ssc-bew-mms-model-mix-adapt-max3

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

• Loss: 1.2699
• Cer: 0.1949
• Wer: 0.5862

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: 1
• eval_batch_size: 6
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 2
• 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_steps: 100
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer	Wer
0.4718	0.4920	200	1.2211	0.2140	0.6304
0.5056	0.9840	400	1.2124	0.2094	0.6213
0.4584	1.4748	600	1.2072	0.1964	0.5913
0.4492	1.9668	800	1.2265	0.1978	0.5935
0.4001	2.4576	1000	1.2681	0.1984	0.5963
0.4182	2.9496	1200	1.2458	0.1952	0.5863
0.3669	3.4403	1400	1.2681	0.1951	0.5868
0.3742	3.9323	1600	1.2672	0.1951	0.5880
0.3686	4.4231	1800	1.2680	0.1969	0.5905
0.346	4.9151	2000	1.2699	0.1949	0.5862

Framework versions

• Transformers 4.52.1
• Pytorch 2.9.1+cu128
• Datasets 3.6.0
• Tokenizers 0.21.4