metadata
base_model: ylacombe/w2v-bert-2.0
tags:
- generated_from_trainer
datasets:
- common_voice_16_0
metrics:
- wer
model-index:
- name: w2v-bert-2.0-mongolian-colab-CV16.0
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: common_voice_16_0
type: common_voice_16_0
config: mn
split: test
args: mn
metrics:
- name: Wer
type: wer
value: 0.3251033282575593
w2v-bert-2.0-mongolian-colab-CV16.0
This model is a fine-tuned version of ylacombe/w2v-bert-2.0 on the common_voice_16_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.5032
- Wer: 0.3251
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-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.7516 | 0.79 | 100 | 2.4041 | 1.0089 |
| 1.0185 | 1.58 | 200 | 0.7642 | 0.6153 |
| 0.5366 | 2.37 | 300 | 0.6518 | 0.5328 |
| 0.4153 | 3.16 | 400 | 0.6116 | 0.4811 |
| 0.353 | 3.95 | 500 | 0.6357 | 0.4806 |
| 0.2876 | 4.74 | 600 | 0.6213 | 0.4434 |
| 0.2389 | 5.53 | 700 | 0.5103 | 0.4243 |
| 0.1735 | 6.32 | 800 | 0.5079 | 0.3753 |
| 0.1419 | 7.11 | 900 | 0.5264 | 0.3638 |
| 0.1031 | 7.91 | 1000 | 0.5454 | 0.3466 |
| 0.0743 | 8.7 | 1100 | 0.5286 | 0.3337 |
| 0.054 | 9.49 | 1200 | 0.5032 | 0.3251 |
Framework versions
- Transformers 4.37.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0