.ipynb_checkpoints/README-checkpoint.md

metadata

language: hi
datasets:
  - common_voice
  - indic tts
  - iiith
metrics:
  - wer
tags:
  - audio
  - automatic-speech-recognition
  - speech
  - xlsr-fine-tuning-week
license: apache-2.0
model-index:
  - name: Hindi XLSR Wav2Vec2 Large 53
results:
  - task:
      name: Speech Recognition
      type: automatic-speech-recognition
      dataset:
        - name: Common Voice hi
          type: common_voice
          args: hi
        - name: Indic IIT (IITM)
          type: indic
          args: hi
        - name: IIITH Indic Dataset
          type: iiith
          args: hi
      metrics:
        - name: Test WER
          type: wer
          value: 19.05

Wav2Vec2-Large-XLSR-53-Hindi

Fine-tuned facebook/wav2vec2-large-xlsr-53 on Hindi using the following datasets:

• Common Voice,
• Indic TTS- IITM and
• IIITH - Indic Speech Datasets

The Hindi CommonVoice data is skewed towards male voices. However the other Indic datasets are well balanced.

Fine-tuned on facebook/wav2vec2-large-xlsr-53 using Hindi dataset :: 30 epochs >> 19.05% WER Resuming from checkpoints trained for another XX epochs >> XX.XX%

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", "hi", split="test")

processor = Wav2Vec2Processor.from_pretrained("skylord/wav2vec2-large-xlsr-hindi") 
model = Wav2Vec2ForCTC.from_pretrained("skylord/wav2vec2-large-xlsr-hindi") 

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 Hindi 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", "hi", split="test") 
wer = load_metric("wer")

processor = Wav2Vec2Processor.from_pretrained("skylord/wav2vec2-large-xlsr-hindi") 
model = Wav2Vec2ForCTC.from_pretrained("skylord/wav2vec2-large-xlsr-hindi")
model.to("cuda")

chars_to_ignore_regex = '[\,\?\.\!\-\'\;\:\"\“\%\‘\”\�Utrnle\_]'
unicode_ignore_regex = r'[dceMaWpmFui\xa0\u200d]' # Some unwanted unicode chars
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"]).sub(unicode_ignore_regex, '', batch["sentence"])
  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: 19.056 %

Training

The Common Voice train, validation, datasets were used for training as well as

The script used for training & wandb dashboard can be found here