fine-tuned-DatasetQAS-TYDI-QA-ID-with-xlm-roberta-large-with-ITTL-with-freeze-LR-1e-05

This model is a fine-tuned version of xlm-roberta-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.9154
• Exact Match: 67.4296
• F1: 80.7483

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Exact Match	F1
6.331	0.5	19	3.7275	5.2817	16.4975
6.331	0.99	38	2.5293	22.3592	33.2570
3.6805	1.5	57	1.5504	45.4225	61.5302
3.6805	1.99	76	1.2025	57.2183	72.1651
3.6805	2.5	95	1.0664	61.0915	75.6496
1.3982	2.99	114	0.9926	63.2042	77.6464
1.3982	3.5	133	0.9823	64.6127	78.3848
0.9533	3.99	152	0.9596	66.1972	79.5651
0.9533	4.5	171	0.9578	67.4296	80.6710
0.9533	4.99	190	0.9376	68.3099	80.8025
0.7418	5.5	209	0.9393	67.4296	79.8821
0.7418	5.99	228	0.9242	67.4296	79.9318
0.7418	6.5	247	0.9154	67.4296	80.7483

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu117
• Datasets 2.2.0
• Tokenizers 0.13.2