xlmRoberta-for-VietnameseQA

This model is a fine-tuned version of xlm-roberta-base on the UIT-Viquad_v2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.8315

Model description

Fine-tuned by Honganh Nguyen (FPTU AI Club).

Intended uses & limitations

More information needed

Training and evaluation data

Credits to Viet Nguyen (FPTU AI Club) for the training and evaluation data.

Training data: https://github.com/vietnguyen012/QA_viuit/blob/main/train.json

Evaluation data: https://github.com/vietnguyen012/QA_viuit/blob/main/trial/trial.json

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 12
• eval_batch_size: 12
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
1.5701	1.0	2534	1.2220
1.2942	2.0	5068	0.9698
1.0693	3.0	7602	0.8315

Framework versions

• Transformers 4.15.0
• Pytorch 1.10.0+cu111
• Datasets 1.17.0
• Tokenizers 0.10.3