namphungdn134/whisper-base-vi

Whisper Base Vi - Nam Phung

This model is a fine-tuned version of openai/whisper-base on the vlsp2020_vinai_100h dataset. It achieves the following results on the evaluation set:

• Loss: 0.4049
• Wer: 20.3964

To use this model, Let's get here: https://github.com/namphung134/np-asr-vietnamese

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 32
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 3500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.7425	0.1492	250	0.7303	34.7993
0.6393	0.2983	500	0.6105	30.0462
0.5636	0.4475	750	0.5404	28.1320
0.5199	0.5967	1000	0.5043	25.5525
0.4806	0.7458	1250	0.4756	24.4785
0.4779	0.8950	1500	0.4581	23.8864
0.414	1.0442	1750	0.4447	22.6037
0.3967	1.1933	2000	0.4336	21.2506
0.3723	1.3425	2250	0.4243	21.8426
0.3886	1.4916	2500	0.4179	21.3605
0.3876	1.6408	2750	0.4128	20.8728
0.3459	1.7900	3000	0.4086	20.7572
0.3546	1.9391	3250	0.4058	20.5909
0.3208	2.0883	3500	0.4049	20.3964

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.0
• Tokenizers 0.21.1