opus-mt-zh-en-ft

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

• Loss: 1.8263
• Model Preparation Time: 0.0084
• Bleu: 21.5105

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: 3e-05
• train_batch_size: 64
• eval_batch_size: 128
• seed: 42
• optimizer: Use adamw_torch_fused with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 9
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Bleu
2.1306	1.0	79841	2.0096	0.0084	19.5707
2.0201	2.0	159682	1.9438	0.0084	20.0972
1.9758	3.0	239523	1.9070	0.0084	20.5279
1.9308	4.0	319364	1.8823	0.0084	20.7887
1.8848	5.0	399205	1.8643	0.0084	20.9881
1.8479	6.0	479046	1.8494	0.0084	21.1376
1.8189	7.0	558887	1.8387	0.0084	21.2779
1.8127	8.0	638728	1.8300	0.0084	21.4233
1.8066	9.0	718569	1.8263	0.0084	21.5102

Framework versions

• Transformers 4.57.3
• Pytorch 2.9.1+cu128
• Datasets 2.18.0
• Tokenizers 0.22.2