Edit model card

wav2vec2-common_voice-tr-demo

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the COMMON_VOICE - TR dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3709
  • Wer: 0.3429

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: 15.0

Training results

Training Loss Epoch Step Validation Loss Wer
No log 0.92 100 3.5988 1.0
No log 1.83 200 3.0083 0.9999
No log 2.75 300 0.8642 0.7579
No log 3.67 400 0.5713 0.6203
3.14 4.59 500 0.4795 0.5338
3.14 5.5 600 0.4441 0.4912
3.14 6.42 700 0.4241 0.4521
3.14 7.34 800 0.4326 0.4611
3.14 8.26 900 0.3913 0.4212
0.2183 9.17 1000 0.4036 0.3973
0.2183 10.09 1100 0.4035 0.3959
0.2183 11.01 1200 0.3807 0.3790
0.2183 11.93 1300 0.3750 0.3650
0.2183 12.84 1400 0.3822 0.3573
0.1011 13.76 1500 0.3747 0.3510
0.1011 14.68 1600 0.3714 0.3454

Framework versions

  • Transformers 4.32.0.dev0
  • Pytorch 2.0.1+cu118
  • Datasets 2.14.0
  • Tokenizers 0.13.3
Downloads last month
36
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for xiaoming-leza/wav2vec2-common_voice-tr-demo

Finetuned
(204)
this model

Dataset used to train xiaoming-leza/wav2vec2-common_voice-tr-demo

Evaluation results