metadata

library_name: transformers
license: apache-2.0
base_model: openai/whisper-small
tags:
  - generated_from_trainer
datasets:
  - Leonel-Maia/fongbe-splitted
metrics:
  - wer
model-index:
  - name: whisper-small-splitted
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Leonel-Maia/fongbe-splitted
          type: Leonel-Maia/fongbe-splitted
        metrics:
          - name: Wer
            type: wer
            value: 0.11257288805358778

whisper-small-splitted

This model is a fine-tuned version of openai/whisper-small on the Leonel-Maia/fongbe-splitted dataset. It achieves the following results on the evaluation set:

Loss: 0.1084
Wer: 0.1126

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: 4
seed: 42
gradient_accumulation_steps: 8
total_train_batch_size: 64
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
num_epochs: 60.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1324	2.1094	500	0.1909	0.3307
0.0383	4.2187	1000	0.1153	0.1131
0.0184	6.3281	1500	0.1084	0.1126
0.0098	8.4374	2000	0.1122	0.1014
0.0076	10.5468	2500	0.1101	0.0966
0.0075	12.6562	3000	0.1156	0.0992

Framework versions

Transformers 4.50.3
Pytorch 2.6.0+cu124
Datasets 3.5.0
Tokenizers 0.21.1