ViT5

This model is a fine-tuned version of VietAI/vit5-base-vietnews-summarization on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5754671692848206
• Rouge-1: 0.2966
• Rouge-2: 0.1572
• Rouge-4: 0.0899
• Rouge-l: 0.2582
• Rouge-w-1.2: 0.1046
• Rouge-s4: 0.1182
• Rouge-su4: 0.1485
• R: 0.2336
• P: 0.4058
• Bleu: 12.7017

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: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge-1	Rouge-2	Rouge-4	Rouge-l	Rouge-w-1.2	Rouge-s4	Rouge-su4	R	P	Bleu
0.6619	1.0	628	0.6740	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.67	2.0	1256	0.6519	0.2971	0.1722	0.1027	0.2659	0.1072	0.1353	0.1628	0.2169	0.4713	11.5398
0.5606	3.0	1884	0.6481	0.3067	0.1597	0.0794	0.2693	0.1064	0.1184	0.1504	0.2311	0.4557	10.4569
0.5006	4.0	2512	0.6499	0.3029	0.1619	0.0815	0.2663	0.1051	0.1193	0.1504	0.229	0.4474	10.9827

Framework versions

• Transformers 4.26.1
• Pytorch 2.0.0
• Datasets 2.1.0
• Tokenizers 0.13.3