vit5-base-v2-paraphase_data

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

• Loss: 1.1037

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.3847	0.4980	3300	1.2644
1.2837	0.9961	6600	1.1851
1.23	1.4941	9900	1.1554
1.2291	1.9922	13200	1.1341
1.1839	2.4902	16500	1.1218
1.1908	2.9882	19800	1.1150
1.1778	3.4863	23100	1.1090
1.1674	3.9843	26400	1.1044
1.1712	4.4823	29700	1.1038
1.1709	4.9804	33000	1.1037

Framework versions

• PEFT 0.10.0
• Transformers 4.49.0
• Pytorch 2.4.1+cu118
• Datasets 4.1.1
• Tokenizers 0.21.0