squad

ALBERT model Pretraining and Fine Tuning using TF2.0

BERT which stands for Bidirectional Encoder Representations from Transformations is the SOTA in Transfer Learning in NLP.

Train Llama 2 & 3 on the SQuAD v2 task as an example of how to specialize a generalized (foundation) model.

JGLUE: Japanese General Language Understanding Evaluation for huggingface datasets