Edit model card

Common Voice 16

This model is a fine-tuned version of glob-asr/wav2vec2-large-xls-r-300m-guarani-small on the Common Voice 16 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2425
  • Cer: 5.9170

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: 8
  • eval_batch_size: 16
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: constant_with_warmup
  • lr_scheduler_warmup_steps: 3000
  • training_steps: 3000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Cer
1.2573 0.4955 500 0.3703 8.4904
0.9205 0.9911 1000 0.3224 7.6296
0.7466 1.4866 1500 0.2938 7.1221
0.6766 1.9822 2000 0.2715 6.6510
0.5782 2.4777 2500 0.2831 7.0497
0.5495 2.9732 3000 0.2425 5.9170

Framework versions

  • Transformers 4.44.1
  • Pytorch 2.3.1+cu121
  • Datasets 2.21.0
  • Tokenizers 0.19.1
Downloads last month
95
Safetensors
Model size
315M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for adrianSauer/wav2vec2-cer-extension

Finetuned
(5)
this model

Dataset used to train adrianSauer/wav2vec2-cer-extension