metadata

library_name: transformers
language:
  - en
license: apache-2.0
base_model: openai/whisper-base
tags:
  - generated_from_trainer
datasets:
  - iFaz/Whisper_Compatible_SER_benchmark
metrics:
  - wer
model-index:
  - name: whisper-SER-base-v1
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Whisper_Compatible_SER_benchmark(Not train_augmented)
          type: iFaz/Whisper_Compatible_SER_benchmark
          args: 'config: en, split: test'
        metrics:
          - name: Wer
            type: wer
            value: 105.45094152626362

whisper-SER-base-v1

This model is a fine-tuned version of openai/whisper-base on the Whisper_Compatible_SER_benchmark(Not train_augmented) dataset. It achieves the following results on the evaluation set:

Loss: 0.8757
Wer: 105.4509

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: 32
eval_batch_size: 1
seed: 42
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: 500
training_steps: 6000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1761	2.4450	1000	0.5625	48.9594
0.0796	4.8900	2000	0.5905	87.2151
0.0201	7.3350	3000	0.7191	125.5203
0.0054	9.7800	4000	0.7985	127.7998
0.0012	12.2249	5000	0.8611	108.0278
0.0008	14.6699	6000	0.8757	105.4509

Framework versions

Transformers 4.48.0
Pytorch 2.5.1+cu121
Datasets 3.2.0
Tokenizers 0.21.0