spark-name-zh-to-en

This model is a fine-tuned version of ihebaker10/spark-name-ka-to-en on the None dataset. It achieves the following results on the evaluation set:

• Loss: 4.3571
• Bleu: 0.4007
• Gen Len: 13.1744

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: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
4.5771	1.0	1600	4.5729	0.6348	48.9281
4.2727	2.0	3200	4.4437	0.3449	13.7372
4.1169	3.0	4800	4.3856	0.3302	14.0606
3.9978	4.0	6400	4.3584	0.3945	13.1856
3.9164	5.0	8000	4.3571	0.4007	13.1744

Framework versions

• Transformers 4.41.0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1