---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: xls-r-asr_af-run1
  results: []
datasets:
- lucas-meyer/asr_af
---

<!-- 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. -->

# xls-r-asr_af-run1

This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the asr_af dataset.
It achieves the following results:
- Wer (Validation): 42.46%
- Wer (Test): 43.31%

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 3
- total_train_batch_size: 12
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 30
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Wer (Train)    |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 11.8837       | 0.22  | 50   | 7.3321          | 1.0    |
| 4.9879        | 0.44  | 100  | 4.0442          | 1.0    |
| 3.5299        | 0.66  | 150  | 3.1941          | 1.0    |
| 3.0621        | 0.88  | 200  | 2.9995          | 1.0    |
| 2.9941        | 1.1   | 250  | 3.0224          | 1.0    |
| 2.9668        | 1.32  | 300  | 2.9470          | 1.0    |
| 2.9351        | 1.54  | 350  | 2.9053          | 1.0    |
| 2.5839        | 1.76  | 400  | 1.9778          | 0.9929 |
| 1.7074        | 1.98  | 450  | 1.3715          | 0.8741 |
| 1.2129        | 2.2   | 500  | 1.2575          | 0.9809 |
| 1.0397        | 2.42  | 550  | 1.0340          | 0.7792 |
| 0.9081        | 2.64  | 600  | 0.9177          | 0.7051 |
| 0.7855        | 2.86  | 650  | 0.7769          | 0.6323 |
| 0.7099        | 3.08  | 700  | 0.6570          | 0.5771 |
| 0.5658        | 3.3   | 750  | 0.6074          | 0.5456 |
| 0.5115        | 3.52  | 800  | 0.5645          | 0.5291 |
| 0.5005        | 3.74  | 850  | 0.5539          | 0.5124 |
| 0.4892        | 3.96  | 900  | 0.5390          | 0.5105 |
| 0.3787        | 4.19  | 950  | 0.5359          | 0.4879 |
| 0.3559        | 4.41  | 1000 | 0.5400          | 0.4904 |
| 0.3349        | 4.63  | 1050 | 0.5019          | 0.4672 |
| 0.3374        | 4.85  | 1100 | 0.4940          | 0.4519 |
| 0.3423        | 5.07  | 1150 | 0.5030          | 0.4364 |
| 0.257         | 5.29  | 1200 | 0.4646          | 0.4243 |
| 0.2431        | 5.51  | 1250 | 0.4803          | 0.4330 |
| 0.2598        | 5.73  | 1300 | 0.5143          | 0.4398 |
| 0.2602        | 5.95  | 1350 | 0.4789          | 0.4201 |


### Framework versions

- Transformers 4.28.0
- Pytorch 2.0.1+cu117
- Datasets 2.14.4
- Tokenizers 0.13.3