Rufiiii/umt5_base_version1

Elite_Enigmas/google/umt5-base

This model is a fine-tuned version of google/umt5-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0961
• Wer: 0.2812

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: 0.0005
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2000
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
7.4871	4.0	4000	2.0314	1.2812
0.9487	8.0	8000	1.0738	0.3125
0.4547	12.0	12000	1.0961	0.2812

Framework versions

• Transformers 4.47.1
• Pytorch 2.4.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0