wav2vec2-large-xlsr-sw_ndizi_782_without_NF4

This model is a fine-tuned version of smutuvi/wav2vec2-large-xlsr-sw on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.0702
• Wer: 0.4635

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: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.3466	5.97	400	1.0027	0.4830
0.7355	11.94	800	1.1488	0.4873
0.4853	17.91	1200	1.4337	0.4678
0.3254	23.88	1600	1.5478	0.4895
0.2251	29.85	2000	1.7495	0.4736
0.1674	35.82	2400	1.8199	0.4751
0.1258	41.79	2800	2.0479	0.4606
0.1096	47.76	3200	2.0702	0.4635

Framework versions

• Transformers 4.37.1
• Pytorch 2.2.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0