---
license: apache-2.0
base_model: mattbonnell/wav2vec2-base-wonders-phonemes
tags:
- generated_from_trainer
datasets:
- transcribed_calls
metrics:
- wer
model-index:
- name: wav2vec2-base-wonders-phonemes
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: transcribed_calls
      type: transcribed_calls
      config: default
      split: None
      args: default
    metrics:
    - name: Wer
      type: wer
      value: 0.15521933166430826
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# wav2vec2-base-wonders-phonemes

This model is a fine-tuned version of [mattbonnell/wav2vec2-base-wonders-phonemes](https://huggingface.co/mattbonnell/wav2vec2-base-wonders-phonemes) on the transcribed_calls dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8215
- Wer: 0.1552

## 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: 0.0001
- train_batch_size: 128
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- total_train_batch_size: 512
- total_eval_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.01
- num_epochs: 3000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step  | Validation Loss | Wer    |
|:-------------:|:------:|:-----:|:---------------:|:------:|
| 0.0025        | 50.0   | 500   | 0.7040          | 0.1589 |
| 0.0017        | 100.0  | 1000  | 0.7759          | 0.1549 |
| 0.0014        | 150.0  | 1500  | 0.7451          | 0.1576 |
| 0.0019        | 200.0  | 2000  | 0.7775          | 0.1596 |
| 0.0022        | 250.0  | 2500  | 0.7332          | 0.1626 |
| 0.0018        | 300.0  | 3000  | 0.7833          | 0.1585 |
| 0.0016        | 350.0  | 3500  | 0.7756          | 0.1553 |
| 0.0016        | 400.0  | 4000  | 0.7216          | 0.1610 |
| 0.0019        | 450.0  | 4500  | 0.7868          | 0.1578 |
| 0.0012        | 500.0  | 5000  | 0.7815          | 0.1567 |
| 0.0021        | 550.0  | 5500  | 0.6860          | 0.1570 |
| 0.0014        | 600.0  | 6000  | 0.7580          | 0.1573 |
| 0.0014        | 650.0  | 6500  | 0.7352          | 0.1583 |
| 0.0016        | 700.0  | 7000  | 0.7849          | 0.1569 |
| 0.0016        | 750.0  | 7500  | 0.7794          | 0.1583 |
| 0.0012        | 800.0  | 8000  | 0.7791          | 0.1573 |
| 0.0012        | 850.0  | 8500  | 0.7973          | 0.1571 |
| 0.001         | 900.0  | 9000  | 0.7465          | 0.1551 |
| 0.0019        | 950.0  | 9500  | 0.8304          | 0.1587 |
| 0.0009        | 1000.0 | 10000 | 0.7757          | 0.1579 |
| 0.0013        | 1050.0 | 10500 | 0.7556          | 0.1586 |
| 0.0009        | 1100.0 | 11000 | 0.7624          | 0.1566 |
| 0.0009        | 1150.0 | 11500 | 0.7917          | 0.1566 |
| 0.0013        | 1200.0 | 12000 | 0.7550          | 0.1598 |
| 0.0009        | 1250.0 | 12500 | 0.7517          | 0.1555 |
| 0.0015        | 1300.0 | 13000 | 0.7491          | 0.1586 |
| 0.0013        | 1350.0 | 13500 | 0.8027          | 0.1594 |
| 0.0014        | 1400.0 | 14000 | 0.7748          | 0.1577 |
| 0.0015        | 1450.0 | 14500 | 0.6983          | 0.1587 |
| 0.0009        | 1500.0 | 15000 | 0.7734          | 0.1583 |
| 0.001         | 1550.0 | 15500 | 0.7885          | 0.1570 |
| 0.0009        | 1600.0 | 16000 | 0.7818          | 0.1577 |
| 0.0009        | 1650.0 | 16500 | 0.8128          | 0.1559 |
| 0.0008        | 1700.0 | 17000 | 0.7904          | 0.1554 |
| 0.0011        | 1750.0 | 17500 | 0.7784          | 0.1587 |
| 0.0009        | 1800.0 | 18000 | 0.7910          | 0.1585 |
| 0.0011        | 1850.0 | 18500 | 0.7434          | 0.1570 |
| 0.001         | 1900.0 | 19000 | 0.7241          | 0.1562 |
| 0.0007        | 1950.0 | 19500 | 0.8213          | 0.1541 |
| 0.0008        | 2000.0 | 20000 | 0.7713          | 0.1541 |
| 0.0008        | 2050.0 | 20500 | 0.8151          | 0.1542 |
| 0.0007        | 2100.0 | 21000 | 0.7770          | 0.1557 |
| 0.0006        | 2150.0 | 21500 | 0.8487          | 0.1553 |
| 0.0007        | 2200.0 | 22000 | 0.8332          | 0.1550 |
| 0.0007        | 2250.0 | 22500 | 0.7931          | 0.1559 |
| 0.0006        | 2300.0 | 23000 | 0.8252          | 0.1552 |
| 0.0006        | 2350.0 | 23500 | 0.8609          | 0.1543 |
| 0.0005        | 2400.0 | 24000 | 0.8398          | 0.1543 |
| 0.0006        | 2450.0 | 24500 | 0.7921          | 0.1535 |
| 0.0006        | 2500.0 | 25000 | 0.8223          | 0.1552 |
| 0.0006        | 2550.0 | 25500 | 0.7902          | 0.1561 |
| 0.0005        | 2600.0 | 26000 | 0.7974          | 0.1564 |
| 0.0007        | 2650.0 | 26500 | 0.7761          | 0.1556 |
| 0.0004        | 2700.0 | 27000 | 0.8105          | 0.1548 |
| 0.0004        | 2750.0 | 27500 | 0.7826          | 0.1542 |
| 0.0003        | 2800.0 | 28000 | 0.8056          | 0.1551 |
| 0.0004        | 2850.0 | 28500 | 0.7956          | 0.1534 |
| 0.0005        | 2900.0 | 29000 | 0.8316          | 0.1544 |
| 0.0004        | 2950.0 | 29500 | 0.8164          | 0.1545 |
| 0.0005        | 3000.0 | 30000 | 0.8215          | 0.1552 |


### Framework versions

- Transformers 4.39.1
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2