metadata
language:
- es
license: apache-2.0
tags:
- hf-asr-leaderboard
- generated_from_trainer
datasets:
- facebook/multilingual_librispeech
metrics:
- wer
model-index:
- name: Whisper Small Es - Sanchit Gandhi
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Multilingual LibriSpeech
type: facebook/multilingual_librispeech
args: 'config: es, split: test'
metrics:
- name: Wer
type: wer
value: 4.0192544548525975
Whisper Small Es - Sanchit Gandhi
This model is a fine-tuned version of openai/whisper-medium on the Multilingual LibriSpeech dataset. It achieves the following results on the evaluation set:
- Loss: 0.0969
- Wer: 4.0193
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: 2
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 32
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
0.1438 | 0.2 | 1000 | 0.1414 | 6.4317 |
0.1294 | 0.4 | 2000 | 0.1139 | 4.7176 |
0.2289 | 0.6 | 3000 | 0.1048 | 4.3266 |
0.1164 | 1.11 | 4000 | 0.0964 | 3.9164 |
0.1609 | 1.31 | 5000 | 0.0969 | 4.0193 |
Framework versions
- Transformers 4.25.0.dev0
- Pytorch 1.12.0
- Datasets 2.6.2.dev0
- Tokenizers 0.12.1