| Property | Value |
|---|---|
| Parameter Count | 109M parameters |
| License | Apache 2.0 |
| Best Accuracy | 80% |
| F2 Score | 89% |
| Paper | Research Paper |
What is twitter_sexismo-finetuned-robertuito-exist2021?
This is a specialized AI model designed to detect sexist content in Spanish-language tweets. Built on the RoBERTuito architecture, it's fine-tuned on the EXIST (sEXism Identification in Social neTworks) dataset. The model represents a significant advancement in content moderation for Spanish social media, with particular emphasis on identifying and flagging sexist remarks.
Implementation Details
The model utilizes the PyTorch framework and implements a fine-tuned version of the RoBERTuito base model. Training was conducted over 8 epochs with carefully optimized hyperparameters including a learning rate of 5E-5 and batch size of 32. The training process specifically optimized for F2 score to prioritize recall over precision in sexism detection.
- AdamW optimizer with epsilon=1e-08
- Linear learning rate scheduler
- Warm-up period of 3 epochs
- Mini-batch size of 32
Core Capabilities
- Binary classification of tweets as sexist or non-sexist
- High recall focus to minimize missed instances of sexism
- Optimized for Spanish language content
- Real-time tweet analysis capability
Frequently Asked Questions
Q: What makes this model unique?
The model's focus on Spanish-language content and its optimization for F2 score make it particularly effective at catching subtle forms of sexism while maintaining good overall accuracy. The emphasis on minimizing false negatives makes it especially suitable for content moderation.
Q: What are the recommended use cases?
The model is ideal for social media platforms, content moderators, and researchers studying online sexism in Spanish-language contexts. It can be integrated into real-time moderation systems or used for batch analysis of social media content.