byt5-small-wikipron-eng-latn-uk-broad-ft-au-broad

This model is a fine-tuned version of bookbot/byt5-small-wikipron-eng-latn-uk-broad on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1618
• Per: 0.3076
• Gen Len: 16.2858

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: 128
• eval_batch_size: 32
• seed: 42
• 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
0.4971	1.0	243	0.1686	0.3211	16.2676
0.1917	2.0	486	0.1600	0.3119	16.2638
0.1713	3.0	729	0.1578	0.3129	16.2714
0.1583	4.0	972	0.1576	0.3095	16.2768
0.1496	5.0	1215	0.1572	0.3107	16.2721
0.1413	6.0	1458	0.1592	0.3096	16.2676
0.1356	7.0	1701	0.1592	0.3094	16.2773
0.1301	8.0	1944	0.1592	0.3084	16.2771
0.1261	9.0	2187	0.1605	0.3076	16.2816
0.123	10.0	2430	0.1618	0.3076	16.2858

Framework versions

• Transformers 4.28.0.dev0
• Pytorch 1.13.1+cu117
• Datasets 2.9.0
• Tokenizers 0.13.2