metadata
language:
- mn
license: apache-2.0
base_model: openai/whisper-large-v3
tags:
- hf-asr-leaderboard
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_16_1
- google/fleurs
metrics:
- wer
model-index:
- name: Whisper Large MN - Ankhbayasgalan Davaadorj
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Common Voice 16.1 & FLEURS
type: mozilla-foundation/common_voice_16_1
config: mn
split: None
args: 'config: mn, split: test+validation'
metrics:
- name: Wer
type: wer
value: 31.994939772289754
Whisper Large MN - Ankhbayasgalan Davaadorj
This model is a fine-tuned version of openai/whisper-large-v3 on the Common Voice 16.1 & FLEURS dataset. It achieves the following results on the evaluation set:
- Loss: 0.5662
- Wer: 31.9949
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: 0.0001
- 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: 10000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0691 | 5.99 | 1000 | 0.4597 | 41.5049 |
| 0.0183 | 11.98 | 2000 | 0.4996 | 38.2982 |
| 0.012 | 17.96 | 3000 | 0.5328 | 38.5402 |
| 0.0091 | 23.95 | 4000 | 0.5619 | 38.1277 |
| 0.004 | 29.94 | 5000 | 0.5439 | 35.2236 |
| 0.0019 | 35.93 | 6000 | 0.5731 | 35.3941 |
| 0.001 | 41.92 | 7000 | 0.5309 | 33.3755 |
| 0.0002 | 47.9 | 8000 | 0.5391 | 32.3140 |
| 0.0 | 53.89 | 9000 | 0.5543 | 32.1984 |
| 0.0 | 59.88 | 10000 | 0.5662 | 31.9949 |
Framework versions
- Transformers 4.38.0.dev0
- Pytorch 2.2.0+cu121
- Datasets 2.17.0
- Tokenizers 0.15.2