Modelpublic

minhvn1433/wav2vec2-base-vi-vlsp2020-ft-ViMD_central_phowhisper

sourceHugging Faceupdated 3mo agoView on Hugging Face

0likes4downloads

Model Card

wav2vec2-base-vi-vlsp2020-ft-ViMDcentralphowhisper

This model is a fine-tuned version of minhvn1433/wav2vec2-base-vi-vlsp2020 on the MINHVN1433/VIMDCENTRALPHOWHISPER - NA dataset. It achieves the following results on the evaluation set:

—Loss: 0.5676
—Wer: 0.1943
—Cer: 0.1165

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: 3e-05
—trainbatchsize: 16
—evalbatchsize: 16
—seed: 42
—optimizer: Use OptimizerNames.ADAMWTORCHFUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
—lrschedulertype: linear
—lrschedulerwarmup_steps: 500.0
—training_steps: 3000
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.4430	3.0675	500	0.5160	0.1970	0.1177
0.3985	6.1350	1000	0.5459	0.2015	0.1164
0.3677	9.2025	1500	0.5584	0.1981	0.1171
0.3444	12.2699	2000	0.5765	0.1957	0.1172
0.3353	15.3374	2500	0.5676	0.1943	0.1165
0.3280	18.4049	3000	0.5872	0.1955	0.1173

Framework versions

—Transformers 5.13.0
—Pytorch 2.11.0+cu128
—Datasets 3.6.0
—Tokenizers 0.22.2