metadata

library_name: transformers
license: apache-2.0
base_model: openai/whisper-large-v3
tags:
  - whisper-event
  - generated_from_trainer
datasets:
  - OUTCOMESAI/medical_speech_corpus
metrics:
  - wer
model-index:
  - name: Whisper Small Medical
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: OUTCOMESAI/medical_speech_corpus zh-en
          type: OUTCOMESAI/medical_speech_corpus
        metrics:
          - name: Wer
            type: wer
            value: 44.25531914893617

Whisper Small Medical

This model is a fine-tuned version of openai/whisper-large-v3 on the OUTCOMESAI/medical_speech_corpus zh-en dataset. It achieves the following results on the evaluation set:

Loss: 0.6201
Wer: 44.2553

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: 5e-07
train_batch_size: 32
eval_batch_size: 16
seed: 42
distributed_type: multi-GPU
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: 50
training_steps: 100
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
7.4337	25.0	50	0.6201	44.2553
5.7447	50.0	100	0.6113	51.2340

Framework versions

Transformers 4.48.0.dev0
Pytorch 2.1.0+cu118
Datasets 3.1.1.dev0
Tokenizers 0.21.0