news_summarization_model

This model is a fine-tuned version of eenzeenee/t5-base-korean-summarization on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2221
• Rouge1: 0.6031
• Rouge2: 0.4193
• Rougel: 0.5896
• Rougelsum: 0.5892
• Gen Len: 42.2714

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 400
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
0.523	0.4762	100	0.3847	0.4833	0.2872	0.4627	0.4654	41.7571
0.3377	0.9524	200	0.2590	0.5709	0.3929	0.5581	0.5564	41.7810
0.2641	1.4286	300	0.2334	0.5889	0.4018	0.5785	0.5781	42.5048
0.2361	1.9048	400	0.2221	0.6031	0.4193	0.5896	0.5892	42.2714

Framework versions

• Transformers 4.55.0
• Pytorch 2.6.0+cu124
• Datasets 4.0.0
• Tokenizers 0.21.4