byt5-small-wikipron-eng-latn-au-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: 0.1875
• Per: 0.3296
• Gen Len: 16.2507

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
2.4225	1.0	243	0.3885	0.5182	16.046
0.386	2.0	486	0.2610	0.4201	16.1472
0.2876	3.0	729	0.2242	0.3699	16.1972
0.2459	4.0	972	0.2073	0.3501	16.25
0.2236	5.0	1215	0.1966	0.3402	16.2254
0.207	6.0	1458	0.1953	0.337	16.2453
0.1971	7.0	1701	0.1879	0.3339	16.2523
0.1888	8.0	1944	0.1879	0.3319	16.2565
0.1829	9.0	2187	0.1869	0.3305	16.2509
0.1783	10.0	2430	0.1875	0.3296	16.2507

Framework versions

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