce a specific prediction regardless of the context.
|
|
|
|
| 70 |
Please refer to our paper for a quantitative evaluation of this mitigation.
|
| 71 |
|
| 72 |
## License
|
|
|
|
| 32 |
| Model | Download |
|
| 33 |
| ------ | -------------------------|
|
| 34 |
| `bert-base-uncased-ear-misogyny` | [Link](https://huggingface.co/MilaNLProc/bert-base-uncased-ear-misogyny) |
|
| 35 |
+
| `bert-base-uncased-ear-mlma` | [Link](https://huggingface.co/MilaNLProc/bert-base-uncased-ear-mlma) |
|
| 36 |
| `bert-base-uncased-ear-misogyny-italian` | [Link](https://huggingface.co/MilaNLProc/bert-base-uncased-ear-misogyny-italian) |
|
| 37 |
|
| 38 |
# Authors
|
|
|
|
| 66 |
|
| 67 |
# Limitations
|
| 68 |
|
| 69 |
+
Entropy-Attention Regularization mitigates lexical overfitting but does not completely remove it. We expect the model still to show biases, e.g., peculiar keywords that induce a specific prediction regardless of the context.
|
| 70 |
+
|
| 71 |
Please refer to our paper for a quantitative evaluation of this mitigation.
|
| 72 |
|
| 73 |
## License
|