wav2vec2-large-xlsr-turkish-demo-colab

This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4659
• Wer: 0.3465

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.4507	4.0	500	1.4016	0.9587
0.5928	8.0	1000	0.4211	0.4360
0.2228	12.0	1500	0.4629	0.3856
0.1292	16.0	2000	0.4531	0.3654
0.0898	20.0	2500	0.4969	0.3700
0.0613	24.0	3000	0.4731	0.3542
0.046	28.0	3500	0.4659	0.3465

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu118
• Datasets 1.18.3
• Tokenizers 0.13.3