base_sami_22k_ftallpseudo_ftlabelled_sami_parliament_alld0

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 189.4120
• Wer: 0.3890
• Cer: 0.1332

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.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_ratio: 0.25
• num_epochs: 60.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
543.4487	1.0	446	189.4152	0.3905	0.1331
468.607	2.0	892	195.9747	0.3961	0.1389
431.6385	3.0	1338	213.4495	0.4097	0.1307
423.9342	4.0	1784	255.1584	0.4358	0.1790
428.2925	5.0	2230	242.8004	0.4474	0.1565
446.372	6.0	2676	294.5761	0.4721	0.1664

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1
• Datasets 3.2.0
• Tokenizers 0.21.0