๐ Wav2Vec2-Large-XLSR-Turkish
This is a fine - tuned facebook/wav2vec2-large-xlsr-53 model on the Turkish Common Voice dataset. It can be used for Turkish speech recognition, and make sure the speech input is sampled at 16kHz.
๐ Quick Start
When using this model, ensure that your speech input is sampled at 16kHz.
โจ Features
- Fine - tuned Model: Based on facebook/wav2vec2-large-xlsr-53, fine - tuned on the Turkish Common Voice dataset.
- Direct Usage: Can be used directly without a language model.
๐ฆ Installation
No specific installation steps are provided in the original document, so this section is skipped.
๐ป 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", "tr", split="test[:2%]")
processor = Wav2Vec2Processor.from_pretrained("cahya-wirawan/wav2vec2-large-xlsr-turkish")
model = Wav2Vec2ForCTC.from_pretrained("cahya-wirawan/wav2vec2-large-xlsr-turkish")
def speech_file_to_array_fn(batch):
speech_array, sampling_rate = torchaudio.load(batch["path"])
resampler = torchaudio.transforms.Resample(sampling_rate, 16_000)
batch["speech"] = resampler(speech_array).squeeze().numpy()
return batch
test_dataset = test_dataset.map(speech_file_to_array_fn)
inputs = processor(test_dataset[:2]["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)
print("Prediction:", processor.batch_decode(predicted_ids))
print("Reference:", test_dataset[:2]["sentence"])
Advanced Usage
The model can be evaluated as follows on the Turkish 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", "tr", split="test")
wer = load_metric("wer")
processor = Wav2Vec2Processor.from_pretrained("cahya-wirawan/wav2vec2-large-xlsr-turkish")
model = Wav2Vec2ForCTC.from_pretrained("cahya-wirawan/wav2vec2-large-xlsr-turkish")
model.to("cuda")
chars_to_ignore_regex = '[\,\?\.\!\-\;\:\"\โ\โ\โ\'\`โฆ\โยปยซ]'
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"])
resampler = torchaudio.transforms.Resample(sampling_rate, 16_000)
batch["speech"] = resampler(speech_array).squeeze().numpy()
return batch
test_dataset = test_dataset.map(speech_file_to_array_fn)
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: 21.13 %
๐ Documentation
Model Information
Property |
Details |
Model Type |
Wav2Vec2 - Large - XLSR - Turkish |
Training Data |
Turkish Common Voice dataset |
Metrics |
Word Error Rate (WER) |
License |
apache - 2.0 |
Model Index
- Name: XLSR Wav2Vec2 Turkish by Cahya
- Results:
- Task:
- Name: Speech Recognition
- Type: automatic - speech - recognition
- Dataset:
- Name: Common Voice tr
- Type: common_voice
- Args: tr
- Metrics:
- Name: Test WER
- Type: wer
- Value: 21.13
๐ง Technical Details
The script used for training can be found here. The Common Voice train
, validation
, other and invalidated datasets are used in the training process.
๐ License
This model is licensed under the apache - 2.0 license.