---
language:
- ne
license: apache-2.0
tags:
- generated_from_trainer
- automatic-speech-recognition
- speech
- openslr
- nepali
datasets:
- spktsagar/openslr-nepali-asr-cleaned
metrics:
- wer
base_model: facebook/wav2vec2-xls-r-300m
model-index:
- name: wav2vec2-large-xls-r-300m-nepali-openslr
  results:
  - task:
      type: automatic-speech-recognition
      name: Nepali Speech Recognition
    dataset:
      name: OpenSLR Nepali ASR
      type: spktsagar/openslr-nepali-asr-cleaned
      config: original
      split: train
    metrics:
    - type: were
      value: 21.27
      name: Test WER
      verified: false
---

<!-- 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-xls-r-300m-nepali-openslr

This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on an [OpenSLR Nepali ASR](https://huggingface.co/datasets/spktsagar/openslr-nepali-asr-cleaned) dataset.
It achieves the following results on the evaluation set:
- eval_loss: 0.1767
- eval_wer: 0.2127
- eval_runtime: 595.3962
- eval_samples_per_second: 36.273
- eval_steps_per_second: 4.535
- epoch: 6.07
- step: 23200

## Model description

Wav2Vec2 is a pretrained model for Automatic Speech Recognition (ASR) and was released in September 2020 by Alexei Baevski, Michael Auli, and Alex Conneau. Soon after the superior performance of Wav2Vec2 was demonstrated on one of the most popular English datasets for ASR, called LibriSpeech, Facebook AI presented a multi-lingual version of Wav2Vec2, called XLSR. XLSR stands for cross-lingual speech representations and refers to model's ability to learn speech representations that are useful across multiple languages.

## How to use?
1. Install transformers and librosa
```
pip install librosa, transformers
```
2. Run the following code which loads your audio file, preprocessor, models, and returns your prediction
```python
import librosa
from transformers import pipeline

audio, sample_rate = librosa.load("<path to your audio file>", sr=16000)
recognizer = pipeline("automatic-speech-recognition", model="spktsagar/wav2vec2-large-xls-r-300m-nepali-openslr")
prediction = recognizer(audio)
```

## Intended uses & limitations

The model is trained on the OpenSLR Nepali ASR dataset, which in itself has some incorrect transcriptions, so it is obvious that the model will not have perfect predictions for your transcript. Similarly, due to colab's resource limit utterances longer than 5 sec are filtered out from the dataset during training and evaluation. Hence, the model might not perform as expected when given audio input longer than 5 sec.

## Training and evaluation data and Training procedure

For dataset preparation and training code, please consult [my blog](https://sagar-spkt.github.io/posts/2022/08/finetune-xlsr-nepali/).

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 0.0003
- 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: 500
- num_epochs: 10
- mixed_precision_training: Native AMP

### Framework versions

- Transformers 4.23.1
- Pytorch 1.11.0+cu113
- Datasets 2.6.0
- Tokenizers 0.13.1