base_sami_22k_cont_pt_ftpseudo_shuff1048wr30esp5

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

• Loss: 301.8862
• Wer: 0.5090
• Cer: 0.1701

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: 0.3
• num_epochs: 60.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
5346.3694	1.0	1080	1150.5342	1.0	1.0
3391.3456	2.0	2160	1049.4224	1.0	0.9995
1439.7179	3.0	3240	304.0005	0.5745	0.1837
662.4384	4.0	4320	300.4795	0.5312	0.1843
544.894	5.0	5400	301.7199	0.5099	0.1691
527.8369	6.0	6480	351.5502	0.5780	0.2033
484.5173	7.0	7560	340.1422	0.5186	0.1878
492.0863	8.0	8640	426.7222	0.5788	0.2232

Framework versions

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