|
|
--- |
|
|
license: mit |
|
|
language: |
|
|
- es |
|
|
- en |
|
|
metrics: |
|
|
- accuracy |
|
|
--- |
|
|
|
|
|
# HelloBERT |
|
|
|
|
|
HelloBERT is a fine-tuned BERT model for text classification, designed to identify greetings (`greeting`) and non-greetings (`not_greeting`) in both English and Spanish. The model is based on `bert-base-uncased`. |
|
|
|
|
|
## Intended Use |
|
|
|
|
|
This model is intended for identifying whether a given text is a greeting or not. It can be used in various applications such as chatbots, automated customer service, and text analysis tools. |
|
|
|
|
|
## Training Data |
|
|
|
|
|
The model was fine-tuned on a custom dataset containing examples of greetings and non-greetings in English and Spanish. You can find the dataset [here](https://github.com/luzalbaposse/HelloBERT/tree/main). |
|
|
|
|
|
## Training Procedure |
|
|
|
|
|
- **Base Model:** `bert-base-uncased` |
|
|
- **Training Steps:** 3 epochs |
|
|
- **Batch Size:** 16 |
|
|
- **Learning Rate:** 2e-5 |
|
|
- **Optimizer:** AdamW |
|
|
- **Evaluation Metric:** Accuracy |
|
|
|
|
|
## How to Use |
|
|
|
|
|
You can use this model directly with the Hugging Face `transformers` library as follows: |
|
|
|
|
|
```python |
|
|
from transformers import AutoModelForSequenceClassification, AutoTokenizer, pipeline |
|
|
|
|
|
# Load the model and tokenizer |
|
|
model = AutoModelForSequenceClassification.from_pretrained("luzalbaposse/HelloBERT") |
|
|
tokenizer = AutoTokenizer.from_pretrained("luzalbaposse/HelloBERT") |
|
|
|
|
|
# Create the pipeline |
|
|
classifier = pipeline('text-classification', model=model, tokenizer=tokenizer) |
|
|
|
|
|
# Example usage |
|
|
text = "Qué ondaaa??" |
|
|
result = classifier(text) |
|
|
print(f"Text: {text} => Label: {result[0]['label']}, Score: {result[0]['score']}") |
|
|
``` |
|
|
|
|
|
## Limitations and Biases |
|
|
|
|
|
While the model performs well on the training and validation sets, it may not generalize perfectly to all types of text inputs, especially those that are significantly different from the training data. Be cautious of potential biases in the training data that may affect model performance. |
|
|
|
|
|
## Author |
|
|
|
|
|
This model was fine-tuned by [Luz Alba Posse](https://huggingface.co/luzalbaposse). |
|
|
|
|
|
## License |
|
|
|
|
|
This model is licensed under the MIT License. |
|
|
|
