byt5-small-wikipron-eng-latn

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

• Loss: 0.1898
• Per: 0.3272
• Gen Len: 16.4158

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: 0.0002
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Per	Gen Len
2.2564	1.0	230	0.4016	0.5235	15.7351
0.3856	2.0	461	0.2648	0.4189	16.3283
0.2861	3.0	692	0.2248	0.3665	16.3982
0.2438	4.0	923	0.2090	0.3452	16.3591
0.2207	5.0	1153	0.2015	0.3403	16.3944
0.2049	6.0	1384	0.1952	0.3342	16.4001
0.193	7.0	1615	0.1908	0.3306	16.4006
0.185	8.0	1846	0.1883	0.3271	16.408
0.18	9.0	2076	0.1894	0.3276	16.4194
0.1751	9.97	2300	0.1898	0.3272	16.4158

Framework versions

• Transformers 4.29.2
• Pytorch 2.0.1+cu117
• Datasets 2.12.0
• Tokenizers 0.13.3