Wav2vec2 Base Voxpopuli Sv Swedish

🚀 Wav2vec 2.0 base-voxpopuli-sv-swedish

This is a finetuned version of Facebook's VoxPopuli-sv base model, leveraging NST and Common Voice data. Without a language model, the Word Error Rate (WER) is 5.62% for the NST + Common Voice test set (2% of total sentences) and 19.15% for the Common Voice test set.

🚀 Quick Start

When using this model, ensure that your speech input is sampled at 16kHz.

💻 Usage Examples

Basic Usage

import torch
import torchaudio
from datasets import load_dataset
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
test_dataset = load_dataset("common_voice", "sv-SE", split="test[:2%]").
processor = Wav2Vec2Processor.from_pretrained("KBLab/wav2vec2-base-voxpopuli-sv-swedish")
model = Wav2Vec2ForCTC.from_pretrained("KBLab/wav2vec2-base-voxpopuli-sv-swedish")
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])

📚 Documentation

Dataset Information

Model Evaluation Results

The model named "Wav2vec 2.0 base VoxPopuli-sv swedish" has the following evaluation results:

• For the NST Swedish ASR Database, the Test WER is 5.619804368919309.
• For the Common Voice dataset (sv - SE), the Test WER is 19.145252414798616.

📄 License

This model is released under the cc - by - nc - 4.0 license.