byt5-small-wikipron-eng-latn-us-broad-p2g

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: 0.2595
• Per: 0.4628
• Gen Len: 8.4996

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
2.4797	1.0	382	0.4371	0.6951	8.4302
0.4823	2.0	764	0.3543	0.5974	8.4338
0.3878	3.0	1146	0.3081	0.545	8.4394
0.3378	4.0	1528	0.2904	0.518	8.449
0.3061	5.0	1910	0.2736	0.5004	8.4612
0.2823	6.0	2292	0.2664	0.4893	8.4734
0.265	7.0	2674	0.2626	0.4747	8.4721
0.2502	8.0	3056	0.2612	0.4697	8.4945
0.2388	9.0	3438	0.2592	0.4633	8.489
0.231	10.0	3820	0.2595	0.4628	8.4996

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.0+cu117
• Datasets 2.11.1.dev0
• Tokenizers 0.13.2