metadata

base_model: openai/whisper-tiny
datasets:
  - fleurs
language:
  - zh
license: apache-2.0
metrics:
  - wer
tags:
  - hf-asr-leaderboard
  - generated_from_trainer
model-index:
  - name: Whisper Tiny Chinese - Chee Li
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: Google Fleurs
          type: fleurs
          config: cmn_hans_cn
          split: None
          args: 'config: zh split: test'
        metrics:
          - type: wer
            value: 38.568340285601195
            name: Wer

Whisper Tiny Chinese - Chee Li

This model is a fine-tuned version of openai/whisper-tiny on the Google Fleurs dataset. It achieves the following results on the evaluation set:

Loss: 0.5500
Wer: 38.5683

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
0.182	4.3668	1000	0.4832	42.5418
0.0473	8.7336	2000	0.5039	38.0568
0.0121	13.1004	3000	0.5371	40.1699
0.0079	17.4672	4000	0.5500	38.5683

Framework versions

Transformers 4.43.4
Pytorch 2.3.1+cu121
Datasets 2.20.0
Tokenizers 0.19.1