medical_en_zh_8_29

This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-zh on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6159
• Bleu: 41.4839
• Gen Len: 77.4048

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: 0.0004
• train_batch_size: 8
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
1.5915	1.02	3000	1.4640	30.8193	76.572
1.2908	2.04	6000	1.2734	32.3053	76.897
1.0814	3.06	9000	1.1348	34.3605	77.2082
0.9083	4.08	12000	1.0246	34.9139	76.7213
0.7507	5.1	15000	0.9336	36.2245	76.6036
0.6046	6.12	18000	0.8291	37.987	77.326
0.4838	7.14	21000	0.7496	38.7572	77.2366
0.3861	8.16	24000	0.6730	40.3566	77.49
0.3203	9.19	27000	0.6159	41.4839	77.4048

Framework versions

• Transformers 4.33.0.dev0
• Pytorch 2.0.0
• Datasets 2.1.0
• Tokenizers 0.13.3