Whisper Medium Commonvoice17 Bn

--- library_name: transformers license: apache-2.0 base_model: openai/whisper-medium tags: - generated_from_trainer datasets: - common_voice_17_0 metrics: - wer model-index: - name: whisper-medium-commonvoice17-bn results: - task: name: Automatic Speech Recognition type: automatic-speech-recognition dataset: name: common_voice_17_0 type: common_voice_17_0 config: bn split: None args: bn metrics: - name: Wer type: wer value: 24.2615 --- # whisper-medium-commonvoice17-bn This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the common_voice_17_0 dataset. It achieves the following results on the evaluation set: - Loss: 0.1528 - Wer: 24.2615 - Cer: 6.9685 ## 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: 8 - eval_batch_size: 8 - seed: 42 - gradient_accumulation_steps: 4 - total_train_batch_size: 32 - optimizer: Use 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: 1000 - training_steps: 6000 ### Training results | Training Loss | Epoch | Step | Validation Loss | Wer | Cer | |:-------------:|:------:|:----:|:---------------:|:-------:|:-------:| | 0.1052 | 1.5072 | 1000 | 0.1141 | 34.6308 | 10.3753 | | 0.0296 | 3.0143 | 2000 | 0.0903 | 27.493 | 7.9059 | | 0.0158 | 4.5215 | 3000 | 0.0958 | 25.8633 | 7.4133 | | 0.0045 | 6.0286 | 4000 | 0.1198 | 25.2698 | 7.2153 | | 0.0028 | 7.5358 | 5000 | 0.1401 | 24.7517 | 7.1214 | | 0.0001 | 9.0430 | 6000 | 0.1528 | 24.2615 | 6.9685 | ### Framework versions - Transformers 4.46.3 - Pytorch 2.4.0+cu121 - Datasets 2.21.0 - Tokenizers 0.20.3