w2v2bert-test
This model was trained from scratch on the None dataset.
It achieves the following results on the evaluation set:
- eval_loss: 1.0736
- eval_model_preparation_time: 0.0141
- eval_wer: 0.6051
- eval_runtime: 1.6517
- eval_samples_per_second: 13.319
- eval_steps_per_second: 1.816
- step: 0
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: 50
- num_epochs: 5
- mixed_precision_training: Native AMP
Framework versions
- Transformers 4.43.4
- Pytorch 2.1.2+cu118
- Datasets 2.15.0
- Tokenizers 0.19.1
