---
library_name: transformers
license: apache-2.0
base_model: facebook/wav2vec2-xls-r-1b
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-xls-r-1b-scandinavian-E4-100h-30-epochs-20250208_v3
  results: []
---

<!-- 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-xls-r-1b-scandinavian-E4-100h-30-epochs-20250208_v3

This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: inf
- Wer: 17.8979
- Cer: 4.8286

## 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: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 5000
- num_epochs: 8
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step  | Validation Loss | Wer     | Cer     |
|:-------------:|:------:|:-----:|:---------------:|:-------:|:-------:|
| 2.6196        | 0.3909 | 500   | inf             | 100.0   | 65.7730 |
| 1.0801        | 0.7819 | 1000  | inf             | 38.9103 | 10.8535 |
| 0.69          | 1.1728 | 1500  | inf             | 28.1298 | 7.8743  |
| 0.5238        | 1.5637 | 2000  | inf             | 24.6450 | 6.9218  |
| 0.4787        | 1.9547 | 2500  | inf             | 23.5140 | 6.6324  |
| 0.41          | 2.3456 | 3000  | inf             | 23.2639 | 6.4844  |
| 0.3868        | 2.7365 | 3500  | inf             | 23.0202 | 6.4139  |
| 0.4987        | 3.1274 | 4000  | inf             | 22.1255 | 6.1609  |
| 0.5173        | 3.5184 | 4500  | inf             | 22.0348 | 6.1361  |
| 0.5159        | 3.9093 | 5000  | inf             | 22.2732 | 6.2526  |
| 0.2712        | 4.3002 | 5500  | inf             | 22.0759 | 6.1940  |
| 0.2565        | 4.6912 | 6000  | inf             | 21.4397 | 5.9482  |
| 0.4129        | 5.0821 | 6500  | inf             | 21.1116 | 5.8439  |
| 0.3925        | 5.4730 | 7000  | inf             | 21.1506 | 5.9633  |
| 0.3943        | 5.8640 | 7500  | inf             | 19.8160 | 5.4331  |
| 0.1971        | 6.2549 | 8000  | inf             | 18.7778 | 5.1171  |
| 0.1898        | 6.6458 | 8500  | inf             | 18.2018 | 4.8967  |
| 0.2553        | 7.0367 | 9000  | inf             | 18.1142 | 4.9026  |
| 0.2291        | 7.4277 | 9500  | inf             | 17.5730 | 4.7604  |
| 0.2655        | 7.8186 | 10000 | inf             | 17.8979 | 4.8286  |


### Framework versions

- Transformers 4.48.3
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0