metadata

library_name: transformers
language:
  - uz
license: apache-2.0
base_model: openai/whisper-medium
tags:
  - generated_from_trainer
datasets:
  - mozilla-foundation/common_voice_17_0
metrics:
  - wer
model-index:
  - name: Whisper Medium Uzbek
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice 17.0
          type: mozilla-foundation/common_voice_17_0
          config: uz
          split: None
          args: 'config: uz, split: test'
        metrics:
          - name: Wer
            type: wer
            value: 38.19985168705969

Whisper Medium Uzbek

This model is a fine-tuned version of openai/whisper-medium on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:

Loss: 0.4148
Wer: 38.1999

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: 64
eval_batch_size: 16
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 200
training_steps: 1500

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.6553	0.5330	250	0.5830	51.2637
0.3945	1.0661	500	0.4612	41.6914
0.3352	1.5991	750	0.4360	42.0931
0.2028	2.1322	1000	0.4155	38.1133
0.1956	2.6652	1250	0.4081	37.6900
0.1202	3.1983	1500	0.4148	38.1999

Framework versions

Transformers 4.49.0
Pytorch 2.5.1+cu124
Datasets 3.3.2
Tokenizers 0.21.0