README.md

metadata

language: ru
datasets:
  - common_voice
metrics:
  - wer
  - cer
tags:
  - audio
  - automatic-speech-recognition
  - speech
  - xlsr-fine-tuning-week
license: apache-2.0
model-index:
  - name: XLSR Wav2Vec2 Russian by Jonatas Grosman
    results:
      - task:
          name: Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice ru
          type: common_voice
          args: ru
        metrics:
          - name: Test WER
            type: wer
            value: 16.79
          - name: Test CER
            type: cer
            value: 3.68

Wav2Vec2-Large-XLSR-53-Russian

Fine-tuned facebook/wav2vec2-large-xlsr-53 on Russian using the Common Voice and CSS10. When using this model, make sure that your speech input is sampled at 16kHz.

The script used for training can be found here: https://github.com/jonatasgrosman/wav2vec2-sprint

Usage

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

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

LANG_ID = "ru"
MODEL_ID = "jonatasgrosman/wav2vec2-large-xlsr-53-russian"
SAMPLES = 5

test_dataset = load_dataset("common_voice", LANG_ID, split=f"test[:{SAMPLES}]")

processor = Wav2Vec2Processor.from_pretrained(MODEL_ID)
model = Wav2Vec2ForCTC.from_pretrained(MODEL_ID)

# Preprocessing the datasets.
# We need to read the audio files as arrays
def speech_file_to_array_fn(batch):
    speech_array, sampling_rate = librosa.load(batch["path"], sr=16_000)
    batch["speech"] = speech_array
    batch["sentence"] = batch["sentence"].upper()
    return batch

test_dataset = test_dataset.map(speech_file_to_array_fn)
inputs = processor(test_dataset["speech"], 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)
predicted_sentences = processor.batch_decode(predicted_ids)

for i, predicted_sentence in enumerate(predicted_sentences):
    print("-" * 100)
    print("Reference:", test_dataset[i]["sentence"])
    print("Prediction:", predicted_sentence)

Reference	Prediction
ОН РАБОТАТЬ, А ЕЕ НЕ УДЕРЖАТЬ НИКАК — БЕГАЕТ ЗА КЛЁШЕМ КАЖДОГО БУЛЬВАРНИКА.	ОН РАБОТАТЬ А ЕЕ НЕ УДЕРЖАТНИКАК БЕГАЕТ ЗА КЛЕШОМ КАЖДОГО БУЛЬВАРНИКА
ЕСЛИ НЕ БУДЕТ ВОЗРАЖЕНИЙ, Я БУДУ СЧИТАТЬ, ЧТО АССАМБЛЕЯ СОГЛАСНА С ЭТИМ ПРЕДЛОЖЕНИЕМ.	ЕСЛИ НЕ БУДЕТ ВОЗРАЖЕНИЙ Я БУДУ СЧИТАТЬ ЧТО АССАМБЛЕЯ СОГЛАСНА С ЭТИМ ПРЕДЛОЖЕНИЕМ
ПАЛЕСТИНЦАМ НЕОБХОДИМО СНАЧАЛА УСТАНОВИТЬ МИР С ИЗРАИЛЕМ, А ЗАТЕМ ДОБИВАТЬСЯ ПРИЗНАНИЯ ГОСУДАРСТВЕННОСТИ.	ПАЛЕСТИНЦАМ НЕОБХОДИМО СНАЧАЛА УСТАНОВИТЬ С НИ МИР С ИЗРАИЛЕМ А ЗАТЕМ ДОБИВАТЬСЯ ПРИЗНАНИЯ ГОСУДАРСТВЕННОВСКИЙ
У МЕНЯ БЫЛО ТАКОЕ ЧУВСТВО, ЧТО ЧТО-ТО ТАКОЕ ОЧЕНЬ ВАЖНОЕ Я ПРИБАВЛЯЮ.	У МЕНЯ БЫЛО ТАКОЕ ЧУВСТВО ЧТО ЧТО-ТО ТАКОЕ ОЧЕНЬ ВАЖНОЕ Е ПРЕДБАВЛЯЕТ
ТОЛЬКО ВРЯД ЛИ ПОЙМЕТ.	ТОЛЬКО ВРЯД ЛИ ПОЙМЕТ

Evaluation

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

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

LANG_ID = "ru"
MODEL_ID = "jonatasgrosman/wav2vec2-large-xlsr-53-russian"
DEVICE = "cuda"

CHARS_TO_IGNORE = [",", "?", "¿", ".", "!", "¡", ";", ":", '""', "%", '"', "�", "ʿ", "·", "჻", "~", "՞", 
                   "؟", "،", "।", "॥", "«", "»", "„", "“", "”", "「", "」", "‘", "’", "《", "》", "(", ")", "[", "]",
                   "=", "`", "_", "+", "<", ">", "…", "–", "°", "´", "ʾ", "‹", "›", "©", "®", "—", "→", "。"]

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

wer = load_metric("wer.py") # https://github.com/jonatasgrosman/wav2vec2-sprint/blob/main/wer.py
cer = load_metric("cer.py") # https://github.com/jonatasgrosman/wav2vec2-sprint/blob/main/cer.py

chars_to_ignore_regex = f"[{re.escape(''.join(CHARS_TO_IGNORE))}]"

processor = Wav2Vec2Processor.from_pretrained(MODEL_ID)
model = Wav2Vec2ForCTC.from_pretrained(MODEL_ID)
model.to(DEVICE)

# Preprocessing the datasets.
# We need to read the audio files as arrays
def speech_file_to_array_fn(batch):
    with warnings.catch_warnings():
        warnings.simplefilter("ignore")
        speech_array, sampling_rate = librosa.load(batch["path"], sr=16_000)
    batch["speech"] = speech_array
    batch["sentence"] = re.sub(chars_to_ignore_regex, "", batch["sentence"]).upper()
    return batch

test_dataset = test_dataset.map(speech_file_to_array_fn)

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

\\twith torch.no_grad():
\\t\\tlogits = model(inputs.input_values.to(DEVICE), attention_mask=inputs.attention_mask.to(DEVICE)).logits

\\tpred_ids = torch.argmax(logits, dim=-1)
\\tbatch["pred_strings"] = processor.batch_decode(pred_ids)
\\treturn batch

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

predictions = [x.upper() for x in result["pred_strings"]]
references = [x.upper() for x in result["sentence"]]

print(f"WER: {wer.compute(predictions=predictions, references=references, chunk_size=1000) * 100}")
print(f"CER: {cer.compute(predictions=predictions, references=references, chunk_size=1000) * 100}")

Test Result:

In the table below I report the Word Error Rate (WER) and the Character Error Rate (CER) of the model. I ran the evaluation script described above on other models as well (on 2021-04-22). Note that the table below may show different results from those already reported, this may have been caused due to some specificity of the other evaluation scripts used.

Model	WER	CER
jonatasgrosman/wav2vec2-large-xlsr-53-russian	16.79%	3.68%
anton-l/wav2vec2-large-xlsr-53-russian	19.49%	4.15%