ssc-lth-mms-model-mix-adapt-max-lowlr

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

• Loss: 0.5966
• Cer: 0.1756
• Wer: 0.4165

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: 0.0003
• train_batch_size: 2
• eval_batch_size: 6
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 100
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer	Wer
0.5919	0.5594	200	0.5687	0.1728	0.4225
0.8336	1.1175	400	0.5684	0.1721	0.4067
0.5713	1.6769	600	0.5492	0.1716	0.4188
1.1595	2.2350	800	0.5365	0.1688	0.4114
1.7524	2.7944	1000	0.5756	0.1698	0.4063
0.64	3.3524	1200	0.5606	0.1696	0.4108
0.614	3.9119	1400	0.5883	0.1748	0.4214
0.6868	4.4699	1600	0.5966	0.1756	0.4165

Framework versions

• Transformers 4.57.2
• Pytorch 2.9.1+cu128
• Datasets 3.6.0
• Tokenizers 0.22.0