YonaKhine

This model is a fine-tuned version of facebook/w2v-bert-2.0 on the OpenSLR-80 Burmese Speech Dataset dataset. It achieves the following results on the evaluation set: - Loss: 0.3886 - Wer: 0.4256 The following hyperparameters were used during training: - learningrate: 5e-05 - trainbatchsize: 16 - evalbatchsize: 8 - seed: 42 - gradientaccumulationsteps: 2 - totaltrainbatchsize: 32 - optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments - lrschedulertype: linear - lrschedulerwarmupsteps: 500 - numepochs: 10 - mixedprecisiontraining: Native AMP | Training Loss | Epoch | Step | Validation Loss | Wer | |:-------------:|:------:|:----:|:---------------:|:------:| | 0.2901 | 4.1678 | 300 | 0.4585 | 0.5053 | | 0.2127 | 8.3357 | 600 | 0.3886 | 0.4256 | 🎤 Audio Sample You can listen to an example where a user says: 📝 Model Transcription > မင်္ဂလာ ပါ တွေ့ရတာ ဝမ်းသပါတယ် ရှင် 📉 Word Error Rate (WER) WER ≈ 0.10 – minor vowel difference only, excellent semantic preservation. 🎧 Audio Playback Download and listen: burmesetest.wav - Transformers 4.51.3 - Pytorch 2.6.0+cu124 - Datasets 3.6.0 - Tokenizers 0.21.1