ssc-pne-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: 1.3348
• Cer: 0.1645
• Wer: 0.5580

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
0.4835	0.2209	200	1.1821	0.1778	0.5903
0.5555	0.4417	400	1.1882	0.1750	0.5897
0.5099	0.6626	600	1.2536	0.1727	0.5835
0.4912	0.8835	800	1.2874	0.1785	0.5886
0.5023	1.1038	1000	1.2388	0.1713	0.5762
0.4751	1.3247	1200	1.2643	0.1734	0.5797
0.4609	1.5456	1400	1.2942	0.1716	0.5752
0.5015	1.7664	1600	1.3325	0.1698	0.5717
0.4386	1.9873	1800	1.3177	0.1688	0.5704
0.4438	2.2076	2000	1.3226	0.1679	0.5660
0.4216	2.4285	2200	1.2883	0.1686	0.5643
0.4447	2.6494	2400	1.2753	0.1685	0.5692
0.4652	2.8702	2600	1.3000	0.1675	0.5671
0.4392	3.0906	2800	1.3135	0.1673	0.5662
0.4564	3.3114	3000	1.3482	0.1656	0.5633
0.3526	3.5323	3200	1.3341	0.1659	0.5632
0.3607	3.7532	3400	1.3237	0.1667	0.5627
0.4063	3.9740	3600	1.3294	0.1657	0.5588
0.3676	4.1944	3800	1.3427	0.1661	0.5602
0.4109	4.4152	4000	1.3362	0.1658	0.5618
0.4045	4.6361	4200	1.3418	0.1651	0.5590
0.3534	4.8570	4400	1.3348	0.1645	0.5580

Framework versions

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