README.md

metadata

language: vi
license: apache-2.0
tags:
  - audio
  - automatic-speech-recognition
  - speech
  - xlsr-fine-tuning-week
datasets:
  - common_voice
  - vivos
metrics:
  - wer
base_model: facebook/wav2vec2-large-xlsr-53
model-index:
  - name: Ted Vietnamese XLSR Wav2Vec2 Large 53
    results:
      - task:
          type: automatic-speech-recognition
          name: Speech Recognition
        dataset:
          name: Common Voice vi
          type: common_voice
          args: vi
        metrics:
          - type: wer
            value: 39.571823
            name: Test WER

Wav2Vec2-Large-XLSR-53-Vietnamese

Fine-tuned facebook/wav2vec2-large-xlsr-53 on Vietnamese using the Common Voice, Vivos dataset and FOSD dataset. When using this model, make sure that your speech input is sampled at 16kHz.

Usage

The model can be used directly (without a language model) as follows:

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

test_dataset = load_dataset("common_voice", "vi", split="test")

processor = Wav2Vec2Processor.from_pretrained("not-tanh/wav2vec2-large-xlsr-53-vietnamese")
model = Wav2Vec2ForCTC.from_pretrained("not-tanh/wav2vec2-large-xlsr-53-vietnamese")

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):
  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)
inputs = processor(test_dataset["speech"][:2], sampling_rate=16_000, return_tensors="pt", padding=True)

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

predicted_ids = torch.argmax(logits, dim=-1)

print("Prediction:", processor.batch_decode(predicted_ids))
print("Reference:", test_dataset["sentence"][:2])

Evaluation

The model can be evaluated as follows on the Vietnamese 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", "vi", split="test")
wer = load_metric("wer")

processor = Wav2Vec2Processor.from_pretrained("not-tanh/wav2vec2-large-xlsr-53-vietnamese")
model = Wav2Vec2ForCTC.from_pretrained("not-tanh/wav2vec2-large-xlsr-53-vietnamese")
model.to("cuda")

chars_to_ignore_regex = r'[,?.!\-;:"“%\'�]'
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: 39.571823%

Training

TODO

The Common Voice train, validation, the VIVOS and FOSD datasets were used for training

The script used for training can be found ... # TODO