Marcusxx/gyeongsan_1_6_model_50000CER

gyeongsan_1_6_model_50000

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

• Loss: 0.1386
• Cer: 10.5431

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: 100
• training_steps: 50000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.1752	0.2252	10000	0.1892	11.5421
0.1559	0.4503	20000	0.1678	11.1123
0.1676	0.6755	30000	0.1538	10.6121
0.1342	0.9006	40000	0.1427	10.6985
0.0933	1.1258	50000	0.1386	10.5431

Framework versions

• Transformers 4.41.2
• Pytorch 2.2.2+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1