byt5-translation-model

This model is a fine-tuned version of google/byt5-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3816

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• total_train_batch_size: 16
• total_eval_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 60

Training results

Training Loss	Epoch	Step	Validation Loss
1.0026	6.8182	1200	0.7811
0.7206	13.6364	2400	0.5618
0.6075	20.4545	3600	0.4895
0.5543	27.2727	4800	0.4482
0.5113	34.0909	6000	0.4200
0.4924	40.9091	7200	0.3995
0.4676	47.7273	8400	0.3878
0.4692	54.5455	9600	0.3816

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu126
• Datasets 4.4.2
• Tokenizers 0.22.1