Wav2vec2 Large Xlsr Japanese 0325 1200

🚀 Wav2Vec2-Large-XLSR-53-Japanese

This model is a fine - tuned version of facebook/wav2vec2-large-xlsr-53 on Japanese using the Common Voice dataset. It's designed for automatic speech recognition tasks. When using this model, ensure that your speech input is sampled at 16kHz.

🚀 Quick Start

This model is fine - tuned on Japanese using the Common Voice dataset. Make sure your speech input is sampled at 16kHz.

✨ Features

• Multilingual Adaptability: Based on the large - scale XLSR - 53 architecture, it can potentially adapt to multiple languages.
• Fine - Tuned for Japanese: Specifically optimized for Japanese speech recognition.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic 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", "ja", split="test[:2%]")

processor = Wav2Vec2Processor.from_pretrained("qqhann/wav2vec2-large-xlsr-japanese-0325-1200")
model = Wav2Vec2ForCTC.from_pretrained("qqhann/wav2vec2-large-xlsr-japanese-0325-1200")

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])

Advanced Usage

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

processor = Wav2Vec2Processor.from_pretrained("qqhann/wav2vec2-large-xlsr-japanese-0325-1200")
model = Wav2Vec2ForCTC.from_pretrained("qqhann/wav2vec2-large-xlsr-japanese-0325-1200")
model.to("cuda")

chars_to_ignore_regex = '[\,\?\.\!\-\;\:\"\“]'  # TODO: adapt this list to include all special characters you removed from the data
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: XX.XX %

📚 Documentation

Model Information

Training

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

The script used for training can be found here

📄 License

This model is licensed under the apache - 2.0 license.