ssc-kcn-xlsr300m-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: 1.9899
• Cer: 0.2907
• Wer: 0.6677

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
0.3955	0.5731	200	1.6748	0.3252	0.7384
0.4201	1.1461	400	1.7072	0.3066	0.7196
0.3296	1.7192	600	1.5138	0.3243	0.7236
0.25	2.2923	800	1.7316	0.3004	0.6943
0.2522	2.8653	1000	1.7450	0.3003	0.7068
0.1796	3.4384	1200	2.0827	0.2903	0.6722
0.1629	4.0115	1400	1.9182	0.2914	0.6766
0.1063	4.5845	1600	1.9899	0.2907	0.6677

Framework versions

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