README.md

---
license: cc-by-nc-4.0
base_model: facebook/mms-1b-all
tags:
- generated_from_trainer
datasets:
- common_voice_6_1
metrics:
- wer
model-index:
- name: wav2vec2-large-mms-1b-turkish-colab
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: common_voice_6_1
      type: common_voice_6_1
      config: tr
      split: test
      args: tr
    metrics:
    - name: Wer
      type: wer
      value: 0.22520682259217648
---

<!-- 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-large-mms-1b-turkish-colab

This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the common_voice_6_1 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1602
- Wer: 0.2252

## 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: 0.001
- train_batch_size: 32
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 4
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Wer    |
|:-------------:|:------:|:----:|:---------------:|:------:|
| 7.6235        | 0.9174 | 100  | 3.6396          | 1.0    |
| 0.9369        | 1.8349 | 200  | 0.1817          | 0.2524 |
| 0.286         | 2.7523 | 300  | 0.1670          | 0.2378 |
| 0.2552        | 3.6697 | 400  | 0.1602          | 0.2252 |


### Framework versions

- Transformers 4.41.0.dev0
- Pytorch 2.2.1+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1