metadata

language:
  - tr
license: apache-2.0
tags:
  - hf-asr-leaderboard
  - generated_from_trainer
datasets:
  - mozilla-foundation/common_voice_11_0
metrics:
  - wer
base_model: openai/whisper-tiny
model-index:
  - name: Whisper Tiny Tr - Canberk Kandemir
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: Common Voice 11.0
          type: mozilla-foundation/common_voice_11_0
          config: tr
          split: None
          args: 'config: tr, split: test'
        metrics:
          - type: wer
            value: 75.91546835885195
            name: Wer

Whisper Tiny Tr - Canberk Kandemir

This model is a fine-tuned version of openai/whisper-tiny on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

Loss: 0.5366
Wer: 75.9155

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: 3e-05
train_batch_size: 32
eval_batch_size: 16
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 64
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine_with_restarts
lr_scheduler_warmup_steps: 500
training_steps: 4000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.3923	0.89	500	0.5935	73.9613
0.2697	1.77	1000	0.5414	53.0923
0.1784	2.66	1500	0.5194	53.8744
0.1081	3.54	2000	0.5317	64.3962
0.0672	4.43	2500	0.5366	75.9155

Framework versions

Transformers 4.37.2
Pytorch 2.1.0+cu121
Datasets 2.17.0
Tokenizers 0.15.2