Hubert-noisy-cv-kakeiken-debug

This model is a fine-tuned version of rinna/japanese-hubert-base on the ORIGINAL_NOISY_KAKEIKEN_W - JA dataset. It achieves the following results on the evaluation set:

• Loss: 5.3683
• Wer: 1.0
• Cer: 1.6786

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: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 12500
• num_epochs: 3.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.2029	1.0	2732	2.6634	1.0	1.5117
0.0342	2.0	5464	4.7376	1.0	1.5111
0.0275	2.9991	8193	5.3590	1.0	1.6778

Framework versions

• Transformers 4.48.0
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.21.0