UGRIP_MST

Multilingual Statement Tuning

Running Evaluation Script

python evalscript.py --model ashabrawy/ST-roberta-base --tokenizer roberta-base

Datasets, Trained Models Repo in HuggingFace

https://huggingface.co/mbzuai-ugrip-statement-tuning

Running the Training Script

python3 statement_tune.py --Exp_name "XLMR" --transformer "xlm-roberta-base" --save "./XLMR2" The Base statement to train the XLMR model on english dataset, can add hyperparameter flags as needed. Supported flags: --tol --test_size --tr_ep --tr_batch --ev_batch --warmup --lr --decay