README.md

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: 4.988756935106611

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: 0.1252
• Wer: 4.9888

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: 8
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 2500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2346	0.2	500	0.1957	8.5131
0.1252	0.4	1000	0.1448	5.7876
0.2076	0.6	1500	0.1361	5.5786
0.2356	0.8	2000	0.1504	6.6611
0.1893	1.0	2500	0.1252	4.9888

Framework versions

• Transformers 4.25.0.dev0
• Pytorch 1.12.0
• Datasets 2.6.2.dev0
• Tokenizers 0.12.1