metadata
language:
- en
license: apache-2.0
base_model: openai/whisper-large-v3
tags:
- hf-asr-leaderboard
- generated_from_trainer
metrics:
- wer
model-index:
- name: Whisper-large-v3-atc
results: []
Whisper-large-v3-atc
This model is a fine-tuned version of openai/whisper-large-v3 on the atc_dataset dataset. It achieves the following results on the evaluation set:
- Loss: 0.2187
- Wer: 7.4754
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: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
0.19 | 0.8446 | 1000 | 0.2450 | 9.4369 |
0.1105 | 1.6892 | 2000 | 0.2092 | 8.4916 |
0.047 | 2.5338 | 3000 | 0.2069 | 7.9794 |
0.017 | 3.3784 | 4000 | 0.2187 | 7.4754 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.2
- Tokenizers 0.19.1