ssc-koo-mms-model-mix-adapt-max

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

• Loss: 0.8345
• Cer: 0.2095
• Wer: 0.7234

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: 0.001
• train_batch_size: 2
• eval_batch_size: 8
• 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
0.8208	0.4756	200	0.9063	0.2333	0.7841
0.9436	0.9512	400	0.9141	0.2212	0.7739
0.7226	1.4257	600	0.8844	0.2240	0.7683
0.8434	1.9013	800	0.8876	0.2151	0.7380
0.6211	2.3757	1000	0.8576	0.2136	0.7403
0.8917	2.8514	1200	0.8440	0.2175	0.7522
0.6976	3.3258	1400	0.8561	0.2091	0.7213
0.6136	3.8014	1600	0.8439	0.2099	0.7287
0.5528	4.2759	1800	0.8574	0.2114	0.7274
0.6116	4.7515	2000	0.8345	0.2095	0.7234

Framework versions

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