Overview
MSey/tiny_BROLLLT_0001.1 is a compact 1.1 billion parameter Llama-based model, uniquely trained on the BRO dataset. Its training incorporates Named Entity Recognition (NER) tags, labels, and tokens, making it highly specialized for tasks involving entity extraction and classification.
Key Capabilities
- Specialized NER: Demonstrates strong performance in identifying and classifying entities, as evidenced by its evaluation metrics.
- High F1 Scores: Achieved an average F1 score of 0.847 across its evaluation, with specific F1 scores of 0.737 for 'DIAG', 0.948 for 'MED', and 0.855 for 'TREAT' entities.
- Efficient Size: At 1.1 billion parameters, it offers a balance between performance and computational efficiency for specialized tasks.
- Contextual Prompting: Utilizes a clear prompt format (
### Context\n{Nachricht}\n\n### Answer) for structured input and output.
Good For
- Named Entity Recognition: Ideal for applications requiring precise identification and classification of entities within text, particularly in domains similar to the BRO dataset.
- Specialized Text Analysis: Suitable for use cases where extracting structured information from unstructured text is critical.
- Resource-Constrained Environments: Its smaller parameter count makes it a viable option for deployment in environments with limited computational resources, while still delivering strong performance on its target tasks.