In this blog post, we will explore how to effectively use a fine-tuned model designed for racism detection in Spanish. This model is based on BETO, a Spanish version of BERT, and tailored using the *Datathon Against Racism* dataset from 2022. Whether you’re a seasoned developer or a curious enthusiast, follow these steps to harness the power of this incredible tool.

Step 1: Set Up Your Environment

Before diving into coding, ensure you have the necessary libraries installed. For this model, you will need the ‘transformers’ library. You can install it using pip:

pip install transformers

Step 2: Load the Model

Once your environment is set up, you can begin to load the pre-trained model. For analogy, think of it as preparing a chef’s kitchen before cooking a gourmet dish; everything should be in place. Here is how to do it:

from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline

model_name = "raw-label-epoch-4"
tokenizer = AutoTokenizer.from_pretrained("dccuchile/bert-base-spanish-wwm-uncased")

full_model_path = f"MartinoMensio/racism-models/{model_name}"
model = AutoModelForSequenceClassification.from_pretrained(full_model_path)

pipe = pipeline("text-classification", model=model, tokenizer=tokenizer)

In this snippet, we are loading the tokenizer and model. The tokenizer is like a translator that turns your ingredients (text) into a format that the model can understand.

Step 3: Classify Texts

Now let’s classify some texts for racism detection. We will run the pipeline with our example sentences:

texts = [
    "y porqué es lo que hay que hacer con los menas y con los adultos también!!!! NO a los inmigrantes ilegales!!!!",
    "Es que los judíos controlan el mundo"
]
print(pipe(texts))

In this code segment, we are feeding our sentences into the model, much like serving a dish to a panel of food critics, who will evaluate its qualities (in this case, the scores reflecting racism detection).

Understanding the Output

The output you will receive will include labels like “racist” or “non-racist” along with a score representing the model’s confidence in its classification. For example:

# [label: racist, score: 0.921501636505127, label: non-racist, score: 0.9459075331687927]

Troubleshooting

If you encounter any issues while using the model, consider the following solutions:

• Model Not Found: Ensure the model name is correct and you have an internet connection.
• Installation Issues: Check that the ‘transformers’ package is correctly installed. Run pip show transformers to verify.
• Text Not Classifying: Ensure the input text is in Spanish and properly formatted.

For more insights, updates, or to collaborate on AI development projects, stay connected with fxis.ai.

Conclusion

And there you have it! With just a few steps, you can leverage the power of this racism detection model. It emphasizes the importance of addressing social issues through technology. At fxis.ai, we believe that such advancements are crucial for the future of AI, as they enable more comprehensive and effective solutions. Our team is continually exploring new methodologies to push the envelope in artificial intelligence, ensuring that our clients benefit from the latest technological innovations.