byt5-small-wikipron-eng-latn-us-broad

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: 0.1668
• Per: 0.2588
• Gen Len: 15.7318

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
1.9462	1.0	382	0.2963	0.3495	15.453
0.2979	2.0	764	0.2144	0.2941	15.6759
0.2319	3.0	1146	0.1898	0.2765	15.7219
0.2042	4.0	1528	0.1799	0.2719	15.7235
0.1879	5.0	1910	0.1744	0.2657	15.7054
0.1763	6.0	2292	0.1729	0.2645	15.7406
0.1677	7.0	2674	0.1693	0.2619	15.7269
0.1613	8.0	3056	0.1668	0.2604	15.7371
0.1559	9.0	3438	0.1667	0.2585	15.7383
0.1534	10.0	3820	0.1668	0.2588	15.7318

Framework versions

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