README.md

metadata

license: mit
tags:
  - generated_from_trainer
base_model: facebook/w2v-bert-2.0
datasets:
  - common_voice_16_0
metrics:
  - wer
model-index:
  - name: w2v-bert-2.0-mongolian-colab-CV16.0
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: common_voice_16_0
          type: common_voice_16_0
          config: mn
          split: test
          args: mn
        metrics:
          - type: wer
            value: 0.32733304328910157
            name: Wer

w2v-bert-2.0-mongolian-colab-CV16.0

This model is a fine-tuned version of facebook/w2v-bert-2.0 on the common_voice_16_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.5090
• Wer: 0.3273

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
1.8026	2.3715	300	0.6395	0.5274
0.3561	4.7431	600	0.5804	0.4247
0.1776	7.1146	900	0.5514	0.3697
0.0764	9.4862	1200	0.5090	0.3273

Framework versions

• Transformers 4.41.0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1