CHUNGNAMADDRSSTranslate_model

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

• Loss: 0.0166
• Cer: 13.1299

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: 10000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.0835	0.7704	1000	0.0637	25.5441
0.0255	1.5408	2000	0.0417	1.1239
0.0073	2.3112	3000	0.0241	45.3216
0.0036	3.0817	4000	0.0196	0.5121
0.0006	3.8521	5000	0.0179	36.8526
0.0007	4.6225	6000	0.0187	3.9286
0.0004	5.3929	7000	0.0171	3.6235
0.0003	6.1633	8000	0.0170	4.4623
0.0005	6.9337	9000	0.0165	16.3369
0.0001	7.7042	10000	0.0166	13.1299

Framework versions

• Transformers 4.41.2
• Pytorch 2.2.2+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1