Watarungurunnn/w2v-bert-2.0-japanese-CV16.0_aynita_1

w2v-bert-2.0-japanese-CV16.0

This model is a fine-tuned version of Watarungurunnn/w2v-bert-2.0-japanese-CV16.0 on the common_voice_16_0 dataset. It achieves the following results on the evaluation set:

• Loss: 1.5149
• Wer: 32.6188

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: 5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1205	1.69	500	1.3602	36.4355
0.2116	3.39	1000	1.4580	35.1067
0.1054	5.08	1500	1.4180	34.6457
0.0661	6.78	2000	1.4557	32.3889
0.0208	8.47	2500	1.5149	32.6188

Framework versions

• Transformers 4.37.2
• Pytorch 2.2.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2