t5-base-finetuned-ancient_chinese-to-modern_chinese

This model is a fine-tuned version of t5-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1221
• Bleu: 84.7874
• Gen Len: 7.4143

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
0.1833	1.0	716	0.1371	83.007	7.5431
0.1528	2.0	1432	0.1286	84.1978	7.4289
0.1414	3.0	2148	0.1279	84.8682	7.4034
0.131	4.0	2864	0.1252	84.6009	7.4209
0.1298	5.0	3580	0.1250	84.7541	7.4146
0.1325	6.0	4296	0.1233	85.0001	7.4097
0.1284	7.0	5012	0.1235	84.7152	7.4122
0.1315	8.0	5728	0.1232	85.2833	7.4097
0.1276	9.0	6444	0.1231	84.7562	7.4104
0.1259	10.0	7160	0.1226	84.684	7.4139
0.1259	11.0	7876	0.1216	84.8757	7.4129
0.1257	12.0	8592	0.1221	84.6458	7.4143
0.1233	13.0	9308	0.1220	84.8371	7.4122
0.1217	14.0	10024	0.1218	84.7984	7.4115
0.1253	15.0	10740	0.1221	84.7874	7.4143

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1