metadata
license: mit
base_model: facebook/w2v-bert-2.0
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: w2v2-bert-urdu
results: []
w2v2-bert-urdu
This model is a fine-tuned version of facebook/w2v-bert-2.0 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6237
- Wer: 0.4732
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-06
- 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: 100
- num_epochs: 2
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 7.3111 | 0.1695 | 50 | 4.0973 | 1.0 |
| 3.6426 | 0.3390 | 100 | 3.0408 | 1.0 |
| 2.7471 | 0.5085 | 150 | 2.0725 | 0.9836 |
| 1.4561 | 0.6780 | 200 | 0.9029 | 0.5519 |
| 0.85 | 0.8475 | 250 | 0.6233 | 0.4219 |
| 0.6703 | 1.0169 | 300 | 0.5772 | 0.4590 |
| 0.6025 | 1.1864 | 350 | 0.5479 | 0.4077 |
| 0.633 | 1.3559 | 400 | 0.6068 | 0.4798 |
| 0.6775 | 1.5254 | 450 | 0.6257 | 0.4787 |
| 0.7196 | 1.6949 | 500 | 0.6241 | 0.4765 |
| 0.6955 | 1.8644 | 550 | 0.6237 | 0.4732 |
Framework versions
- Transformers 4.40.2
- Pytorch 2.2.1+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1