Llama2-MedTuned-7b

Model Description

Llama2-MedTuned-7b is an instruction-tuned version of the Llama2 7B model, specifically adapted for biomedical language processing tasks. It has been fine-tuned on a dataset consisting of approximately 200,000 instruction-focused samples, covering a range of biomedical and clinical NLP tasks such as Named Entity Recognition (NER), Relation Extraction (RE), and Medical Natural Language Inference (NLI).

Instruction Tuning Procedure

This model underwent instruction tuning, a process where the model is fine-tuned with detailed instructions to enhance its ability to interpret and execute specific tasks in the biomedical domain. The tuning involved the use of a comprehensive instruction-based dataset, tailor-made to align with the requirements of biomedical NLP tasks.

Model Capabilities

Llama2-MedTuned-7b demonstrates an enhanced understanding of biomedical contexts, effectively handling NER, RE, and NLI tasks. It showcases improved accuracy in generating structured outputs suitable for evaluation using conventional metrics.

Architecture

The architecture of Llama2-MedTuned-7b is based on the autoregressive transformer model Llama2 7B. This model maintains the original transformer layers and attention mechanisms, specifically adjusted to cater to the linguistic intricacies of the biomedical field.

Citation

If you utilise Llama2-MedTuned-7b in your research or application, please consider citing our paper:

@article{rohanian2024exploring,
  title = {Exploring the Effectiveness of Instruction Tuning in Biomedical Language Processing},
  author = {Rohanian, Omid and Nouriborji, Mohammadmahdi and Kouchaki, Samaneh and Nooralahzadeh, Farhad and Clifton, Lei and Clifton, David A},
  journal = {Artificial Intelligence in Medicine},
  volume = {158},
  pages = {103007},
  year = {2024},
  publisher = {Elsevier},
  doi = {10.1016/j.artmed.2024.103007},
  url = {https://www.sciencedirect.com/science/article/pii/S0933365724002495},
  issn = {0933-3657}
}

Support

If this model helps your work, you can keep the project running with a one-off or monthly contribution:
https://github.com/sponsors/nlpie-research