mbart-large-50-finetuned-ua-gec-2.1

This model is a fine-tuned version of facebook/mbart-large-50 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5376
• Rouge1: 18.2963
• Rouge2: 10.2365
• Rougel: 18.2593
• Rougelsum: 18.2759
• Gen Len: 28.6107

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
0.0765	1.0	1010	0.4070	18.2963	10.2365	18.2593	18.2759	28.522
0.046	2.0	2020	0.4710	18.2963	10.2365	18.2593	18.2759	28.578
0.0291	3.0	3030	0.4885	18.2833	10.2052	18.2454	18.263	28.5793
0.0188	4.0	4040	0.5145	18.2963	10.2365	18.2593	18.2759	28.6127
0.0117	5.0	5050	0.5376	18.2963	10.2365	18.2593	18.2759	28.6107

Framework versions

• Transformers 4.29.2
• Pytorch 2.0.1+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3