T2R2: Train, test, record, repeat: incremental framework for NLP model training
[ 1 ] Wydział Informatyki i Telekomunikacji, Politechnika Poznańska | [ 2 ] Instytut Informatyki, Wydział Informatyki i Telekomunikacji, Politechnika Poznańska | [ S ] student | [ P ] pracownik
2024
rozdział w monografii naukowej / referat
angielski
- deep learning
- iterative training
- MLOps
- NLP
EN In this paper, we introduce an iterative training loop framework for natural language processing models, facilitating the seamless integration of various data-building tools for training, testing, and validation sets. Our approach empowers researchers and practitioners with flexible model training capabilities, enhanced by connectors to an assortment of modern NLP resources. Additionally, the framework integrates MLOps features such as automatic versioning, ensuring reproducibility, and streamlining the model development lifecycle. By enabling continuous refinement and evaluation, our solution paves the way for more robust and accurate NLP models that can be adapted to dynamic real-world datasets.
218 - 225
copyright
witryna wydawcy
ostateczna wersja opublikowana
w momencie opublikowania
20