This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_11_0 dataset. It achieves the following results on the evaluation set:

Loss: 0.8692
Wer: 0.9708
Cer: 0.1439

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: 0.0003
train_batch_size: 16
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 32
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
num_epochs: 40
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
4.5645	9.8765	400	0.8602	0.9935	0.2244
0.2858	19.7531	800	0.8357	0.9838	0.1709
0.1076	29.6296	1200	0.8204	0.9708	0.1525
0.0659	39.5062	1600	0.8692	0.9708	0.1439

Framework versions

Transformers 4.41.2
Pytorch 2.3.0+cu121
Datasets 2.20.0
Tokenizers 0.19.1

jayashreedevi2020/wav2vec2-large-xls-r-300m-assamese_speech_to_IPA_with_wer_cer_metric

wav2vec2-large-xls-r-300m-assamese_speech_to_IPA_with_wer_cer_metric

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

jayashreedevi2020/
wav2vec2-large-xls-r-300m-assamese_speech_to_IPA_with_wer_cer_metric