Jinho11/repo_name

Kim_Sohn_Whisper

This model is a fine-tuned version of openai/whisper-base on the jinho_data dataset. It achieves the following results on the evaluation set:

• Loss: 0.6363
• Cer: 15.4890

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: 1e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.3344	2.43	1000	0.5535	19.6948
0.1546	4.85	2000	0.5602	16.2194
0.0505	7.28	3000	0.6095	15.7197
0.0266	9.71	4000	0.6363	15.4890

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0