ko-news-summarization

This model is a fine-tuned version of psyche/KoT5-summarization on an unknown dataset. It achieves the following results on the evaluation set:

• Train Loss: 0.9267
• Validation Loss: 1.2897
• Train Rougel: tf.Tensor(0.39195082, shape=(), dtype=float32)
• Epoch: 6

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:

• optimizer: {'name': 'Adam', 'learning_rate': 2e-05, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
• training_precision: float32

Training results

Train Loss	Validation Loss	Train Rougel	Epoch
1.2735	1.2266	tf.Tensor(0.38460648, shape=(), dtype=float32)	0
1.2050	1.2290	tf.Tensor(0.38383868, shape=(), dtype=float32)	1
1.1428	1.2320	tf.Tensor(0.38434482, shape=(), dtype=float32)	2
1.0866	1.2497	tf.Tensor(0.38514885, shape=(), dtype=float32)	3
1.0311	1.2612	tf.Tensor(0.3883608, shape=(), dtype=float32)	4
0.9740	1.2713	tf.Tensor(0.39306718, shape=(), dtype=float32)	5
0.9267	1.2897	tf.Tensor(0.39195082, shape=(), dtype=float32)	6

Framework versions

• Transformers 4.25.1
• TensorFlow 2.9.2
• Datasets 2.8.0
• Tokenizers 0.13.2