README.md

metadata

language: mn
datasets:
  - common_voice mn
metrics:
  - wer
tags:
  - audio
  - automatic-speech-recognition
  - speech
  - xlsr-fine-tuning-week
license: apache-2.0
model-index:
  - name: XLSR Wav2Vec2 Mongolian V1 by Bayartsogt
    results:
      - task:
          name: Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice mn
          type: common_voice
          args: mn
        metrics:
          - name: Test WER
            type: wer
            value: 35.33

Wav2Vec2-Large-XLSR-53-Mongolian-v1

Fine-tuned facebook/wav2vec2-large-xlsr-53 on Mongolian using the Common Voice.

When using this model, make sure that your speech input is sampled at 16kHz.

Evaluation

The model can be evaluated as follows on the Mongolian test data of Common Voice.

import torch
import torchaudio
from datasets import load_dataset, load_metric
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
import re

test_dataset = load_dataset("common_voice", "mn", split="test")
wer = load_metric("wer")

processor = Wav2Vec2Processor.from_pretrained("bayartsogt/wav2vec2-large-xlsr-mongolian-v1")
model = Wav2Vec2ForCTC.from_pretrained("bayartsogt/wav2vec2-large-xlsr-mongolian-v1")
model.to("cuda")

chars_to_ignore_regex = '[\!\"\'\,\.\«\»\?\-]'
resampler = torchaudio.transforms.Resample(48_000, 16_000)

# Preprocessing the datasets.
# We need to read the aduio files as arrays
def speech_file_to_array_fn(batch):
     batch["sentence"] = re.sub(chars_to_ignore_regex, '', batch["sentence"]).lower()
     speech_array, sampling_rate = torchaudio.load(batch["path"])
     batch["speech"] = resampler(speech_array).squeeze().numpy()
     return batch

test_dataset = test_dataset.map(speech_file_to_array_fn)

# Preprocessing the datasets.
# We need to read the aduio files as arrays
def evaluate(batch):
     inputs = processor(batch["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)

     with torch.no_grad():
          logits = model(inputs.input_values.to("cuda"), attention_mask=inputs.attention_mask.to("cuda")).logits

     pred_ids = torch.argmax(logits, dim=-1)
     batch["pred_strings"] = processor.batch_decode(pred_ids)
     return batch

result = test_dataset.map(evaluate, batched=True, batch_size=8)

print("WER: {:2f}".format(100 * wer.compute(predictions=result["pred_strings"], references=result["sentence"])))

Test Result: 35.33 %

Training

The Common Voice train dataset was used for training as well as ... and ...