chungnam_large_model

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

• Loss: 0.0542
• Cer: 31.5683

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: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.1181	1.6556	250	0.1939	72.2014
0.0217	3.3113	500	0.0617	74.5324
0.0045	4.9669	750	0.0495	29.1223
0.0019	6.6225	1000	0.0553	40.1439
0.0001	8.2781	1250	0.0544	27.1079
0.0001	9.9338	1500	0.0541	28.0288
0.0001	11.5894	1750	0.0541	30.8201
0.0001	13.2450	2000	0.0542	31.5683

Framework versions

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