Whisper large V3 to Kaggle Albanian

This model is a fine-tuned version of openai/whisper-large-v3 on the rishabhjain16/kaggle_albanian default dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3189
  • Wer: 23.8109

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: 1e-05
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • training_steps: 5000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
0.2778 0.1 500 0.4945 36.9839
0.4343 0.2 1000 0.4290 35.1835
0.3523 0.3 1500 0.3780 30.6514
0.3833 0.4 2000 0.3436 28.0318
0.2029 1.09 2500 0.3360 26.5128
0.1797 1.19 3000 0.3351 25.8927
0.1687 1.29 3500 0.3249 25.8628
0.1603 1.39 4000 0.3089 23.9877
0.0762 2.08 4500 0.3345 24.5804
0.0676 2.18 5000 0.3189 23.8109

Framework versions

  • Transformers 4.37.2
  • Pytorch 1.14.0a0+44dac51
  • Datasets 2.17.1
  • Tokenizers 0.15.2
Downloads last month
21
Safetensors
Model size
1.54B params
Tensor type
F32
·
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 rishabhjain16/whisper_l3_to_kaggle_sq

Finetuned
(350)
this model

Evaluation results